Can NOT get a PC to run Xylotex without freaking out, laptop is OK!?

[Riceburner98]

Ok, spent many hundreds of $$ here now over the last few years trying to get a PC together to run my Xylotex board, same bad results every time! (although now I have a screamer for CAD/CAM which is good..) Tried 4 or 5 different motherboards, processors, parallel port cards, power supplies everything! Completely different computers have the exact same problem. Moving any axis on my Taig mill, I get a random "extra" pulse every 1 second or so that just freaks out the motors. Sometimes they recover, sometimes they lock up. I can see the extra pulse on the 'scope, so it's there for real. I've been using my HP Pavilion laptop for the past 2+ years now to run the board no problem, but have yet to get a PC to work. Mach2, Mach3, Win2000, WinXP, Pentiums, AMD's, all give me the same crap. The laptop is getting slower and slower, it's only a matter of time before it dies, need to figure out what's up before that happens! I've tried not grounding the PC to the wall, directly grounding it to the Xylotex box, nothing. Anyone have a specific method of connecting thier PC to their Xylotex board that works for sure? I'm going to start smashing stuff soon, and that's never good!

[pminmo]

is the board old enough that it doesn't have the conditioning filters on it?

[Riceburner98]

Thanks for the reply! I don't *think* it's an older one.. It was made sometime around 4/2005, forget what revision it is.. I've got the controller box removed from it's usual spot so I could take a look to be sure - what would they be? I can physically see the extra pulse / spike on the parallel port step lines with the 'scope, which isn't there when I use the laptop. I can't remember now how much voltage the extra pulse is, but I seem to recall it being a larger pulse than the normal step pulse stream. Wondering if I could diode or opto-isolate the step lines temporarily to see if the pulse is coming from the PC's or the Xylotex. Can't see how the Xylotex would be causing it since it doesn't happen with the laptop, but 4 or 5 totally different PC's..? I think I'm going to try and bring the controller box and a spare stepper into work to try on a few new-ish IBM PC's to see if it does the same. At least I have a proper 'scope setup there. Frustrating to say the least!

[pminmo]

Does the datasheet for yours say filterd and buffered? I don't know when they added the filters.

In all probability its an induced spike on a logic gnd that happens based on your 0V layout or caused by wire routing. The laptop in all probability doesn't couple the parallel ports gnd pins back to AC safety. If your gnd wiring ties multiple places for safety gnd, especially on seperate AC breaker circuits could also be an issue. Make sure that motor wires don't physically route near the logic or switch lines as well.

Some explanation here: http://pminmo.com/PMinMOwiki/index.php5?title=Grounding

[ger21]

If you run the Mach3 Driver Test, does it show a flat line?

[Riceburner98]

After the first few PC's failed with the pulsing problem I started leaning towards a ground problem, but have so far been unable to find it. The first controller box was a standard "Deepgroove" setup from eBay, the wiring was adequate, but not great. I rebuilt my controller box into what's shown in the pictures below, to hopefully eliminate its construction as a problem.
All stepper wiring is 12gauge stranded wiring, the only place it crosses any low-voltage wiring is directly under where it leaves the panel, those are the enable switches and the wiring is perpendicular, also I didn't have the enable switches on the old box. When the stepper wiring leaves the box, it's again 12gauge, now shielded with a braided sleeve, tied only to the chassis ground of the control box. (which is only tied to earth ground) Not sure if the power supply ties the return to earth ground, but it's a farily good industrial supply similar to what we use at work, which are ground isolated even between multiple output voltages.
The Xylotex board isn't connected to chassis ground AFAIK. (Multimeter is outside in the freezing cold car, I'll have to check that tomorrow) The only GND connection @ the parallel port cable should be the individual ground pins, as it looks like the connector housing isn't electrically connected to anything. (and I removed it from the metal box back panel just to be sure) Can't think of anything else at the moment... As far as I can tell I have it wired as per the diagram on your grounding page. I even tried not grounding the PC to see if it was a ground loop there, no dice.

(Pictures->) The front panel of my 'box is a bit busier looking than it really is right now, for future expansion. The right-angle AMP connectors above the stepper Molex connectors are for future limit and home switches (not connected now), the switches and LED's below the Molex connectors are individual motor enables so I can turn off each channel individually. The "joystick" port is for a future analog joystick control, wich will use a seperate circuit board to generate step and direction pulses directly into the Xylotex board. The bunch of 1/8" 4-pin stereo jacks are for "something" in the future. Only one connected now is for the E-stop button. Then there's the main ON/OFF toggle switch which cuts the 120v to the power supply. Next to that is the "master enable" switch which is tied to all the enables, so I can quickly disable all the channels for quick manual moves with the knobs, or when changing tools or stuff like that. The box is completely sealed except for the fan ports on the back, and pull all incoming air through the holes in the top and side, so all incoming air passes through the stepper driver heatsinks. Those holes are sealed / ducted to the outside of my CNC 'cabinet' to keep the incoming air clean.

'WHEW!' That was a long explanation! Just wanted to share a little about my box, I was pretty happy with it when I finished. The enclosure itself was a bit of magic also... I was looking around work for a spare enclosure I could use, or spare metal to make one. Nobody had seen anything like I needed. Then the next day this box showed up on my chair in my cube, which perfectly fit everything! Even matched the round contours of the power supply that I already had. Scary. I asked everyone I had asked the previous day, and nobody knew where it came from. People randomly leave stuff in my cube like that all the time, but nothing as perfectly timed as this. I found it online and I think it was originally like a $300 enclosure for something or other.. It was still shrink-wrapped, but covered in dust from sitting somewhere on a shelf for years. Enough about that I guess! (just added a picture of what the "shop" usually looks like while in action. What a mess! I can close the doors on the cabinet while the mill is in action if it gets too loud. It's in my spare bedroom in my apartment.)

(Just got notification of the other reply while typing this... As far as I recall, the driver test showed a near flat line, but I'll have to re-run it to double-check. I'll try it on the 1.4Ghz laptop after I close all my open programs later today, and on the 800mhz AMD and the 3Ghz Pentium and report back.)

[pminmo]

Try disconnecting the cable for your limit switches and see if that makes a difference. you may have to change the software to fake out the limits.

[Riceburner98]

I actually don't have any limit switches connected, only the plugs are there on the front. (there's no wires attached to the plugs.. Someday!) I just ran the driver test on all 3 machines I have now. The 1.8Ghz Pavilion laptop is the furthest thing from a flat line I could imagine. It looks like a voice pattern from someone talking into a microphone! This is the one computer that actually works! The 800Mhz AMD is pretty close to a flat line, with a few small peaks here and there. The 3Ghz Pentium (1G ram) looks about the same as the AMD, with a couple spikes here and there. One thing I don't think I checked before was to run Mach without the Xylotex attached and the 'scope on the parallel port of the PC's. That should tell me if it's some kind of pulse coming from being connected to the Xylotex, or if it's definitely the PC. If all 5 PC's have been bad in the same exact way I'll be freaked out though. I only have 2 other computers here, both laptops running Win98. I also have a couple touchscreen industrial PC's, but they're running 400Mhz "Caruso" processors and they just plain crash when I try to run Mach. Wish I could set them up with PCanywhere or something to control another PC running Mach (like just use them as a touchscreen display / mouse) because they're only about 1.5" thick and look like framed pictures. I could mount one in the door of the cabinet, which I think was thier intention in the industrial world. (Advantech industrial computers) They are fun to play with Paintbrush though!