RE: filmscanners: LED Illumination for Film Scanners
> 4. Ability to control intensity of each color > illuminant separately -- eg., the "Analog Gain" > control in NikonScan. The Leaf does that by using three scans, and controlling each scan... > What I don't yet understand is how the illuminant > is evenly distributed over the film width, or how > the uniformity of a LED illuminant compares with > cold-cathode or fluorescent light sources. Yep, that was my main concern, but not between LEDs so much as just for one LED unto it self! There is also space between LEDs, so I just don't see how this can be made to give near the evenness of illumination that you can get with a cold light source (non-point also). > However, similar issues exist for cold cathode > and fluorescent lamps. Not near as drastically though. The light is typically diffused...but I guess you could do that to the LEDs too. > (And you have to admit, a LED power supply is a > lot simpler than the high-voltage, high-frequency > supply needed for fluorescent bulbs.) And the fact that they run a LOT cooler. But I do have a question as to wavelength. > A further advantage to the LED illumination > scheme is that it works with a monochrome CCD sensor. He he, or of the three pass scanner, but I really don't consider that an advantage in my case ;-) > The monochrome sensor has one less source of > non-uniformity, since it doesn't have any > color filtering over the sensor itself. (I > imagine these filters themselves must introduce > some degree of non-uniformity.) Yes, but the LED it self isn't going to be uniform even across one LED... > The Leaf also "works around" this issue by > doing three scans -- presumably using a different > filter with each pass. But the Leaf can't > control illuminant intensity or spectral content > the way that the Nikon scanners do. I don't quite know what you mean by that... How does the Nikon control spectral content? They don't change the color of the LEDs, do they? The color is fixed... The Leaf does control the lamp brightness, and the lamp brightness of a non-point linear diffuse light source is going to be quite a bit more even than an array of LEDs...
RE: filmscanners: LED Illumination for Film Scanners
[rafe b:] >> The Leaf also "works around" this issue by >> doing three scans -- presumably using a different >> filter with each pass. But the Leaf can't >> control illuminant intensity or spectral content >> the way that the Nikon scanners do. [Austin:] >I don't quite know what you mean by that... How does the Nikon control >spectral content? They don't change the color of the LEDs, do they? The >color is fixed... My mistake. I suppose if Nikon is using just R, G and B leds, there's no tweaking of spectral content. That idea comes from the Kodak white paper I cited earlier, where LEDs of several different hues were employed. >The Leaf does control the lamp brightness, and the lamp brightness of a >non-point linear diffuse light source is going to be quite a bit more even >than an array of LEDs... I don't know how they do it, but apparently they and many others do it with a good deal of success. Nikon's been using LEDs in their film scanners for several years. If there were a systematic problem with this, I think we'd have heard about it by now. It could well be that the LEDs are far enough from the focal plane so that they appear diffuse. That's purely a guess on my part. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
> It could well be that the LEDs are far enough from > the focal plane so that they appear diffuse. That's > purely a guess on my part. That makes sense. LEDs typically have a lense of some kind... Do they have a diffuser over the light source that you know of? I still have my druthers about using LEDs as a film scanner light source though, but I really haven't researched it any, and it is interesting.
Re: filmscanners: LED Illumination for Film Scanners
> > It could well be that the LEDs are far enough from > > the focal plane so that they appear diffuse. That's > > purely a guess on my part. > > That makes sense. LEDs typically have a lense of some kind... Do they have > a diffuser over the light source that you know of? I still have my druthers > about using LEDs as a film scanner light source though, but I really haven't > researched it any, and it is interesting. LED's can be manufactured to a wide variety of specifications, and that's what makes them so versatile. A couple of years ago someone on the darkroom newsgroup was working on an LED light source for enlarger heads, utilizing clusters of high-intensity LED's. I don't know what happened to the project, but at the time a lot of people were really excited about the technology and the initial results showed a lot of promise. http://www.trailing-edge.com/www/led.html
RE: filmscanners: LED Illumination for Film Scanners
> > What I don't yet understand is how the illuminant > > is evenly distributed over the film width, > > Lots of LEDS, spaced to give even illumination. But that's the point...you CAN'T space them to give even illumination. Just a single LED is unevenly illuminated in and of it self! It's typically a mounded plastic piece, which is really not very consistent.
Re: filmscanners: LED Illumination for Film Scanners
On Tue, 19 Jun 2001 07:30:02 -0400 (EDT) Raphael Bustin ([EMAIL PROTECTED]) wrote: > What I don't yet understand is how the illuminant > is evenly distributed over the film width, Lots of LEDS, spaced to give even illumination. Regards Tony Sleep http://www.halftone.co.uk - Online portfolio & exhibit; + film scanner info & comparisons
RE: filmscanners: LED Illumination for Film Scanners
> A couple of years ago someone on the darkroom newsgroup was working on an > LED light source for enlarger heads, utilizing clusters of high-intensity > LED's. I don't know what happened to the project, but at the time a lot of > people were really excited about the technology and the initial results > showed a lot of promise. > > http://www.trailing-edge.com/www/led.html Yes, and look at how uneven the lighting is. For an enlarger, that may turn out to be a very bad idea, simply because you can't adjust each individual LED for even illumination. At least with a CCD, you can adjust the gain for each sensor element to get even illumination.
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: > > > A couple of years ago someone on the darkroom newsgroup was working on an > > LED light source for enlarger heads, utilizing clusters of high-intensity > > LED's. I don't know what happened to the project, but at the time a lot of > > people were really excited about the technology and the initial results > > showed a lot of promise. > > > > http://www.trailing-edge.com/www/led.html > > Yes, and look at how uneven the lighting is. For an enlarger, that may turn > out to be a very bad idea, simply because you can't adjust each individual > LED for even illumination. At least with a CCD, you can adjust the gain for > each sensor element to get even illumination. All that you would need to do is shoot this array through some sort of diffusing device. After all, light bulbs aren't exactly evenly illuminated and those have been used for years in enlargers... Isaac
RE: filmscanners: LED Illumination for Film Scanners
At 09:44 AM 6/20/01 -0400, Austin wrote: >Yes, and look at how uneven the lighting is. For an enlarger, that may turn >out to be a very bad idea, simply because you can't adjust each individual >LED for even illumination. At least with a CCD, you can adjust the gain for >each sensor element to get even illumination. You *have* to adjust the gain for each sensor element. It's a given. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
> Austin Franklin wrote: > > > > > A couple of years ago someone on the darkroom newsgroup was > working on an > > > LED light source for enlarger heads, utilizing clusters of > high-intensity > > > LED's. I don't know what happened to the project, but at the > time a lot of > > > people were really excited about the technology and the > initial results > > > showed a lot of promise. > > > > > > http://www.trailing-edge.com/www/led.html > > > > Yes, and look at how uneven the lighting is. For an enlarger, > that may turn > > out to be a very bad idea, simply because you can't adjust each > individual > > LED for even illumination. At least with a CCD, you can adjust > the gain for > > each sensor element to get even illumination. > > All that you would need to do is shoot this array through > some sort of > diffusing device. And that was the question, was there a diffuser of some sort, and that has gone unanswered. > After all, light bulbs aren't exactly evenly > illuminated and those have been used for years in enlargers... Yes, and incandescent light bulbs used in an enlarger get a condenser system, which is entirely different than a cold light head, which is basically what most scanners use for illumination.
Re: filmscanners: LED Illumination for Film Scanners
> > A couple of years ago someone on the darkroom newsgroup was working on an > > LED light source for enlarger heads, utilizing clusters of high-intensity > > LED's. I don't know what happened to the project, but at the time a lot of > > people were really excited about the technology and the initial results > > showed a lot of promise. > > > > http://www.trailing-edge.com/www/led.html > > Yes, and look at how uneven the lighting is. Quoting a recent post on this list: "And you can tell this from a 72dpi JPEG image?" > For an enlarger, that may turn > out to be a very bad idea, simply because you can't adjust each individual > LED for even illumination. At least with a CCD, you can adjust the gain for > each sensor element to get even illumination. Why can't you adjust each LED for illumination levels? For instance, the designer could quite easily vary the voltage levels throughout the array to compensate for enlarger lens light fall-off. And taking things one step further, a dense LED array positioned closer to the negative could even be programmed to provide some degree of selective dodging/burning/variable constrast control. With an appropriate control mechanism, a user could adjust for dead even lighting across the easel for a specific lens/format size combination.
RE: filmscanners: LED Illumination for Film Scanners
> > > A couple of years ago someone on the darkroom newsgroup was working on > an > > > LED light source for enlarger heads, utilizing clusters of > high-intensity > > > LED's. I don't know what happened to the project, but at the > time a lot > of > > > people were really excited about the technology and the > initial results > > > showed a lot of promise. > > > > > > http://www.trailing-edge.com/www/led.html > > > > Yes, and look at how uneven the lighting is. > > Quoting a recent post on this list: "And you can tell this from a > 72dpi JPEG > image?" Not relevant. Here we are not talking about image quality garnered from a 72PPI image. Go to the referenced URL above. Anyone can see that the LEDs used in that array have very large gaps between them, and the illumination is VERY uneven...even if it was 10PPI. Go about halfway down the page. > > For an enlarger, that may turn > > out to be a very bad idea, simply because you can't adjust each > individual > > LED for even illumination. At least with a CCD, you can adjust the gain > for > > each sensor element to get even illumination. > > Why can't you adjust each LED for illumination levels? For instance, the > designer could quite easily vary the voltage levels throughout > the array to > compensate for enlarger lens light fall-off. Yes, but that's not the issue. > And taking things one step > further, a dense LED array positioned closer to the negative could even be > programmed to provide some degree of selective dodging/burning/variable > constrast control. I doubt it. The control isn't that fine since the array is so course. Also it would make the unevenness worse. > With an appropriate control mechanism, a user could > adjust for dead even lighting across the easel for a specific lens/format > size combination. Not with that array shown, and not without a VERY diffuse diffuser. Making an LED array as shown is not a high level of sophistication. If it truly worked well, I would believe that it would be being used by the enlarger manufacturers, or at least offered in the aftermarket. I don't believe either is true, or does anyone know differently? The use of an LED array as a light source for a film enlarger is entirely different than using an LED array for a scanner, where you have individual sensors that you can adjust to accommodate for the unevenness of the LEDs.
RE: filmscanners: LED Illumination for Film Scanners
At 05:01 PM 6/20/01 -0400, Austin wrote: >Yes, and incandescent light bulbs used in an enlarger get a condenser >system, which is entirely different than a cold light head, which is >basically what most scanners use for illumination. Ah, but the "dynamics" of a conventional enlarger are rather different from a film scanner, no? In an enlarger, exposures take seconds, or tens of seconds. Fluctuations in the lamp intensity will average out, but they cannot cause "banding" in the print. Spatial non-uniformities (in an enlarger) will be dealt with by the diffuser or condenser. Any non-uniformities (at the print) can be attributed to a poor diffuser or condenser design. In a film scanner, fluctuations in intensity, or spatial non-uniformities that vary in time (during the course of the scan) will cause banding. It's not a hypothetical situation -- I've seen this effect in both of my previous film scanners, which used fluorescent light sources. This is a rare and anomalous situation (at least in the better scanners) but it does occur. I wonder if the Leaf benefits from that ridiculously huge bulb -- as opposed to the dinky little 4 inch F4T5 tube used in the older SprintScans and Microtek machines. rafe b.
