On Sun, Jul 20, 2003 at 11:21:28AM -0500, Charles Steinkuehler wrote: > Matt Schalit wrote:
> > > http://bund.dk/blinder > >That's awesome. Really good stuff. Close to what I > >was posting about. You even have some sort of > >rudimentary input, the kill switches you added to > >the blinds. Thanks :) > >Charles hasn't posted on this thread, but he may not > >have seen it. Too bad because he built a Battle Bot > >and has quite a bit of computer interfacing > >experience. > > OK, OK...hint taken. :) :D > >I got in A's in Analog Lab, Digital Lab, and in Computer > >Interfacing, but he's actually doing it in real life > >having to deal with Voltage converters, grounding planes, > >filters, the whole gamut. How much fun is that? A lot? > >What's cool about the GP-3 is that it does the stepper > >motor control already, along with tons of other good > >stuff. I was sort of surprised it was only $35 US. > >I could run a farm with that little thing. > > For cheap & easy DAC output from your PC, use the pre-existing ports. I was unemployed at the time, so reason #1 weighed heavily. I wouldn't really have agreed with #2 at the time, but then, I don't know squat about electronics. (Sure I can solder, but...) > The audio output ports for dynamic signals (although you won't typically > get DC output levels without shorting across the output caps and adding > some level-shifting circuitry to deal with the typical 1.5-2.5 V offset). I would have thought exactly that; If it needs to become analog, why not use a soundcard? > NOTE: If you really want to smoothly drive stepper moters with a PC, > you either need smart hardware or you'll quickly find out why linux (and > most other general purpose OS's) are not considered "real-time". The > stepper motors will *NOT* be happy when the pulses don't come out at > exactly the wrong time, and you *WILL* notice the "glitch" in smooth > movement, by feel and/or by sound. A bit of timing uncertiancy is not > that bad (especially if you can compensate for it with the software > generating the drive waveforms...motors have a *MUCH* slower response > time than modern CPUs), but you can't just call the system sleep routine > for a time delay and expect nice, clean waveform timing on output. Actually sleep comes in too large chunks. What I do is read the parallelport a calculated number of times per 'full-sequence', to get the slowdown. I read somewhere that this read takes approx. X microseconds so I just multiplied X by so and so much. It's not actually that precise, but at 5:30 in the morning, believe me, I don't notice ;) > This is really an instance where a $2 8-bit MCU with a couple of PWM or > timer outputs can solve a problem better than a GHz+ CPU. :) Well... In this particular case (a ~25 MHz 486) I doubt it will ever be smooth as in "silk"... But it's not really that important for this application. > > > I just picked up development a couple of days ago. Basically the bit > > > drives the stepper motor is a script that pushes out bit-sequences of > > > parallelport. But being a script (and not particularly well written, at > > > that) it makes for some jerky motion of the motor, which I think is in > > > responsible for the two mechanical failures it has suffered in the year > > > been running. (Plastic fatique). > See NOTE:, above. Noted. > >I sort of thought that you just tell those stepper motors > >to slew and use an infra-red transmitter/pickup system that > >counts the times the IR beam is interrupted as it shines through > >the slits in the stepper's axel mounted circular disk. On the contrary, as Charles explains below; > >And I thought you pretty much used the stepping feature > >for "fine" positioning.... > > Stepper motors require proper sequencing of the drive waveforms. They > are not real picky about signal timing if they're not actually moving > very fast, but you *DO* need to provide for acceleration/deceleration > ramping and if the motor is actually rotating with any speed, it will > not react well to out of phase drive signals, potentially putting lots > of strain on the drive electronics and/or mechanical drive system. Ramping... Never thought of that... That's a very nice piece of input. I shall keep it in mind, when I get around to the testruns of the binary I'm building. I'm pretty anxious to see how fast full-tilt is going to be ;) > Also, I'm not sure what you mean by the "stepping feature" only being > used for fine positioning. You *HAVE* to "step" stepper motors...that's > how they work. You can't simply hook a stepper motor up to power and > get it to turn (at least not more than one step). Getting a stepper > motor to rotate *REQUIRES* a properly timed dynamic drive signal. > > If you're referring to "micro-stepping", or driving the stepper coils > with something other than full voltage, this technique is useful even if > you don't need it for fine positioning. If you think about how a > stepper motor works, the "steps" correspond to different arrangements of > the drive coils being fully energized. At all the points in-between the > natural steps of the motor, the different drive coils need to be > partially energized (think sine-waves). While it's not required that > you drive stepper motor coils with a sine-wave like signal, it will > proivde smooter rotation, especially at lower speeds. Hmmm... but no. Not this time around ;) > >I love the real world, learn-by-destruction feedback > >you're getting. I'd like to model that in autocad :) > > There's nothing like the real world to teach you how everything REALLY > works (or just as importantly doesn't work). :) So very, very true /Jon -- If we can't be free, at least we can be cheap! ------------------------------------------------------- This SF.net email is sponsored by: VM Ware With VMware you can run multiple operating systems on a single machine. WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the same time. Free trial click here: http://www.vmware.com/wl/offer/345/0 ------------------------------------------------------------------------ leaf-user mailing list: [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user SR FAQ: http://leaf-project.org/pub/doc/docmanager/docid_1891.html
