On Thursday 24 September 2009, Sven Wesley wrote: >I'll see what I can do, I have an aftermarket parallel port though. I >decided to take the cards out from the box and make a new PID tuning, I >think it's not correct. Anyway, something is fishy. The documentation is >saying that far higher step rates should be no problem at all. >BTW, SPI is not the same thing as Kirk's "dumb drivers", right? These cards >are supporting SPI communication as well.
SPI? I'm not familiar with that acronym. One of the perils of being a retired old fart even if I am a C.E.T. The aftermarket parport would I expect be strong enough for the job, no one here has mentioned finding a poor one yet. So that means we are back to timing of the pulses, and noise considerations. The interface may need a certain 'settling time' which could be several microseconds before it is ready to accept a step signal and treat it as the direction chosen. Find the docs if you haven't already, and verify the requirements. There are ways of setting that but I've never had to do it, I just chose xylotex output and all that seems to be set correctly for my little toy mill. I did find a few months back while I was attempting to optimize my .ini by hand, that if I forgo the rapid acceleration, it moves smoother on complex stuff, and can go 2x faster on long moves. So it takes about the same time to execute the default logo carving, but the corners and other details seem to be, as far as the video trace is concerned, a lot more accurately displayed. That part is 100% a video speed problem with my machine, it has an old TNT video card, using the vesa driver. emc does NOT need super fawncy video from a $300 card that needs its proprietary driver to get a glxgears output above 150 fps. That proprietary driver has been known to screw up emc's heartbeat terribly, locking out the interrupts that are part and parcel of emc's smoothness when running steppers directly, wrecking parts in the process. The vesa driver, while not as speedy as one would like to play the latest whiz bang game, doesn't appear to muck with the interrupts. Both nvidia drivers, and the one time I tried an ati 9200se card with the ati driver was a bigger disaster than the ndvidia drivers were. The machine is actually moving within .0005" of the gcode described path, that is the error bandwidth I have setup. The machine itself isn't that close. Harbor Freight Micromill, with bigger tables and a home made Z axis, virtually a given requirement as the z screw is behind the post and binds the slider very badly even when the heads weight is counterbalanced well. I moved the screw to the front of the post, and can now put 20x the force on a drill bit than I could before, 155 lbs. I can drill holes with gcode now. <http://gene.homelinux.net:85/gene/emc> for some pix of my mess. However, I don't think this has anything to do with your problem, but since all video cards can be driven by the vesa driver, it won't take (usually) anything but an edit to xorg.conf, and an X restart to check it. >2009/9/24 Gene Heskett <[email protected]> > >> On Thursday 24 September 2009, Sven Wesley wrote: >> >Yes, maybe. What should I look for? >> >> The voltage levels your parport is capable of outputting, and whether >> there may be extra noises on the signals. The scope will need to be dc >> coupled of >> course, and have good bandwidth. The 100mhz dual trace that most techs >> carry >> is about right for this sort of stuff. >> >> Your parport should be capable of outputting a logic zero that is well >> below >> .2 volts, which gives about a .4 or .5 volt margin for noise. >> Conversely, the logic 1's should be 2.9 volts or better ttto have that >> same half a volt for noise margin, and here, generally speaking, the >> pullup ability is rather >> puny, so noise on the line becomes more important. There is a voltage >> range >> from about .6 volts to around 2.4 volts where the chip makers will not >> say what the output will be, so the driving voltages must be more widely >> separated in order to be running at guaranteed valid signaling voltages. >> >> Sure, one can measure this with a multimeter, but it doesn't show you the >> noise that screws things up. The scope will. And some parport driver >> chips >> are known to have very low driving abilities. The cure there is to use >> an aftermarket card, which will generally have much stronger drivers on >> it. >> >> 'star' grounding schemes are also good practice. Just tying all the >> grounds >> together all over the place will often cause ground loops, which can pick >> up >> large amounts of noise & muck with things in strange ways. The shield >> coverings of my motor cabling are grounded only at the driver terminal >> ends, >> and not to the machine, which has its own frame ground via the power cord >> and >> the variable speed spindle controller. I can even do EDM with an 80 >> volt, 6 >> amp capable supply running at 3 amps, so noisy my 75 year old ears >> object, but none of that noise gets into the control circuits badly >> enough to effect >> the machine. >> >> >2009/9/24 Gene Heskett <[email protected]> >> > >> >> On Wednesday 23 September 2009, Sven Wesley wrote: >> >> >I'm still loosing on Z by the way, I'm starting to suspect a faulty >> >> >> >> driver. >> >> >> >> Do you perchance have access to an oscilloscope? >> > -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) The NRA is offering FREE Associate memberships to anyone who wants them. <https://www.nrahq.org/nrabonus/accept-membership.asp> /* * Please skip to the bottom of this file if you ate lunch recently * -- Alan */ -- from Linux kernel pre-2.1.91-1 ------------------------------------------------------------------------------ Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