Re: filmscanners: LED Illumination for Film Scanners
> > And taking things one step > > further, a dense LED array positioned closer to the negative could even be > > programmed to provide some degree of selective dodging/burning/variable > > constrast control. > > I doubt it. The control isn't that fine since the array is so course. Also > it would make the unevenness worse. Notice my words, I said to "some" degree. This isn't much different than a contact printing box made up of a dozen or so incandescent lamps that can be individually switched on or off. In which case you are purposely seeking uneven illumination. > > With an appropriate control mechanism, a user could > > adjust for dead even lighting across the easel for a specific lens/format > > size combination. > > Not with that array shown, and not without a VERY diffuse diffuser. Making > an LED array as shown is not a high level of sophistication. If it truly > worked well, I would believe that it would be being used by the enlarger > manufacturers, or at least offered in the aftermarket. I don't believe > either is true, or does anyone know differently? I said "dense LED array". The degree of sophistication in that enlarger light source was with the control electronics, and high-output LED's with very tight spectral parameters. It provided a very precisely controlled, low-heat light source that seemed ideal for enlargers. The downside, and probably one of the reasons there was no commercial development, is that the high-output LED's were very expensive at the time, and given the shrinking high-end enlarger market, probably not too much hope of commercial success. > The use of an LED array as a light source for a film enlarger is entirely > different than using an LED array for a scanner, where you have individual > sensors that you can adjust to accommodate for the unevenness of the LEDs. How have you determined that the LED's are uneven?
RE: filmscanners: LED Illumination for Film Scanners
> This isn't much different than a > contact printing box made up of a dozen or so incandescent lamps > that can be > individually switched on or off. In which case you are purposely seeking > uneven illumination. It's entirely different. Incandescent lamps used for such are diffuse and are not near as focused as LEDs. Using commercially available standard parts, you need individual LEDs because you need three colors. You can only get the array just so dense, and dense isn't necessarily good. If you have too much light, it's too difficult to control. You don't want to get the array too close to the film, because that would exaggerate the gaps between the LEDs. > I said "dense LED array". The degree of sophistication in that enlarger > light source was with the control electronics, and high-output LED's with > very tight spectral parameters. It provided a very precisely controlled, > low-heat light source that seemed ideal for enlargers. It's a nice idea, but it fails in the implementation. It just doesn't work as well as you may believe it could work. > How have you determined that the LED's are uneven? It's a fact that they are uneven. They have gaps between them, since they are individual lights, and their housed in plastic that is not very even optically. Why don't you just look at the images that were provided with the light source we've been discussing. It is blatantly obvious that they provide uneven illumination. Of course, you can diffuse them, and you can get them to be more even. The problem with that is that you lose some of the supposed "control" you are touting that they can have, since you are increasing/overlapping the area each LED covers. Your tradeoff is evenness of illumination vs control. LEDs have been around for a very long time, and they are reasonably inexpensive, as well as very easy to control. I am sure that if this was such a great idea, and the implementation worked near as well as you believe, it would have been done some 15-20 years ago as a commercial venture, but, alas, it wasn't.
RE: filmscanners: LED Illumination for Film Scanners
> Ah, but the "dynamics" of a conventional enlarger > are rather different from a film scanner, no? It depends on what dynamics you are talking about. Yes, as you say the exposure time is less, but that's about it. > I wonder if the Leaf benefits from that ridiculously > huge bulb -- as opposed to the dinky little 4 inch > F4T5 tube used in the older SprintScans and Microtek > machines. Being that it scans 4x5, it needs to use the "cleanest" area of a bulb at least 4+ inches wide. The ends of bulbs tend not to be near as uniform as the middle. I would agree it certainly is a lot larger than it needs to be as far as illumination area goes! Perhaps it has to do with controllability...that it is just easier to control a larger tube than a smaller one, or perhaps it was the best bulb available at the time. I know it's a custom bulb, made from a stock bulb...it has an area of the bulb unfrosted. I'll ask next time I talk with one of the original Leaf folks. It also doesn't really matter that the bulb is long or short, mechanically that is, it certainly isn't the determining factor in the size of the unit. BTW, I figured out a way to get 5080 PPI scans from it in MF mode...whether they are any good or not, I don't know, but it'll be interesting to try it out! It's only true 5080 in one direction (direction of stage travel), the other (CCD width) is just a simple avg across two adjacent pixels to double the data...
RE: filmscanners: LED Illumination for Film Scanners
> Has anyone actually seen the illumination system used in a Nikon > scanner? How large are the LED "elements"? Are all four "colors" > integrated into one LED element (R,G,B, IR), or are they individual? > Are they diffused via some material or light chamber? Sounds like we need a volunteer! Rafe? Got your screwdriver handy ;-)
Re: filmscanners: LED Illumination for Film Scanners
Obviously, how evenly illuminated a field of light sources appears to be has to do with at what distance (or magnification) they are being viewed at, and through what media the light is being shown through. There are all sorts of methods of mixing and baffling and diffusing that can blend illumination. One could say a color CRT isn't evenly illuminated either, as it is bunch of lines or spots of varying phosphors and a shadow mask through which the electron guns activate the phosphors. Yet, from a few feet away a good quality and properly aligned, converged and degaussed color CRT can look like a very pure field of white (or whatever other color) light. So, without knowing how large the LEDs are, how bright they are, how close together they can be placed, how even their output, how closely they will be viewed and ultimately what their light is "processed through" I have no reason to believe they cannot produce even illumination for a specific application. Has anyone actually seen the illumination system used in a Nikon scanner? How large are the LED "elements"? Are all four "colors" integrated into one LED element (R,G,B, IR), or are they individual? Are they diffused via some material or light chamber? Art Austin Franklin wrote: >>> What I don't yet understand is how the illuminant >>> is evenly distributed over the film width, >> >> Lots of LEDS, spaced to give even illumination. > > > But that's the point...you CAN'T space them to give even illumination. Just > a single LED is unevenly illuminated in and of it self! It's typically a > mounded plastic piece, which is really not very consistent.
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: > > > Austin Franklin wrote: > > > > > > > A couple of years ago someone on the darkroom newsgroup was > > working on an > > > > LED light source for enlarger heads, utilizing clusters of > > high-intensity > > > > LED's. I don't know what happened to the project, but at the > > time a lot of > > > > people were really excited about the technology and the > > initial results > > > > showed a lot of promise. > > > > > > > > http://www.trailing-edge.com/www/led.html > > > > > > Yes, and look at how uneven the lighting is. For an enlarger, > > that may turn > > > out to be a very bad idea, simply because you can't adjust each > > individual > > > LED for even illumination. At least with a CCD, you can adjust > > the gain for > > > each sensor element to get even illumination. > > > > All that you would need to do is shoot this array through > > some sort of > > diffusing device. > > And that was the question, was there a diffuser of some sort, and that has > gone unanswered. > > > After all, light bulbs aren't exactly evenly > > illuminated and those have been used for years in enlargers... > > Yes, and incandescent light bulbs used in an enlarger get a condenser > system, which is entirely different than a cold light head, which is > basically what most scanners use for illumination. Well halogen bulbs have the same uneveness problems as regular bulbs, but with the use of a mixing chamber, they can give the same results a "cold" head can... I can see some potential advantages in an enlarger with LEDs in this arangement, primarly concerning heat. I'm not sure what sort of advantage LEDs would have in a scanner, maybe long life with minimal shift (light wise)? Isaac
RE: filmscanners: LED Illumination for Film Scanners
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Austin Franklin Sent: Wednesday, June 20, 2001 9:00 PM To: [EMAIL PROTECTED] Subject: RE: filmscanners: LED Illumination for Film Scanners > LEDs have been around for a very long time, and they are reasonably > inexpensive, as well as very easy to control. I am sure that if this was > such a great idea, and the implementation worked near as well as you > believe, it would have been done some 15-20 years ago as a commercial > venture, but, alas, it wasn't. An LED light source for enlargers was not done 15-20 years ago because it was not possible. Blue LED's did not exist as anything other than laboratory curiosities until within the last 5 years. High volume commercial production of blue LED's has come about within the last 3 years, and high-brightness versions have only been available for a about a year. Here's some history: http://www.sciam.com/2000/0800issue/0800profile.html http://ledmuseum.home.att.net/1990.htm LED's are available from many manufacturers with many forms of diffusion and light distribution patterns. With the right combination of molded LED lenses, diffusion materials, and possibly optics a very good enlarger source could be constructed. With the current decline of chemical imaging such a source would possibly not be commercially viable, but it is technically feasible. As far as the uneven character of the source shown at http://www.trailing-edge.com/www/led.html , it's no worse than the three lamp sources found in RGB color heads in some additive enlargers. The individual bulbs used there are spaced a couple of inches apart and yet their light is mixed to provide an even, color controlled source. The light source shown on the above site is obviously a prototype; it uses what appear to be narrow beam clear lensed LED's. If diffuse lensed wide angle LED's were used along with a secondary diffuser or lens, you'd see much more even illumination. There's no reason that desire and a little innovative engineering could not produce an acceptable source. Economics might be a tougher nut to crack though. Cliff Ober
RE: filmscanners: LED Illumination for Film Scanners
> > LEDs have been around for a very long time, and they are reasonably > > inexpensive, as well as very easy to control. I am sure that > if this was > > such a great idea, and the implementation worked near as well as you > > believe, it would have been done some 15-20 years ago as a commercial > > venture, but, alas, it wasn't. > > > An LED light source for enlargers was not done 15-20 years ago because it > was not possible. Blue LED's did not exist as anything other than > laboratory > curiosities until within the last 5 years. Sure you could have done that 15-20 years ago. Use filters...red, green and blue filters certainly were around 15-20 years ago. > With the current decline of chemical imaging such a > source would possibly not be commercially viable, but it is technically > feasible. Possibly, but it really depends on what you mean by technically feasible. What are your design goals? The discussion was a controllable LED light source, and I still contend that is not really as feasible as "the discussion" was claiming it was. Of course you can make an LED light source that will evenly illuminate, but that was not the design goal being discussed. They are also harder TO control evenness, since there are so many different lamps...and they do have tolerances, and unlike the scanner, you can't adjust the input values of the individual sensors to make up for this unevenness. That is, unless you wanted to go through some exhaustive calibration process. I don't know how much these drift, but they may have the need to be re-calibrated in the field. Also, remember scanners are only one dimensional, an enlarger is two dimensional. > As far as the uneven character of the source shown at > http://www.trailing-edge.com/www/led.html , it's no worse than the three > lamp sources found in RGB color heads in some additive enlargers. The > individual bulbs used there are spaced a couple of inches apart and yet > their light is mixed to provide an even, color controlled source. But the "tout" for using LEDs is that they can be individually controlled, which they, of course, can be, but the problem is the more you diffuse them, the less area control you have.
RE: filmscanners: LED Illumination for Film Scanners
> > An LED light source for enlargers was not done 15-20 years ago > because it > > was not possible. Blue LED's did not exist as anything other than > > laboratory > > curiosities until within the last 5 years. > > Sure you could have done that 15-20 years ago. Use > filters...red, green and > blue filters certainly were around 15-20 years ago. >From what I can tell from my old LED catalogs, I am mistaken here. I though that there were clear LEDs "back then", which there were...but the emitting color is either red or green, and obviously, that won't work. Even if this was such a good idea, it would have been done 5 years ago, and, as far as I can tell, it wasn't. It isn't a difficult engineering project to develop, it's just that, I believe, it really doesn't work all that well. Digital imaging has really only come of age in the past two to three years, and there certainly was a window of opportunity (and I believe still would be) for an LED enlarging light source, if it really worked well.
RE: filmscanners: LED Illumination for Film Scanners
> http://ledmuseum.home.att.net/1990.htm BTW, thanks for that link! I did find it most interesting, and certainly a great source for information on LEDs. I would never have imagined that someone would devote so much time and thought to LEDs. As a side note, when looking through my LED file, I found some brochures that showed there were blue filters for LEDs back in at least the mid 80's. Now how well they would work, I don't know, but they were apparently available.
