On Sat, Nov 21, 2009 at 06:45:37PM -0600, Jon Elson wrote: > > /proc/version reads : > Linux version 2.6.15-magma (r...@ubuntu) (gcc version 4.0.3 (Ubuntu > 4.0.3-1ubuntu5)) #1 Fri Ju > n 9 20:51:19 EEST 2006 > > This system was built from a linuxcnc iso image, you can see how old the > kernel is.
Great, that's probably the best-tested EMC2 kernel ever. I am positive it does not have a problem with PS2 keyboards. > The keyboard is PS/2, and so it the mouse (actually trackball). > > I have not changed or updated the kernel. So, the only difference is > the update from > an early 2009 EMC2 version to the development head from about 4 days > ago, when > I was testing out the homing klunk change. OK, so you updated from some oldish cvs checkout to the current git master. It's a little hard to tell exactly what changed then. But I don't think we've touched the jog arrow code for a while. > I suppose it could just be a bum keyboard or motherboard, and the spate > of "stuck > keys" could just be a fluke, as I had been seeing this very occasionally > before. It would sure be nice if you could say whether this happens with a different keyboard and/or computer. I know changing the computer isn't trivial, but maybe changing the keyboard is. > No, I never jog by mouse. I do most jogging around with the MPG now, > but when > I want a manual feed across a part at constant rate, I use the jog keys. Good to know - it's not any of those obscure mouse/kb interactions. > > When it happens, does it jog for a while extra and then > > stop, or does it go forever? > I think it would go on forever. It definitely was not just a GUI lag, > it continued for several > seconds in all 3 events I had. After the several seconds did you estop or hit a soft limit or something else? What do you do to stop it, or does it stop on its own? [later: I see this is answered below - is it always true that you stopped it by tapping another key?] > /proc/cpuinfo shows 730 MHz Pentium III, 256 MB, and it has an Intel 82810E > graphics chip. OK, not the fastest machine, but probably not the slowest in use either. > As for doing/seeing, I was just doing many manual jogs with the keyboard > arrow keys. I would step down in Z with the MPG, then hold an X arrow > to make a cutting pass across the material. Then step down and make another > pass the opposite direction. If there might be a pattern to it, it > seemed all the > "runaways" were in the minus direction, but the sample set is pretty > small to > be sure that is of any significance. Interesting! Do you think using the wheel is a necessary part of it? I know there is interaction between wheel jogging and keyboard jogging, but it sounds like you were using wheel jogging on Z and keyboard jogging on X. In this case I think they are completely separate (but I'd have to check the code to be sure). This is a kind of interaction I hadn't considered and I bet it is much more lightly-tested than keyboard alone. > The runaway was just that the axis would continue moving when I took my > finger off the key. I could stop it by momentarily tapping one of the > jog arrow > keys, something I learned from the old TkEMC bug. Now that is a very interesting finding... I'm not sure how it fits in. > Well, I will try out another keyboard and see if I see a difference. Great, and when you test some more, concentrate on wheel vs keyboard interaction. I have a nagging feeling that's part of it. If you can reproduce it, we can fix it! Thanks for these really great details. Chris ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