Re: filmscanners: LED Illumination for Film Scanners
> It's entirely different. Incandescent lamps used for such are diffuse and > are not near as focused as LEDs. Using commercially available standard > parts, you need individual LEDs because you need three colors. You can only > get the array just so dense, and dense isn't necessarily good. If you have > too much light, it's too difficult to control. You don't want to get the > array too close to the film, because that would exaggerate the gaps between > the LEDs. All of the parameters of concern can be specifically tailored within the manufacturing of the LED's, which may or may not be commercially available standard parts. If you want a high density array, just specify parts that will provide the illumination level needed. > It's a nice idea, but it fails in the implementation. It just doesn't work > as well as you may believe it could work. How do you know how well it could work? Personally, I can't say for sure because I've never tried it. But I recall the reports from people who tried the various prototypes to be extremely positive. > It's a fact that they are uneven. They have gaps between them, since they > are individual lights, and their housed in plastic that is not very even > optically. Why don't you just look at the images that were provided with > the light source we've been discussing. It is blatantly obvious that they > provide uneven illumination. Whether the light array has gaps or not is of little importance. What matters is how even the illumination is at the target. You cannot tell how even the illumination is by looking at the light source itself. You need to measure it at the target. Whether you're using two, ten, or a hundred LED's doesn't make a difference, as long as the target is evenly illuminated. > Of course, you can diffuse them, and you can get them to be more even. The > problem with that is that you lose some of the supposed "control" you are > touting that they can have, since you are increasing/overlapping the area > each LED covers. Your tradeoff is evenness of illumination vs control. Diffusion, if used, can be built into the LED lens, or it can be provided by supplementary material. I'd assume that in an LED array there would be some amount of overlap involved to provide even illumination.
RE: filmscanners: LED Illumination for Film Scanners
At 11:05 PM 6/20/01 -0400, someone wrote: > >> Has anyone actually seen the illumination system used in a Nikon >> scanner? How large are the LED "elements"? Are all four "colors" >> integrated into one LED element (R,G,B, IR), or are they individual? >> Are they diffused via some material or light chamber? [Austin:] >Sounds like we need a volunteer! Rafe? Got your screwdriver handy ;-) Er, how about someone with an old, out-of-warranty LS-30 or LS-2000? rafe b.
Re: filmscanners: LED Illumination for Film Scanners
At 11:18 PM 6/20/01 -0400, Isaac Crawford wrote: >Well halogen bulbs have the same uneveness problems as regular bulbs, >but with the use of a mixing chamber, they can give the same results a >"cold" head can... I can see some potential advantages in an enlarger >with LEDs in this arangement, primarly concerning heat. I'm not sure >what sort of advantage LEDs would have in a scanner, maybe long life >with minimal shift (light wise)? The *problem* I see with cold-cathode and fluorescents is that they can flicker. I'm not exactly sure why this happens. Is it a power-supply problem, or is it a property of the bulb, or a combination of the two? I do know that fluorescents are based on plasma physics -- essentially, a low-pressure mercury vapor is made to conduct current and emit photons as a result. One can easily imagine local variations in the plasma (and very dynamic ones, at that!) which cause non-uniformities in the light output. In fact, you can do more than imagine it... most of us have seen it occur. Plasmas are incredibly beautiful and complex phenomena. There is no comparable issue with tungsten lighting or LED lighting that I'm aware of -- in these cases, dynamic variations are nearly always attributable to the power supply. With tungsten lighting you have thermal time constants that slow down the dynamics quite a bit. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
On Thu, 21 Jun 2001 00:59:33 -0400 Austin Franklin ([EMAIL PROTECTED]) wrote: > I would never have imagined that > someone would devote so much time and thought to LEDs. I know I can't :-) Regards Tony Sleep http://www.halftone.co.uk - Online portfolio & exhibit; + film scanner info & comparisons
RE: filmscanners: LED Illumination for Film Scanners
On Wed, 20 Jun 2001 13:18:04 -0400 Austin Franklin ([EMAIL PROTECTED]) wrote: > But that's the point...you CAN'T space them to give even illumination. > Just > a single LED is unevenly illuminated in and of it self! It's typically > a > mounded plastic piece, which is really not very consistent. Nikon are unlikely to be using mass-produced general-purpose .5c LED's on a bit of breadboard . I've not seen the strip array used, but ISTM small precision-LED's with suitable front lenses will present a virtual point source. Careful spacing is then all that is required. Or, for a bit more ease of production, interpose a diffuser and/or move the LED's a bit further away... whatever, it works fine. Regards Tony Sleep http://www.halftone.co.uk - Online portfolio & exhibit; + film scanner info & comparisons
RE: filmscanners: LED Illumination for Film Scanners
At 09:59 PM 6/20/01 -0400, Austin wrote: >LEDs have been around for a very long time, and they are reasonably >inexpensive, as well as very easy to control. I am sure that if this was >such a great idea, and the implementation worked near as well as you >believe, it would have been done some 15-20 years ago as a commercial >venture, but, alas, it wasn't. By the same logic, Austin -- Nikon has been using LEDs as illuminants in scanners for several years now, yet you seem convinced that there's something fundamentally wrong with this strategy. While I've heard many complaints about Nikon scanners on this list and others, over the years -- banding has not been one of them. You're right, LEDs have been around for, what -- 30 years or so now. The variations in packaging and encapsulation are amazing -- there are clear lens-like packages, and translucent diffuser-like packages. There are monolithic arrays of all shapes and sizes, also. There are board-mount styles and surface-mount. I can easily imagine packages and monolithich LED arrays that allow for very, very close stacking of LEDs, so that you could have, say, 20 or 50 or even 100 LEDs per inch. Use your imagination, Austin. Is it possible that some engineer at Kodak or Nikon has thought of a trick with LEDs that hasn't occurred to us yet?? Though I do share your curiosity on just how that trick works... maybe a patent search could help here. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
On Wed, 20 Jun 2001 21:59:33 -0400 Austin Franklin ([EMAIL PROTECTED]) wrote: > It's entirely different. Incandescent lamps used for such are diffuse > and > are not near as focused as LEDs. Using commercially available standard > parts, you need individual LEDs because you need three colors. Bah, small minded Luddites abound ;-) Of course the real advance will be to use high-intensity TFT panels as a scanner lightsource, strobing RGB sequentially and self-masking in a closed-loop with the CCD. A calibration prescan will allow user and/or automated control of curves, to attenuate or boost TFT pixel brightness on an individual basis, thereby ensuring maximum shadow detail (all kept above the noise floor), unblowable highlights, balancing of bright areas like skies at scan time, and an OD range of about 12.6 A shame TFT doesn't have room for IR, but I'm sure that can be fixed. Interested scanner manufacturers should forward large bundles of used $50 notes to me ASAP Regards Tony Sleep http://www.halftone.co.uk - Online portfolio & exhibit; + film scanner info & comparisons
RE: filmscanners: LED Illumination for Film Scanners
> The *problem* I see with cold-cathode and fluorescents is that > they can flicker. I'm not exactly sure why this happens. Typically it is caused by the observer being on some psycho conducive substance ;-) The Leaf uses a "tri-band phosphor fluorescent lamp", which I would guess doesn't have the fluctuation problem you mention, or if it does, it does not seem to matter.
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: > > > http://ledmuseum.home.att.net/1990.htm > > BTW, thanks for that link! I did find it most interesting, and certainly a > great source for information on LEDs. I would never have imagined that > someone would devote so much time and thought to LEDs. > > As a side note, when looking through my LED file, I found some brochures > that showed there were blue filters for LEDs back in at least the mid 80's. > Now how well they would work, I don't know, but they were apparently > available. I don't know if this is related or not, but check out the Chromira printer by ZBE http://www.zbe.com/... It uses LEDs to produce 425 ppi prints onto RA4 materials. The prints are awesome to say the least... Isaac
Re: filmscanners: LED Illumination for Film Scanners
> I can easily imagine packages and monolithich LED arrays > that allow for very, very close stacking of LEDs, so that > you could have, say, 20 or 50 or even 100 LEDs per inch. For years Olympus has been making "laser" printers that claim to use LED's instead of a laser/mirror assembly. I always wondered how they managed to pack 300 or 600 LED's into the space of an inchso I too am curious about this. -- Todd Radel - [EMAIL PROTECTED] SCHWAG.ORG - Where Freaks and Geeks Come Together http://www.schwag.org/ PGP key available at http://www.schwag.org/~thr/pgpkey.txt
Re: filmscanners: LED Illumination for Film Scanners
At 08:00 AM 6/21/2001 -0400, rafe b wrote: >At 11:18 PM 6/20/01 -0400, Isaac Crawford wrote: > > >Well halogen bulbs have the same uneveness problems as regular bulbs, > >but with the use of a mixing chamber, they can give the same results a > >"cold" head can... I can see some potential advantages in an enlarger > >with LEDs in this arangement, primarly concerning heat. I'm not sure > >what sort of advantage LEDs would have in a scanner, maybe long life > >with minimal shift (light wise)? > > >The *problem* I see with cold-cathode and fluorescents is that >they can flicker. I'm not exactly sure why this happens. Is >it a power-supply problem, or is it a property of the bulb, or >a combination of the two? > >I do know that fluorescents are based on plasma physics -- >essentially, a low-pressure mercury vapor is made to conduct >current and emit photons as a result. One can easily imagine >local variations in the plasma (and very dynamic ones, at >that!) which cause non-uniformities in the light output. > >In fact, you can do more than imagine it... most of us have >seen it occur. Plasmas are incredibly beautiful and complex >phenomena. > >There is no comparable issue with tungsten lighting or LED >lighting that I'm aware of -- in these cases, dynamic variations >are nearly always attributable to the power supply. With >tungsten lighting you have thermal time constants that slow >down the dynamics quite a bit. Fluorescents flicker at the AC line frequency--60 Hz in the US. This is because, as you say, the fluorescent light is a plasma device. The discharge turns on and off at the line frequency. It is not a continuous discharge (either in time or in wavelength). The UV from the mercury vapor discharge tube causes the inner coating of fluorescent material to, er, fluoresce. The composition of the coating determines what wavelengths (I would normally use "frequency" but I don't want to confuse it with "AC line frequency") will be emitted by the tube. Stan === Photography by Stan McQueen: http://www.smcqueen.com
RE: filmscanners: LED Illumination for Film Scanners
At 12:26 AM 6/21/01 -0400, Austin wrote: >Even if this was such a good idea, it would have been done 5 years ago, and, >as far as I can tell, it wasn't. It isn't a difficult engineering project >to develop, it's just that, I believe, it really doesn't work all that well. >Digital imaging has really only come of age in the past two to three years... You mean, well after the LeafScans were introduced? rafe b.
RE: filmscanners: LED Illumination for Film Scanners
> > But that's the point...you CAN'T space them to give even illumination. > > Just > > a single LED is unevenly illuminated in and of it self! It's typically > > a > > mounded plastic piece, which is really not very consistent. > > Nikon are unlikely to be using mass-produced general-purpose .5c LED's on > a bit of breadboard . Of course, but that comment was aimed at using them for an enlarger light source, not the Nikon.
Re: filmscanners: LED Illumination for Film Scanners
on 6/21/01 5:03 PM, Austin Franklin at [EMAIL PROTECTED] wrote: > No, not for a scanner, for an enlarger! I don't know whether they are a > good or bad idea for a scanner, but I do know that for B&W, I would rather > be using a ND filter, or something with that wavelength. ??? an ND filter doesn't filter by wavelength, it simply reduces intensity. Of all wavelengths to (ideally) the same degree. So whatever the spectrum of the source is, the ND filter leaves it reduced in amplitude but unchanged in frequency. red light stays red etc -- John Brownlow http://www.pinkheadedbug.com
RE: filmscanners: LED Illumination for Film Scanners
> >LEDs have been around for a very long time, and they are reasonably > >inexpensive, as well as very easy to control. I am sure that if this was > >such a great idea, and the implementation worked near as well as you > >believe, it would have been done some 15-20 years ago as a commercial > >venture, but, alas, it wasn't. > > > By the same logic, Austin -- Nikon has been using LEDs > as illuminants in scanners for several years now, yet > you seem convinced that there's something fundamentally > wrong with this strategy. No, not for a scanner, for an enlarger! I don't know whether they are a good or bad idea for a scanner, but I do know that for B&W, I would rather be using a ND filter, or something with that wavelength. I haven't really done the research yet, but I have some suspicion about using RGB for B&W, that it isn't up to snuff. > Use your imagination, Austin. Is it possible that > some engineer at Kodak or Nikon has thought of a > trick with LEDs that hasn't occurred to us yet?? > Though I do share your curiosity on just how that > trick works... maybe a patent search could help here. I don't have a problem with a 1d array of LEDs for a scanner, since the scanner CAN adjust for uneven illumination, but an enlarger can't. That's what I've been talking about...though, as you, I would like to know more about the use in a scanner.
RE: filmscanners: LED Illumination for Film Scanners
> >Even if this was such a good idea, it would have been done 5 > years ago, and, > >as far as I can tell, it wasn't. It isn't a difficult > engineering project > >to develop, it's just that, I believe, it really doesn't work > all that well. > >Digital imaging has really only come of age in the past two to > three years... > > > You mean, well after the LeafScans were introduced? Not quite what I meant, more for the masses is what I meant.
Re: filmscanners: LED Illumination for Film Scanners
On Thu, 21 Jun 2001, Stan McQueen wrote: > Fluorescents flicker at the AC line frequency--60 Hz in the US. This is > because, as you say, the fluorescent light is a plasma device. The > discharge turns on and off at the line frequency. It is not a continuous > discharge (either in time or in wavelength). The UV from the mercury vapor > discharge tube causes the inner coating of fluorescent material to, er, > fluoresce. The composition of the coating determines what wavelengths (I > would normally use "frequency" but I don't want to confuse it with "AC line > frequency") will be emitted by the tube. Stan, I was thinking of more pathological behavior. A healthy bulb, with a healthy, regulated power supply, is not really my concern here. I understand that household flourescents are driven by 60 Hz, (in the USA) but I also know that the bulbs inside some film scanners are driven at much higher frequencies, and those frequencies are not well controlled. But I have seen, on household flourescent lamps, situations where the plasma seems to be moving about and varying in intensity. It's generally when either the bulb or its "ballast" is near the end of its life. And I've seen banding effects in my older scanners (both the Microtek and the Polaroid) which could only be explained by time-variant spatial non-uniformity of the lamp's brightness. A moderate spatial non-uniformity would be acceptable, IF it were time-invariant, at least during the course of one scan. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
> > No, not for a scanner, for an enlarger! I don't know whether they are a > > good or bad idea for a scanner, but I do know that for B&W, I > would rather > > be using a ND filter, or something with that wavelength. > > ??? > > an ND filter doesn't filter by wavelength, it simply reduces intensity. > > Of all wavelengths to (ideally) the same degree. > > So whatever the spectrum of the source is, the ND filter leaves it reduced > in amplitude but unchanged in frequency. > > red light stays red etc Er, right. I understand how an ND filter works. I guess I didn't quite say that right. But to the actual point, I'd like to understand if you get a "better" B&W scan by using an ND filter, vs using one or more of the RGB channels and converting to B&W.
RE: filmscanners: LED Illumination for Film Scanners
> >> Sure you could have done that 15-20 years ago. > >> Use filters...red, green and blue filters > >> certainly were around 15-20 years ago. > > HUH??? Please explain to me how to filter monochromatic red light from an > LED to get blue light. If you can figure out how do that you'll > be up for a > Nobel prize in physics. I did post a correction to this yesterday, saying I was thinking of clear LEDS, but that is not the emitter color, so it wouldn't work as I expected...BUT...I have brochures from back in the mid 80's that offer filters for LEDs...in varying colors, and one of the colors IS blue.
Re: filmscanners: LED Illumination for Film Scanners
If I'm not mistaken, all A.C powered fluorescent bulbs "flicker" at 60/50 cycle (in those places that use 60 or 50 cycle AC, that is). In fact so do incandescent bulbs, for that matter, but in both cases, aspects of the design reduce the visibility of this. I think television screens also do so at some frequency..., and we know that computer monitors do, and the higher the frequency the less obvious, at least on a conscious level. In the case of fluorescents (and televisions and CRTs), the phosphors have a long enough lag activation period to "bridge" this rather short fraction of a second. In the case of incandescents, the filament remains hot enough to bridge that difference in electrical flow. I plead total ignorance in how DC powered bulbs fare, however. Regarding the Leaf scanners. I knew they were off my want list when I saw the bulb for one being sold on ebay, as a separate auction item ;-) Art Austin Franklin wrote: >> The *problem* I see with cold-cathode and fluorescents is that >> they can flicker. I'm not exactly sure why this happens. > > > Typically it is caused by the observer being on some psycho conducive > substance ;-) > > The Leaf uses a "tri-band phosphor fluorescent lamp", which I would guess > doesn't have the fluctuation problem you mention, or if it does, it does not > seem to matter.
Re: filmscanners: LED Illumination for Film Scanners
Tony Sleep wrote: > On Wed, 20 Jun 2001 21:59:33 -0400 Austin Franklin > ([EMAIL PROTECTED]) wrote: > > >> It's entirely different. Incandescent lamps used for such are diffuse >> and >> are not near as focused as LEDs. Using commercially available standard >> parts, you need individual LEDs because you need three colors. > > > Bah, small minded Luddites abound ;-) Of course the real advance will be > to use high-intensity TFT panels as a scanner lightsource, strobing RGB > sequentially and self-masking in a closed-loop with the CCD. A > calibration prescan will allow user and/or automated control of curves, to > attenuate or boost TFT pixel brightness on an individual basis, thereby > ensuring maximum shadow detail (all kept above the noise floor), > unblowable highlights, balancing of bright areas like skies at scan time, > and an OD range of about 12.6 Gee, 'big' minds seem to think alike. I was actually looking into this myself. I think the problem is that the TFT display doesn't provide the light source, just the color, so you are still having to provide a good constant color light source (cold cathode, I suppose). Also, those LCD panels do have problems with color accuracy (which is why no one does color managed work with a LCD panel) and also there are still failed "pixels", something one would like to avoid in mission critical applications like scanning. By the way, some of the index prints that are made by photo labs which come with APS rolls, for instance, use a negative digital mask made from a color LCD screen contact printed onto the photo paper. Resolution is not great, but its an ingenious idea, IMHO. > > A shame TFT doesn't have room for IR, but I'm sure that can be fixed. I suspect this could be accomplished through using two different light sources. > > Interested scanner manufacturers should forward large bundles of used $50 > notes to me ASAP Now, don't tell me you've given up on Euros already! Art
RE: filmscanners: LED Illumination for Film Scanners
Austin, That message was somehow delayed in the system for many hours... What company made those filters? - I'd like to look up the details (I have an extensive electronics data library going back about thirty years that includes a very large amount of optoelectronics). Cliff Ober -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Austin Franklin Sent: Thursday, June 21, 2001 5:25 PM To: [EMAIL PROTECTED] Subject: RE: filmscanners: LED Illumination for Film Scanners > >> Sure you could have done that 15-20 years ago. > >> Use filters...red, green and blue filters > >> certainly were around 15-20 years ago. > > HUH??? Please explain to me how to filter monochromatic red light from an > LED to get blue light. If you can figure out how do that you'll > be up for a > Nobel prize in physics. I did post a correction to this yesterday, saying I was thinking of clear LEDS, but that is not the emitter color, so it wouldn't work as I expected...BUT...I have brochures from back in the mid 80's that offer filters for LEDs...in varying colors, and one of the colors IS blue.
RE: filmscanners: LED Illumination for Film Scanners
Yeah, me too, that's why I have that info. The one I have in front of me is VCC, and it's called the "ClipLite" and the "CubeLite". Red, amber, green, blue, yellow and clear. For 3mm and 5mm LEDs. > > What company made those filters? - I'd like to look up the details (I have > an extensive electronics data library going back about thirty years that > includes a very large amount of optoelectronics). > > Cliff Ober > > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED]]On Behalf Of Austin Franklin > Sent: Thursday, June 21, 2001 5:25 PM > To: [EMAIL PROTECTED] > Subject: RE: filmscanners: LED Illumination for Film Scanners > > > > > >> Sure you could have done that 15-20 years ago. > > >> Use filters...red, green and blue filters > > >> certainly were around 15-20 years ago. > > > > HUH??? Please explain to me how to filter monochromatic red > light from an > > LED to get blue light. If you can figure out how do that you'll > > be up for a > > Nobel prize in physics. > > I did post a correction to this yesterday, saying I was thinking of clear > LEDS, but that is not the emitter color, so it wouldn't work as I > expected...BUT...I have brochures from back in the mid 80's that offer > filters for LEDs...in varying colors, and one of the colors IS blue. > >
RE: filmscanners: LED Illumination for Film Scanners
> Regarding the Leaf scanners. I knew they were off my want list when I > saw the bulb for one being sold on ebay, as a separate auction item ;-) At least you CAN get them with no problem. Check how much the LED array is for the Nikon, and you may reconsider!
Re: filmscanners: LED Illumination for Film Scanners
At 05:39 PM 6/21/2001 -0400, you wrote: >On Thu, 21 Jun 2001, Stan McQueen wrote: > > > Fluorescents flicker at the AC line frequency--60 Hz in the US. This is > > because, as you say, the fluorescent light is a plasma device. The > > discharge turns on and off at the line frequency. It is not a continuous > > discharge (either in time or in wavelength). The UV from the mercury vapor > > discharge tube causes the inner coating of fluorescent material to, er, > > fluoresce. The composition of the coating determines what wavelengths (I > > would normally use "frequency" but I don't want to confuse it with "AC > line > > frequency") will be emitted by the tube. > > >Stan, I was thinking of more pathological behavior. > >A healthy bulb, with a healthy, regulated power >supply, is not really my concern here. I understand >that household flourescents are driven by 60 Hz, >(in the USA) but I also know that the bulbs inside some >film scanners are driven at much higher frequencies, >and those frequencies are not well controlled. > >But I have seen, on household flourescent lamps, >situations where the plasma seems to be moving >about and varying in intensity. It's generally >when either the bulb or its "ballast" is near >the end of its life. > >And I've seen banding effects in my older >scanners (both the Microtek and the Polaroid) >which could only be explained by time-variant >spatial non-uniformity of the lamp's brightness. > >A moderate spatial non-uniformity would be >acceptable, IF it were time-invariant, at least >during the course of one scan. > > >rafe b. === Photography by Stan McQueen: http://www.smcqueen.com
Re: filmscanners: LED Illumination for Film Scanners
At 05:39 PM 6/21/2001 -0400, rafe b wrote: >On Thu, 21 Jun 2001, Stan McQueen wrote: > > > Fluorescents flicker at the AC line frequency--60 Hz in the US. This is > > because, as you say, the fluorescent light is a plasma device. The > > discharge turns on and off at the line frequency. It is not a continuous > > discharge (either in time or in wavelength). The UV from the mercury vapor > > discharge tube causes the inner coating of fluorescent material to, er, > > fluoresce. The composition of the coating determines what wavelengths (I > > would normally use "frequency" but I don't want to confuse it with "AC > line > > frequency") will be emitted by the tube. > >Stan, I was thinking of more pathological behavior. > >A healthy bulb, with a healthy, regulated power >supply, is not really my concern here. I understand >that household flourescents are driven by 60 Hz, >(in the USA) but I also know that the bulbs inside some >film scanners are driven at much higher frequencies, >and those frequencies are not well controlled. > >But I have seen, on household flourescent lamps, >situations where the plasma seems to be moving >about and varying in intensity. It's generally >when either the bulb or its "ballast" is near >the end of its life. [Sorry about the earlier post, I accidentally hit "send" before I had typed in my response.] Yes, bulbs that visibly flicker are usually suffering from under-voltage or possibly vacuum leaks that cause the plasma discharge to be weak and intermittent. Stan === Photography by Stan McQueen: http://www.smcqueen.com
Re: filmscanners: LED Illumination for Film Scanners
At 03:36 PM 6/21/2001 -0700, you wrote: >If I'm not mistaken, all A.C powered fluorescent bulbs "flicker" at 60/50 >cycle (in those places that use 60 or 50 cycle AC, that is). In fact so >do incandescent bulbs, for that matter, but in both cases, aspects of the >design reduce the visibility of this. I think television screens also do >so at some frequency..., and we know that computer monitors do, and the >higher the frequency the less obvious, at least on a conscious level. > >In the case of fluorescents (and televisions and CRTs), the phosphors have >a long enough lag activation period to "bridge" this rather short fraction >of a second. In the case of incandescents, the filament remains hot >enough to bridge that difference in electrical flow. For most, perhaps all, bulbs, the phosphors do not actually have a long enough lag time to "bridge" the gap. Persistence of vision is the explanation for why it seems continuous. 50 or 60 Hz is far faster than the 24 frames-per-second that was (still is, for all I know) used for motion pictures. The thermal "inertia" of the incandescent bulb prevents it from flickering. That is, it can't possibly cool down enough to stop glowing as the voltage crosses the zero point. There just isn't enough time before the voltage climbs back up on the other side. (As you said.) Stan === Photography by Stan McQueen: http://www.smcqueen.com
Re: filmscanners: LED Illumination for Film Scanners
on 6/21/01 8:00 AM, rafeb at [EMAIL PROTECTED] wrote: > At 11:18 PM 6/20/01 -0400, Isaac Crawford wrote: > >> Well halogen bulbs have the same uneveness problems as regular bulbs, >> but with the use of a mixing chamber, they can give the same results a >> "cold" head can... I can see some potential advantages in an enlarger >> with LEDs in this arangement, primarly concerning heat. I'm not sure >> what sort of advantage LEDs would have in a scanner, maybe long life >> with minimal shift (light wise)? > > > The *problem* I see with cold-cathode and fluorescents is that > they can flicker. I'm not exactly sure why this happens. Is > it a power-supply problem, or is it a property of the bulb, or > a combination of the two? It is a property of the bulb. It is using 60 Hz pulses of AC to excite phosporescent matter, so one pulse may not excite as much as the previous or following pulse, and we see at 60 Hz... Jim Snyder
Re: filmscanners: LED Illumination for Film Scanners
> > Regarding the Leaf scanners. I knew they were off my want list when I > > saw the bulb for one being sold on ebay, as a separate auction item ;-) > > At least you CAN get them with no problem. Check how much the LED array is > for the Nikon, and you may reconsider! You've made your aversion to LED's well know. But you must know that they are not a consumable item, like the Leafscan bulb. It's inevitable that at some point you'll need to buy a bulb for your scanner. I don't know where you get them now, but in a few years the only place you'll be able to get them will probably be ebay. On the other hand, I've never heard of and LED array burning out. It's considered a permanent light source. Score 1 for Nikon, 0 for Leafscan.
RE: filmscanners: LED Illumination for Film Scanners
> > > Regarding the Leaf scanners. I knew they were off my want list when I > > > saw the bulb for one being sold on ebay, as a separate > auction item ;-) > > > > At least you CAN get them with no problem. Check how much the LED array > is > > for the Nikon, and you may reconsider! > > You've made your aversion to LED's well know. Not only LEDs, but scanning B&W in RGB. > But you must know that they > are not a consumable item, like the Leafscan bulb. It's > inevitable that at > some point you'll need to buy a bulb for your scanner. I don't know where > you get them now, but in a few years the only place you'll be able to get > them will probably be ebay. On the other hand, I've never heard of and LED > array burning out. It's considered a permanent light source. > > Score 1 for Nikon, 0 for Leafscan. No one, but you, appears to be keeping score. This wasn't a Leaf vs Nikon "discussion". Because YOU don't know where to get Leafscan bulbs, doesn't mean they are not readily available. In fact, they are available from CreoScitex, who owns Leaf, as well as any other parts for the Leafscan. It is still an entirely supported product. The bulbs are reasonably cheap, VERY easy to replace, and take a LONG time to burn out, depending on use. Of the over 100 Leafscan users I know, I know of no one who has replaced a bulb at least since I have known them, though people do typically keep a spare. Not to say someone may have, and I just didn't know about it...it's not like we all sit around and discuss bulb life. My guess is you are not an electrical engineer, or you would know that LEDs do have a life span. Because you haven't heard of them burning out, doesn't mean they don't burn out. In fact, their typical MTBF is rated for 1000 hours. Incandescent light bulbs are rated for 1000 hours. Aside from having written and reviewed quite a few MTBF and MTTR studies on designs that included LEDS, I recently replaced 4 of the 6 LEDs in my radio/CD/Cassette in my 1989 Range Rover, so I DO know they do burn out. Though I have the highest respect for the Nikon scanner, if you should have trouble with the LEDs, it would probably be an expensive and difficult replacement. Personally, that would not deter me at all from buying that scanner. What deters me is I already have a scanner that does exactly what I want it to. I will say I am disappointed that the resolution limit on the Nikon and the Polaroid is only 4000PPI for 35mm. I believe they don't change magnification for different formats. That gives a maximum print size for 35mm for the Nikon and the Polaroid, with 240PPI output to the print driver, of 16" x 24" , vs 21" x 31" for the Leaf. Nothing really wrong with that, but some 35mm films and lenses have enough resolution to take advantage of the higher resolution scans. The Leafscan has been around for more than 10 years, and still most of them are in use today. They are VERY trouble free, and they are built to be that way. They certainly have their shortcomings, but they also have their advantages, and easily hold their own with most any modern CCD scanner. They certainly are not for everyone.
RE: filmscanners: LED Illumination for Film Scanners
Hey let's keep this clean and vaguely accurate even if it is OT... >My guess is you are not an electrical engineer, or you would know that LEDs >do have a life span. Because you haven't heard of them burning out, doesn't >mean they don't burn out. In fact, their typical MTBF is rated for 1000 >hours. I am another engineer(!) (not that this is relevant to reading a manufacturer's spec) and LEDs don't have MTBFs of 1000 hours! One of the great advantages of using LEDs in a scanner is the enormous lifespan of the light source... this was also the original driver for the mooted LED enlarger lamp that you have been discussing - lifespan *and* the consistency of light i.e. unchanging spectral characteristics. In fact the MTBF of ordinary boring nothing special LEDs is around 100x your stated figure and good ones (presumably like those used in scanners) are 1000x. I quote from the first google-located site I found... "If packaged properly, LEDs emit light for a much longer time period than almost every other alternative light source technology. ... The mean time between failure (MTBF) of high quality LEDs properly packaged, is on the order of millions of hours. " Or this second site I found... "The long term dependability of Precision Optical Performance AlInGaP LED lamps is an important consideration for those who specify LED traffic signals and LED variable message signs (VMS). Precision Optical Performance LED Lamps are T-1 3 /4 plastic package devices that exhibit a nominal Mean Time Between (possible catastrophic) Failure, MTBF, greater than 1.2 million hours at the operating temperature of +74°C (+165°F). At operating temperatures below 0°C (32°F), MTBF is in excess of 10 million hours. Therefore, MTBF need not be a concern." Let's say the first LED dies in my scanner after 1/10 th of its MTBF, then I'll get 100,000 hours out of it or 50 years if I use the scanner 5 hours a day. Not bad eh! (Caveat - this was an example only - I don't know what the figure is for the actual LEDs used in Nikon scanners, but I am sure it is a lot higher than 1000hours). Julian Julian Robinson in usually sunny, smog free Canberra, Australia
RE: filmscanners: LED Illumination for Film Scanners
On Fri, 22 Jun 2001, Austin Franklin wrote: > > My guess is you are not an electrical engineer, or you would know that LEDs > do have a life span. Because you haven't heard of them burning out, doesn't > mean they don't burn out. In fact, their typical MTBF is rated for 1000 > hours. Incandescent light bulbs are rated for 1000 hours. Aside from > having written and reviewed quite a few MTBF and MTTR studies on designs > that included LEDS, I recently replaced 4 of the 6 LEDs in my > radio/CD/Cassette in my 1989 Range Rover, so I DO know they do burn out. 1000 hours MTBF can't be right, Austin. Optocouplers and fiberoptics are ubiquitous these days, and that figure would imply a rate of failure that's totally out of line with reality. Hell, the entire telecom industry is built on fiberoptics these days, and the transmitters are always LEDs of some sort. Virtually every single (modern) mechanical actuator these days is either stepper driven or DC-servo driven, and the latter type usually use quadrature-encoded OPTICAL position sensors. (Eg., any inkjet printer made by Epson, Lexmark, or HP uses optical encoding for the carriage position.) Just one example from among tens of thousands, and maybe a few hundred that I could cite. The "sender" in the optocoupler is a LED (two of 'em, for quadrature encoding.) I'd believe 100,000 hours MTBF, maybe, but clearly it'll be a function of several variables -- mostly the operating current. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
Austin, > Yeah, me too, that's why I have that info. The one I have in front of me is > VCC, and it's called the "ClipLite" and the "CubeLite". Red, amber, green, > blue, yellow and clear. For 3mm and 5mm LEDs. >> >> What company made those filters? - I'd like to look up the details (I have >> an extensive electronics data library going back about thirty years that >> includes a very large amount of optoelectronics). >> I'm familiar with VCC, having spec'd their products many times over the years. The early blue (and green) lenses were intended for use with small incandescent bulbs with the same T1-3/4 form-factor that was adopted by LED manufacturers, not for use with LEDs. Here's the type: http://www.abclights.com/midflant134.html > My guess is you are not an electrical engineer, or you would know that LEDs > do have a life span. Because you haven't heard of them burning out, doesn't > mean they don't burn out. In fact, their typical MTBF is rated for 1000 > hours. Incandescent light bulbs are rated for 1000 hours. Aside from > having written and reviewed quite a few MTBF and MTTR studies on designs > that included LEDS, I recently replaced 4 of the 6 LEDs in my > radio/CD/Cassette in my 1989 Range Rover, so I DO know they do burn out. Some of your recent statements of technical "fact" seem to be casting a bit of a shadow on your own credentials as an engineer; once again here are sites with valid data: http://ftp.agilent.com/pub/semiconductor/led_lamps/abi018.pdf http://www.uniroyalopto.com/aenmlife.html http://www.wch.com/led.htm http://www.safe.no/various/ledline.html MTBF of an LED is wholly dependent upon power dissipation, current (in pulse applications) and/or operating temperature. An MTBF figure of 100,000 hours is more the norm for LEDs operated within their specified limits, with many times that possible when the LED is run below rated power or pulse maximums. Cliff Ober
RE: filmscanners: LED Illumination for Film Scanners
> I am another engineer(!) (not that this is relevant to reading a > manufacturer's spec) and LEDs don't have MTBFs of 1000 hours! The one catalog I pulled off my shelf gave that figure. It seemed inordinately low to me, but it certainly was 1000. > the > consistency of light i.e. unchanging spectral characteristics. That has nothing to do with even illumination, which is different than consistent illumination. > In fact > the MTBF of ordinary boring nothing special LEDs is around 100x your > stated figure Whose did you find had that? Lumex stated 10,000 hours, which is 10x the original source I quoted. > "If packaged properly, LEDs emit light for a much longer time period than > almost every other alternative light source technology. ... The mean time > between failure (MTBF) of high quality LEDs properly packaged, is on the > order of millions of hours. " We don't know what LEDs the Nikon scanner uses, nor will we unless someone either takes their scanner apart, or buys a spare LED array. The original point was that they do have a life span, as anything does. The claim was that the LEDs in the Nikon are "permanent", and I do not believe they are. They may be for some people's use, but for others, it will be an item that they will probably have to replace over the usable life of the scanner.
Re: filmscanners: LED Illumination for Film Scanners
> My guess is you are not an electrical engineer, or you would know that LEDs > do have a life span. Because you haven't heard of them burning out, doesn't > mean they don't burn out. In fact, their typical MTBF is rated for 1000 > hours. Incandescent light bulbs are rated for 1000 hours. My guess is that you forgot a few zeros on the LED MTBF.
RE: filmscanners: LED Illumination for Film Scanners
> > My guess is you are not an electrical engineer, or you would > know that LEDs > > do have a life span. Because you haven't heard of them burning > out, doesn't > > mean they don't burn out. In fact, their typical MTBF is rated for 1000 > > hours. Incandescent light bulbs are rated for 1000 hours. Aside from > > having written and reviewed quite a few MTBF and MTTR studies on designs > > that included LEDS, I recently replaced 4 of the 6 LEDs in my > > radio/CD/Cassette in my 1989 Range Rover, so I DO know they do burn out. > > > 1000 hours MTBF can't be right, Austin. I'm only going by what the catalog says, and I didn't write it. Stanley LED catalog, p. 24: "Operating Life JIS C 7035 Ta=25C, IF=Max, t=1000Hrs." It very clearly says "1000". The Lumex web site says their Life Test is 10,000 hours, which sounds a lot better. Your scanner should have a cited MTBF and MTTR, what are they?
RE: filmscanners: LED Illumination for Film Scanners
On Fri, 22 Jun 2001, Julian Robinson wrote: > Hey let's keep this clean and vaguely accurate even if it is OT... Austin went just a bit over the edge with that 1000 hour MTBF figure. Having designed many circuits and systems around HP LED displays, optocouplers, fiberoptic transceivers, etc., I was troubled by that number. HP is now Agilent. Check out www.semiconductor.agilent.com and do a quick search using the search phrase "LED reliability" The first PDF I found cites a particular type of LED, used in highway warning signs, with a MTBF of 1.5 million hours (before catastrophic failure). For simple aging, the paper cites 100,000 hours of operation before a 25% reduction in light output. A second paper gives mildly contradictory data. This one (discussing LEDs for instrument cluster lighting) gives 10 million hours MTFB (catastrophic failure) and 50,000 hours of operation before a 75% reduction in output. Either way, I think most of us won't have to worry much about LEDs failing in our Nikon scanners. The scanners will be long obsolete before the LEDs die. 2000 hours is roughly a year's worth of work- days (weekdays) at 8 hours per day. I suppose if you figure in hard mechanical shock (like in Austin's Land Rover) the numbers might go down a bit. Time to fix the potholes in your driveway, Austin, or get new shock absorbers for that beast. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
LOL Lawrence > > I suppose if you figure in hard mechanical > shock (like in Austin's Land Rover) the > numbers might go down a bit. Time to fix > the potholes in your driveway, Austin, > or get new shock absorbers for that beast. > > > > rafe b. >
Re: filmscanners: LED Illumination for Film Scanners
Raphael Bustin wrote: > > On Fri, 22 Jun 2001, Julian Robinson wrote: > > >> Hey let's keep this clean and vaguely accurate even if it is OT... > > > > Austin went just a bit over the edge with that 1000 > hour MTBF figure. > Oh, what's a few orders of magnitude amongst engineers, anyway? ;-) > > I suppose if you figure in hard mechanical > shock (like in Austin's Land Rover) the > numbers might go down a bit. Time to fix > the potholes in your driveway, Austin, > or get new shock absorbers for that beast. > What type of display is used in things like VCR, tape deck and microwave displays? It looks like it is almost a type of gas plasma/fluorescent type of thing. Many of my older devices with those type of displays now have considerable and uneven loss of brightness. I sort of recall LEDs having pretty poor reliability many moons ago, when they were mainly seen in NASA spacecraft, those large wrist watches and Texas Instrument calculators, but I think 30 years has had it's effect upon the design, eh? Art
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: >>> My guess is you are not an electrical engineer, or you would >> >> know that LEDs >> >>> do have a life span. Because you haven't heard of them burning >> >> out, doesn't >> >>> mean they don't burn out. In fact, their typical MTBF is rated for 1000 >>> hours. Incandescent light bulbs are rated for 1000 hours. Aside from >>> having written and reviewed quite a few MTBF and MTTR studies on designs >>> that included LEDS, I recently replaced 4 of the 6 LEDs in my >>> radio/CD/Cassette in my 1989 Range Rover, so I DO know they do burn out. >> >> >> 1000 hours MTBF can't be right, Austin. Didn't the word humility crop up in a message I sent you about a year ago? ;-) Also, I'm guessing LEDs made in 1987-88, when your Range Rover was likely put together, were not the same ones used today, about 12 years later > I'm only going by what the catalog says, and I didn't write it. Stanley LED > catalog, p. 24: > I never met a Stanley parts catalogue I couldn't trust ;-) ;-) > "Operating Life JIS C 7035 Ta=25C, IF=Max, t=1000Hrs." Austin, this is why it sometimes is helpful to engage your obviously very capable brain rather than relying totally on written "fact". I'm sure given a little thought, you'd have recognized the silliness of the 1000 hr number, considering how often we replace incandescent household light bulbs. (then again, lately companies like GE seem to be happy with 100 hr bulbs or ones that just fail on power up, by that's another matter). Can you imagine what problems in equipment failures would exist if LEDs lasted on average 1000 hr only? I think Rafe gave some good examples of the huge number of devices we use which rely upon LEDs to sense locations, positioning, actuation or switching (how about most computer mouses, for instance). Tell you one thing, I'm not buying any Stanley LEDs! (they must have gotten a great buy on these!) ;-) > > It very clearly says "1000". The Lumex web site says their Life Test is > 10,000 hours, which sounds a lot better. > I also expect, like incandescents and other illumination sources, LEDs can be built to different specs and be run under different electronic designs. For instance, I know that with the halogen bulbs used in projectors, a "25 hr" bulb will last only 5-11 hours run at 125-130 volts, will last 11-19 hours run at 120-125 volts, they last upwards of 19-31 hrs run at 115-120 and if you use a dimmer to bring them down to about 105-110, they will last well over 50 hours, or nearly one order of magnitude from running at 125-130 volts. > Your scanner should have a cited MTBF and MTTR, what are they? Yes, that would be interesting, if they consider the light source separately, and not the whole unit. Art
RE: filmscanners: LED Illumination for Film Scanners
> >> > >> 1000 hours MTBF can't be right, Austin. > > Didn't the word humility crop up in a message I sent you about a year > ago? ;-) This issue isn't about humility at all. I was merely citing directly from a manufacturers data sheet! I did not make up the number, nor did I say I liked it, and I certainly had no plans on spending this inordinate amount of time on such a simply absurd discussion! I posted the URL of a scan of the data sheet to show that I didn't just make it up, as it appears some people are assuming. > Also, I'm guessing LEDs made in 1987-88, when your Range Rover was > likely put together, were not the same ones used today, about 12 > years later You could very well be right. > Austin, this is why it sometimes is helpful to engage your obviously > very capable brain rather than relying totally on written "fact". I'm > sure given a little thought, you'd have recognized the silliness > of the 1000 hr number, considering how often we replace incandescent > household light bulbs. Of course I thought it was quite low. I never said that I liked that number, it was the number the manufacturer stated in their specs. I used that number as a point of data showing that LEDs DO have MTBFs that CAN be quite low. But...since, as I've said before, this is all speculation, since no one here really knows what LEDs the Nikon scanner uses. > I think Rafe gave > some good examples of the huge number of devices we use which rely upon > LEDs to sense locations, positioning, actuation or switching (how about > most computer mouses, for instance). Again, that all has nothing to do with my point, as I said above. We all know that LEDs CAN have very high MTBFs, that is not, and never was, under dispute. They also can have very low MTBFs too. You have to KNOW what the particular parts used are before you can pin down what the MTBF REALLY is instead of speculate. > Tell you one thing, I'm not buying any Stanley LEDs! (they must have > gotten a great buy on these!) ;-) Me either! I'll try to use those billion hour ones that Rafe cites are used for traffic signs. They might be a bit large for my type of projects though...
RE: filmscanners: LED Illumination for Film Scanners
> Austin went just a bit over the edge with that 1000 > hour MTBF figure. I don't know quite what you meant by that comment. It comes across that you believe I am somehow making up the 1000 hour number I cited? Why on earth would I do that? Here is the product spec I got that information from: http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg The sheet says that 1000 Hrs. is what they GUARANTEE for "Operating Life" given the test conditions they state. That's what any designer is going to design it to unless they do their own MTBF tests. There is no doubt that there are LEDs available that (according to the manufacturers) have far longer MTBF, but since no one here knows what LEDs Nikon used, we don't know what the MTBF for the LEDs Nikon used is. You can cite all the specs you want, but unless you cite the spec for that LED Nikon used you really don't know. Do you know that the MTBF numbers you cited, were for a similar type of LED that would be used by the Nikon? > I suppose if you figure in hard mechanical > shock (like in Austin's Land Rover) the > numbers might go down a bit. Time to fix > the potholes in your driveway, Austin, > or get new shock absorbers for that beast. It's a Range Rover, the shocks are fine (relatively new gas Bilsteins) and my driveway doesn't have "pot holes" it does get washed out during heavy rainstorms. Luckily, I have a tractor with a grader to take care of it when someone complains enough.
RE: filmscanners: LED Illumination for Film Scanners
[EMAIL PROTECTED] (Austin Franklin) wrote: > > > LEDs have been around for a very long time, and they are reasonably > > > inexpensive, as well as very easy to control. I am sure that > > if this was > > > such a great idea, and the implementation worked near as well as you > > > believe, it would have been done some 15-20 years ago as a > > > commercial > > > venture, but, alas, it wasn't. > > > > > > An LED light source for enlargers was not done 15-20 years ago > > because it > > was not possible. Blue LED's did not exist as anything other than > > laboratory > > curiosities until within the last 5 years. > > Sure you could have done that 15-20 years ago. Use filters...red, > green and > blue filters certainly were around 15-20 years ago. LEDs are monochromatic light sources and can't be filtered to another colour.
Re: filmscanners: LED Illumination for Film Scanners
Moreno wrote: >And taking things one step further, a dense LED array positioned closer to >the negative could even be programmed to provide some degree of selective dodging/burning/variable constrast control. With an appropriate control mechanism, a user could adjust for dead even lighting across the easel for a specific lens/format size combination. This is a very signifficant "AhHa!" IME. I'm actually surprised that no programs, to date, are using that possibility. As I might have said sometime ago, we users *do* have some ideas worth considering, from time to time. ;-) Best regards--LRA >From: "Moreno Polloni" <[EMAIL PROTECTED]> >Reply-To: [EMAIL PROTECTED] >To: <[EMAIL PROTECTED]> >Subject: Re: filmscanners: LED Illumination for Film Scanners >Date: Wed, 20 Jun 2001 15:05:09 -0700 > > > > A couple of years ago someone on the darkroom newsgroup was working on >an > > > LED light source for enlarger heads, utilizing clusters of >high-intensity > > > LED's. I don't know what happened to the project, but at the time a >lot >of > > > people were really excited about the technology and the initial >results > > > showed a lot of promise. > > > > > > http://www.trailing-edge.com/www/led.html > > > > Yes, and look at how uneven the lighting is. > >Quoting a recent post on this list: "And you can tell this from a 72dpi >JPEG >image?" > > > For an enlarger, that may turn > > out to be a very bad idea, simply because you can't adjust each >individual > > LED for even illumination. At least with a CCD, you can adjust the gain >for > > each sensor element to get even illumination. > >Why can't you adjust each LED for illumination levels? For instance, the >designer could quite easily vary the voltage levels throughout the array to >compensate for enlarger lens light fall-off. And taking things one step >further, a dense LED array positioned closer to the negative could even be >programmed to provide some degree of selective dodging/burning/variable >constrast control. With an appropriate control mechanism, a user could >adjust for dead even lighting across the easel for a specific lens/format >size combination. > _ Get your FREE download of MSN Explorer at http://explorer.msn.com
RE: filmscanners: LED Illumination for Film Scanners
> Moreno wrote: > >And taking things one step further, a dense LED array positioned > closer to > >the negative could even be > programmed to provide some degree of selective dodging/burning/variable > constrast control. With an appropriate control mechanism, a user could > adjust for dead even lighting across the easel for a specific lens/format > size combination. > > This is a very signifficant "AhHa!" IME. I'm actually surprised that no > programs, to date, are using that possibility. As I might have > said sometime > ago, we users *do* have some ideas worth considering, from time to time. > ;-) If you could make it very very dense and were able to calibrate it somehow (which is an big task in and of it self to calibrate a 2d area this size with sufficient resolution), possibly, but I believe it won't work very well in a real implementation. You have to diffuse the LEDs in order to make them illuminate evenly, but in doing so, individual LED coverage will overlap substantially, as well as their area of coverage becoming larger. This means your area of control becomes much smaller.
RE: filmscanners: LED Illumination for Film Scanners
> Some of your recent statements of technical "fact" seem to be > casting a bit > of a shadow on your own credentials as an engineer; That comment is really out of line. If you want to question my credentials, please do so privately. There is only ONE statement I made that is in question, that I am aware of, and that is the 1000 hours life of an LED. Other than that, what other statements of technical "fact" are you referring to? > once again here are > sites with valid data: And once again, here is a scan of the information I sourced that also contains valid data: http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg
Re: filmscanners: LED Illumination for Film Scanners
Rafe wrote: >It's generally >when either the bulb or its "ballast" is near >the end of its life. Thanks for adding that note, Rafe. I didn't *think* I was going nuts, or at least not just yet, and I'm seeing that effect in an 18-month-old HP scanner. Banding on the edges, just where you'd expect it with a bad ballast. Best regards--LRA >From: Raphael Bustin <[EMAIL PROTECTED]> >Reply-To: [EMAIL PROTECTED] >To: [EMAIL PROTECTED] >Subject: Re: filmscanners: LED Illumination for Film Scanners >Date: Thu, 21 Jun 2001 17:39:31 -0400 (EDT) >> >On Thu, 21 Jun 2001, Stan McQueen wrote: > > > Fluorescents flicker at the AC line frequency--60 Hz in the US. This is > > because, as you say, the fluorescent light is a plasma device. The > > discharge turns on and off at the line frequency. It is not a continuous > > discharge (either in time or in wavelength). The UV from the mercury >vapor > > discharge tube causes the inner coating of fluorescent material to, er, > > fluoresce. The composition of the coating determines what wavelengths (I > > would normally use "frequency" but I don't want to confuse it with "AC >line > > frequency") will be emitted by the tube. > > >Stan, I was thinking of more pathological behavior. > >A healthy bulb, with a healthy, regulated power >supply, is not really my concern here. I understand >that household flourescents are driven by 60 Hz, >(in the USA) but I also know that the bulbs inside some >film scanners are driven at much higher frequencies, >and those frequencies are not well controlled. > >But I have seen, on household flourescent lamps, >situations where the plasma seems to be moving >about and varying in intensity. It's generally >when either the bulb or its "ballast" is near >the end of its life. > >And I've seen banding effects in my older >scanners (both the Microtek and the Polaroid) >which could only be explained by time-variant >spatial non-uniformity of the lamp's brightness. > >A moderate spatial non-uniformity would be >acceptable, IF it were time-invariant, at least >during the course of one scan. > > >rafe b. > > > _ Get your FREE download of MSN Explorer at http://explorer.msn.com
RE: filmscanners: LED Illumination for Film Scanners
> > > An LED light source for enlargers was not done 15-20 years ago > > > because it > > > was not possible. Blue LED's did not exist as anything other than > > > laboratory > > > curiosities until within the last 5 years. > > > > Sure you could have done that 15-20 years ago. Use filters...red, > > green and > > blue filters certainly were around 15-20 years ago. > > LEDs are monochromatic light sources and can't be filtered to another > colour. If blue LEDs were not available, outside the lab, as stated above, until 5 years ago, they why were there blue lenses available for LEDS in at least the early 90's if they did nothing?
Re: filmscanners: LED Illumination for Film Scanners
Give it a rest, guys. Please. Maris - Original Message - From: "Austin Franklin" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, June 22, 2001 5:11 PM Subject: RE: filmscanners: LED Illumination for Film Scanners | | > Austin went just a bit over the edge with that 1000 | > hour MTBF figure. | | I don't know quite what you meant by that comment. It comes across that you | believe I am somehow making up the 1000 hour number I cited? Why on earth | would I do that? | | Here is the product spec I got that information from: | | http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg | | The sheet says that 1000 Hrs. is what they GUARANTEE for "Operating Life" | given the test conditions they state. That's what any designer is going to | design it to unless they do their own MTBF tests. | | There is no doubt that there are LEDs available that (according to the | manufacturers) have far longer MTBF, but since no one here knows what LEDs | Nikon used, we don't know what the MTBF for the LEDs Nikon used is. You can | cite all the specs you want, but unless you cite the spec for that LED Nikon | used you really don't know. | | Do you know that the MTBF numbers you cited, were for a similar type of LED | that would be used by the Nikon? | | > I suppose if you figure in hard mechanical | > shock (like in Austin's Land Rover) the | > numbers might go down a bit. Time to fix | > the potholes in your driveway, Austin, | > or get new shock absorbers for that beast. | | It's a Range Rover, the shocks are fine (relatively new gas Bilsteins) and | my driveway doesn't have "pot holes" it does get washed out during heavy | rainstorms. Luckily, I have a tractor with a grader to take care of it when | someone complains enough. | |
RE: filmscanners: LED Illumination for Film Scanners
> I'm familiar with VCC, having spec'd their products many times over the > years. The early blue (and green) lenses were intended for use with small > incandescent bulbs with the same T1-3/4 form-factor that was > adopted by LED > manufacturers, not for use with LEDs. Not the ones I was referencing. They specifically say "Lensmounts for 3mm and 5mm LEDs". The catalog is titled "LED Lenses and Mounting Components". Here is a scan of the VCC sheet for the ClipLite from this catalog (dated 4/92): http://www.darkroom.com/MiscDocs/VCCClipLite4LEDs.jpg I believe the timeline you previously claimed for when blue LEDs were available is possibly wrong. At least according to the LED Museum link you provided in the same post. It appears it is white LEDs that were available about five years ago (1996). Blue LEDs appear to have been available somewhat early in the 1990s. That would make sense that these LED covers were intended (though it's not stated they were) for like colored LEDs. If blue covers were for blue LEDs, that would make sense that blue LEDs would have been reasonably available at the latest by very early 1992 (when the VCC catalog shows blue covers), if not a year or more before.
Re: filmscanners: LED Illumination for Film Scanners
Hi, Art, you wrote: >I also expect, like incandescents and other illumination sources, LEDs can be built to different specs and be run under different electronic designs. For instance, I know that with the halogen bulbs used in projectors, a "25 hr" bulb will last only 5-11 hours run at 125-130 volts, will last 11-19 hours run at 120-125 volts, they last upwards of 19-31 hrs run at 115-120 and if you use a dimmer to bring them down to about 105-110, they will last well over 50 hours, or nearly one order of magnitude from running at 125-130 volts. This is so darned off-topic that I'm going off-list, but aren't halogen bulbs supposed to be used with direct current? That would presuppose an ac-dc converter, which could/should deliver optimum power for the device, unless you're using one tagged to a variable battery-charger. Granted that Manufacturer Specs are almost always overstated, halogens are the most reliable light devices I know of. LED's are another matter, and idealy they're run from DC as well, although it isn't a requirement since they modulate the power themselves, AFAIK. OTOH, they come in all flavors, and one isn't the same as another. Whatever--I just wondered how Austin got us so far off-topic. ;-) Best regards--Lynn _ Get your FREE download of MSN Explorer at http://explorer.msn.com
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: > > > Some of your recent statements of technical "fact" seem to be > > casting a bit > > of a shadow on your own credentials as an engineer; > > That comment is really out of line. If you want to question my credentials, > please do so privately. There is only ONE statement I made that is in > question, that I am aware of, and that is the 1000 hours life of an LED. > Other than that, what other statements of technical "fact" are you referring > to? I think he is refering to your jibe of: > My guess is you are not an electrical engineer, or you would > know that LEDs > do have a life span. Because you haven't heard of them burning > out, doesn't > mean they don't burn out. In fact, their typical MTBF is rated > for 1000 > hours. Not only is the "typical" life of LEDs far longer than what you have asserted, you were pretty snide when you pointed out your "superior" knowledge. So I guess all the above poster was pointing out is what goes around comes around...:-) > > > once again here are > > sites with valid data: > > And once again, here is a scan of the information I sourced that also > contains valid data: > > http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg Yes, but how did you extrapolate this into "typical" performance for all LEDs? Isaac
RE: filmscanners: LED Illumination for Film Scanners
>LEDs are monochromatic light sources and can't be filtered to another colour. Seems to me I've seen LEDs in at least 6 different original colors, and I wasn't paying that close attention. Red, blue, amber and green are the most common. FTM, any white light source can be filtered. _ Get your FREE download of MSN Explorer at http://explorer.msn.com
RE: filmscanners: LED Illumination for Film Scanners
> > In fact, their typical MTBF is rated > > for 1000 > > hours. > Not only is the "typical" life of LEDs far longer than what you have > asserted, You are right, ALL LEDs are not typically rated for 1000 hours. The typical was meant only for the LEDs I was referencing, not for all LEDs. Saying "their" was clearly my mistake. All I meant to point out was that there are LEDs that have as low an MTBF as 1000 hours. > Yes, but how did you extrapolate this into "typical" performance for > all LEDs? Via mistake.
RE: filmscanners: LED Illumination for Film Scanners
> Whatever--I just wondered how Austin got us so far off-topic. ;-) You give me too much credit here! I believe it was the enlarger light source that was what brought this way off topic. I believe that honor goes to Sr. Polloni. OK, I'll take some credit. None the less, at least for me, and despite any statements that were "rough around the edges" it has been a very interesting discussion.
RE: filmscanners: LED Illumination for Film Scanners
Austin wrote (re selective burning w/film scanners) >If you could make it very very dense and were able to calibrate it somehow >(which is an big task in and of it self to calibrate a 2d area this size >with sufficient resolution), possibly, but I believe it won't work very >well in a real implementation. You have to diffuse the LEDs in order to >make them illuminate evenly, but in doing so, individual LED coverage will >overlap substantially, as well as their area of coverage becoming larger. >This means your area of control becomes much smaller. I think you may be unnecessarily complicating the problem. Let's say software and the scanner driver could control what areas were exposed, and for what duration, in selected areas. This would be the equivalent of the darkroom technician's "Dodge & Burn," but would occur during the scan. A common problem is the photo with deep shadow and very bright highlights. A scanner with DR 3.3 will screw one end of the histogram or the other in this case, just as the film did. With selective exposure, the operator could select one area or another for a different exposure rate from the scanner. On balance, you'd have DR 3.3 in the highlights and DR 3.3 in the shadows, but they would have different DMin/DMax values in the selected areas--different white/black points, so to speak. Do you see what I'm getting at, here? I'm not a good enough programmer to do such a thing, but given controlable variables in a scanner--which some now in fact have--it sould be done. Whether it's done with LED's or head speed is immaterial, it's doable. Just another idea from the Rust Belt. Best regards--LRA > > > Moreno wrote: > > >And taking things one step further, a dense LED array positioned > > closer to > > >the negative could even be > > programmed to provide some degree of selective dodging/burning/variable > > constrast control. With an appropriate control mechanism, a user could > > adjust for dead even lighting across the easel for a specific >lens/format > > size combination. > > > > This is a very signifficant "AhHa!" IME. I'm actually surprised that no > > programs, to date, are using that possibility. As I might have > > said sometime > > ago, we users *do* have some ideas worth considering, from time to >time. > > ;-) > >If you could make it very very dense and were able to calibrate it somehow >(which is an big task in and of it self to calibrate a 2d area this size >with sufficient resolution), possibly, but I believe it won't work very >well >in a real implementation. You have to diffuse the LEDs in order to make >them illuminate evenly, but in doing so, individual LED coverage will >overlap substantially, as well as their area of coverage becoming larger. >This means your area of control becomes much smaller. > _ Get your FREE download of MSN Explorer at http://explorer.msn.com
Re: filmscanners: LED Illumination for Film Scanners
That didn't go off-list, did it? :-( Sorry--LRA >From: "Lynn Allen" <[EMAIL PROTECTED]> >Reply-To: [EMAIL PROTECTED] >To: [EMAIL PROTECTED] >Subject: Re: filmscanners: LED Illumination for Film Scanners >Date: Sat, 23 Jun 2001 02:55:41 - > >Hi, Art, you wrote: >>I also expect, like incandescents and other illumination sources, LEDs >can be built to different specs and be run under different electronic >designs. For instance, I know that with the halogen bulbs used in >projectors, a "25 hr" bulb will last only 5-11 hours run at 125-130 >volts, will last 11-19 hours run at 120-125 volts, they last upwards of >19-31 hrs run at 115-120 and if you use a dimmer to bring them down to >about 105-110, they will last well over 50 hours, or nearly one order of >magnitude from running at 125-130 volts. > >This is so darned off-topic that I'm going off-list, but aren't halogen >bulbs supposed to be used with direct current? That would presuppose an >ac-dc converter, which could/should deliver optimum power for the device, >unless you're using one tagged to a variable battery-charger. > >Granted that Manufacturer Specs are almost always overstated, halogens are >the most reliable light devices I know of. LED's are another matter, and >idealy they're run from DC as well, although it isn't a requirement since >they modulate the power themselves, AFAIK. OTOH, they come in all flavors, >and one isn't the same as another. > >Whatever--I just wondered how Austin got us so far off-topic. ;-) > >Best regards--Lynn >_ >Get your FREE download of MSN Explorer at http://explorer.msn.com > _ Get your FREE download of MSN Explorer at http://explorer.msn.com
RE: filmscanners: LED Illumination for Film Scanners
> Austin wrote (re selective burning w/film scanners) > >If you could make it very very dense and were able to calibrate > it somehow > >(which is an big task in and of it self to calibrate a 2d area this size > >with sufficient resolution), possibly, but I believe it won't work very > >well in a real implementation. You have to diffuse the LEDs in order to > >make them illuminate evenly, but in doing so, individual LED > coverage will > >overlap substantially, as well as their area of coverage > becoming larger. > >This means your area of control becomes much smaller. > > I think you may be unnecessarily complicating the problem. Let's say > software and the scanner driver Oops...I think you took this in the wrong context. My discussion was aimed at enlargers using LEDs as a light source, not scanners. Enlargers require a 2d array, not a 1d array. With the scanner, it is very easy to calibrate along the 1d axis...and the points are discrete. For an enlarger, they are not discrete...and they are 2d.
Re: filmscanners: LED Illumination for Film Scanners
Austin Franklin wrote: > > > > In fact, their typical MTBF is rated > > > for 1000 > > > hours. > > Not only is the "typical" life of LEDs far longer than what you have > > asserted, > > You are right, ALL LEDs are not typically rated for 1000 hours. The typical > was meant only for the LEDs I was referencing, not for all LEDs. Saying > "their" was clearly my mistake. All I meant to point out was that there > are LEDs that have as low an MTBF as 1000 hours. > > > Yes, but how did you extrapolate this into "typical" performance for > > all LEDs? > > Via mistake. Fair enough... I vote we kill this topic before we get any more snippy..:-) Myself included... God what time is it anyway?:-) Isaac
RE: filmscanners: LED Illumination for Film Scanners
On Thu, 21 Jun 2001 07:51:13 -0400 rafeb ([EMAIL PROTECTED]) wrote: > I can easily imagine packages and monolithich LED arrays > that allow for very, very close stacking of LEDs, so that > you could have, say, 20 or 50 or even 100 LEDs per inch. ISTR reading somewhere that the LS1000 used an array of 128 LED's. May be me hallucinating tho'. The number 128 looks weird as it's not divisible by 3 (R,G & B), but quite likely the array is RGBG... Regards Tony Sleep http://www.halftone.co.uk - Online portfolio & exhibit; + film scanner info & comparisons
Re: filmscanners: LED Illumination for Film Scanners
From: Isaac Crawford <[EMAIL PROTECTED]> > Not only is the "typical" life of LEDs far longer than what you have > asserted, you were pretty snide when you pointed out your "superior" > knowledge. So I guess all the above poster was pointing out is what goes > around comes around...:-) Austin is frequently so biased that he is (by default) wrong. Please don't ask him which car is the *absolute best* in the history of the world, as we'll likely then have to suffer through a three day Range Rover rant. :) Dave
RE: filmscanners: LED Illumination for Film Scanners
[EMAIL PROTECTED] (Austin Franklin) wrote: > > > > An LED light source for enlargers was not done 15-20 years ago > > > > because it > > > > was not possible. Blue LED's did not exist as anything other than > > > > laboratory > > > > curiosities until within the last 5 years. > > > > > > Sure you could have done that 15-20 years ago. Use filters...red, > > > green and > > > blue filters certainly were around 15-20 years ago. > > > > LEDs are monochromatic light sources and can't be filtered to another > > colour. > > If blue LEDs were not available, outside the lab, as stated above, > until 5 > years ago, they why were there blue lenses available for LEDS in at > least > the early 90's if they did nothing? > > My guess would be that perhaps a conventional red LED (not a laser LED, which by definition is strictly monochromatic) has a wider spectral response than I gave it credit for. You wouldn't be able to get much light by filtering something that is mainly red though. The blue LEDs recently developed have most of their output in the blue part of the spectrum, and are rather better suited for turning into blue laser diodes. I know laser diodes aren't used in film scanners, but the driving force in LED development is for lasers in shorter and shorter wavelengths to allow optical storage of higher and higher capacity.
Re: filmscanners: LED Illumination for Film Scanners
I have no idea if Nikon uses them, but blue LEDs have a particularly short lifespan. In any case I dont think occasional users like us will ever burn out an LED array or a tube. Definitely not me, I keep my scanner off when not in use. "Austin Franklin" <[EMAIL PROTECTED]> wrote: >My guess is you are not an electrical engineer, or you would know that LEDs >do have a life span. Because you haven't heard of them burning out, doesn't >mean they don't burn out. Herm Astropics http://home.att.net/~hermperez
Re: filmscanners: LED Illumination for Film Scanners
At 02:55 AM 6/23/01 -, Lynn Allen wrote: >Whatever--I just wondered how Austin got us so far off-topic. ;-) It wasn't just Austin... a few of us were playing along. And it wasn't all the awfully off-topic; it was all about how scanners work. Granted a bit more speculation than one would like, since facts (about scanner internals) are often scarce. rafe b.
RE: filmscanners: LED Illumination for Film Scanners
Hi, all! I'm a little behind in my reading of this list, but I thought I'd respond to this one. One thing that hasn't been mentioned (that I noticed) is that this spec from Stanley gives the operating life as 1000 hrs at IF=max. Usually, we power an LED with something on the order of 20mA current. This is the TYPICAL forward current. LEDs can be made to run somewhat brighter by pushing more current through them. This IF=max probably means a forward current of over 50mA. Beyond that, the LEDs in my experience undergo a wavelength shift (color shift) and begin to get dim after a short period of time. (I've designed and build many different LED strobe devices for previous employers). One can get around this by pulsing the LED. I've pushed over 500mA through an HP ultra bright LED by pulsing it at on the order of one microsecond with a repitition rate of 1kHz or so and seen no degradation of the LED over the lifetime of the device (several years). YOu just don't want the LED to heat up too much, or it dies quickly. Anyway, Stanley was simply giving a conservative estimate of the longevity of their LED when powered by the highest allowable current. Of course, at a sane current drive, they will last for hundreds of thousands of hours, at least!!! On a historical note, back in the late '80s and early '90s, blue LEDs were very dim. They were made from Silicon Carbide, and put out less than 100mCandela while good red or green LEDs put out in excess of 1000mC. I used a particular HP red diode that put out 3-4 Candela! from a T-1 package (small size). In the mid '90s I saw an example of a (then) US$50 blue ultribright diode. It put out at least one full Candela of power, but it was too expensive for my medical diagnostic device application. This weekend, I went to the drugstore to buy some film , and found a blueish-white LED flashlight for sale for ~$8.00 It is blinding in it's intensity!!! Clearly the state of the art is moving forward at quite a rapid pace! Hope this helps!! Guy Clark --
Re: filmscanners: LED Illumination for Film Scanners
Just two days ago, I was at a local retailer who showed me the new HP scanner/copier (which is basically an inkjet printer and a scanner on top). It was only $399 CAN, and they have reduced the footprint to that of a small inkjet printer. Pretty amazing. The part that relates to this discussion, is that they were just setting it up, and that included a calibration process for the color and black ink heads. It is now an automatic function. The printer printed a set of varying matrixes and lines in both black and yellow. The print head also had a very bright blue LED which went on, and I expect some type of sensor, which read the resultant printout. When the blue LED was on, the yellow printing probably became gray, and the system probably tried to find the placement of the yellow and black lines which made for the highest contrast (which would be when the yellow and black lines printed on top of each other exactly). I assume the unit than used this information to either physically adjust the head positioning, or to change the printing pattern to use certain nozzles and delays between the black and color cart, so that the ink would be printed onto the paper in registration. It was a pretty neat idea, all made affordable by a blue LED. Art Clark Guy wrote: > > On a historical note, back in the late '80s and early '90s, blue LEDs were > very dim. They were made from Silicon Carbide, and put out less than > 100mCandela while good red or green LEDs put out in excess of 1000mC. I > used a particular HP red diode that put out 3-4 Candela! from a T-1 package > (small size). In the mid '90s I saw an example of a (then) US$50 blue > ultribright diode. It put out at least one full Candela of power, but it > was too expensive for my medical diagnostic device application. This > weekend, I went to the drugstore to buy some film , and found a > blueish-white LED flashlight for sale for ~$8.00 It is blinding in it's > intensity!!! Clearly the state of the art is moving forward at quite a > rapid pace! > > Hope this helps!! > > Guy Clark > > > --
RE: filmscanners: LED Illumination for Film Scanners - Apology
Austin, Please forgive my comment; you're right, it was out of line, and I'm sorry to have offended. Cliff Ober -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Austin Franklin Sent: Friday, June 22, 2001 8:17 PM To: [EMAIL PROTECTED] Subject: RE: filmscanners: LED Illumination for Film Scanners > Some of your recent statements of technical "fact" seem to be > casting a bit > of a shadow on your own credentials as an engineer; That comment is really out of line. If you want to question my credentials, please do so privately. There is only ONE statement I made that is in question, that I am aware of, and that is the 1000 hours life of an LED. Other than that, what other statements of technical "fact" are you referring to? > once again here are > sites with valid data: And once again, here is a scan of the information I sourced that also contains valid data: http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg
RE: filmscanners: LED Illumination for Film Scanners - Apology
Cliff, thank you I appreciate it. Since you seem to know quite a bit about LEDs, what do you believe Nikon uses for an LED light source for this new scanner?
Re: filmscanners: LED Illumination for Film Scanners - Apology
Cliff Ober wrote: > > Austin, > > Please forgive my comment; you're right, it was out of line, and I'm sorry > to have offended. > > Cliff Ober I've got an even better idea, let's take this whole LED business, which was once interesting and is now degenerating into "Did not" "Did too" silliness, offline. I'm as guilty as anyone else, and I promise that this is my last post on this topic... Isaac > > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED]]On Behalf Of Austin Franklin > Sent: Friday, June 22, 2001 8:17 PM > To: [EMAIL PROTECTED] > Subject: RE: filmscanners: LED Illumination for Film Scanners > > > Some of your recent statements of technical "fact" seem to be > > casting a bit > > of a shadow on your own credentials as an engineer; > > That comment is really out of line. If you want to question my credentials, > please do so privately. There is only ONE statement I made that is in > question, that I am aware of, and that is the 1000 hours life of an LED. > Other than that, what other statements of technical "fact" are you referring > to? > > > once again here are > > sites with valid data: > > And once again, here is a scan of the information I sourced that also > contains valid data: > > http://www.darkroom.com/MiscDocs/StanleyLEDTestData.jpg
OT Discussion was Re: filmscanners: LED Illumination for Film Scanners
Guys, could we please take the LED discussions off the list? While they may be interesting to the engineers amongst us, I don't think they're of much interest to those who are subscribed to discuss filmscanning? I think we can all agree that the mechanical components of a scanner are more likely to fail than the LEDs, and the LEDs are likely to outlast the useful life of the rest of the scanner hardware. If you disagree - please email me *off* the list. :) Rob