[JC50]

I built a DIY CNC router (gantry type) a couple of years ago. Built from MDF, 1" and 3/4" gas pipe, aluminum runners, skate bearings, etc. I purchased a Keling bundle of a Gecko 540, PSU, and 3 370oz motors (about $500 total). Got it working somewhat.
My problem was always repeatability. Meaning, I would make the first cuts, then when my G-code would go back and make the second and third deeper cuts, when it should have gone back to the same spot to start making the deeper cuts, it would not. When I was making only one shallow cut, the letters or whatever I was cutting would not be symmetrical. It wouldn't be off much, but enough to mess up my work. The computer I was using was a 2.4 GHz Intel system with 1GB memory and I also tried several different video cards (never used the built in video on the motherboard). I tried the trial Mach3 with WinXP and Linux/EMC, and always had high latency (approx 22K ns or whatever it measured in, if I remember correctly). After asking questions on this forum, we all decided I was losing steps. Not wanting to invest in another computer, I finally gave up and dismantled the whole system.

Now, I'm ready to try again. So my questions. Was I correct in that the high latency was causing the motor step loss, which in turn was causing my repeatability problems.....and what are some suggestions for a decent computer and operating system to run the system. I'm willing to spend a small amount on a computer, but I'm just scared whatever I end up with will have the same latency problem. As the topic title says, I need some advice.

P.S. I still have the original computer system stored in my attic just waiting to go back into action if anyone can offer some advice on how to lower the latency.

[hemsworthlad]

To be honest there's too meny variables to give any advice at this time without being connected to a machine.

I would recommend you build the new machine and see how it performs this time round.
If the parallel port was giving 5v and not 3.3v then I doubt latency was the problem and there's so meny other problems which could have caused it.

Carve one had a similiar problem and if I remember mariss fremarnis found the problem and it was something like the A+ A- & B+ B- wires on one of the motors was wrong way round causeing the pulse's to never reach full width before next pulse came along, effectivley causing missing steps.? Check C1 build.!

Good luck.

[CarveOne]

To be honest there's too meny variables to give any advice at this time without being connected to a machine.

I would recommend you build the new machine and see how it performs this time round.
If the parallel port was giving 5v and not 3.3v then I doubt latency was the problem and there's so meny other problems which could have caused it.

Carve one had a similiar problem and if I remember mariss fremarnis found the problem and it was something like the A+ A- & B+ B- wires on one of the motors was wrong way round causeing the pulse's to never reach full width before next pulse came along, effectivley causing missing steps.? Check C1 build.!

Good luck.

Mariss asked me to check the DIR signals at the two X axis G203V drive modules (X and A) to verify that the voltages were 0v (< 0.3vdc) in one direction and 5v (> 4vdc) in the other direction. They were not meeting those conditions and the voltages were abnormal for TTL logic. I found that there were a couple of check boxes in the Mach3 Output Pins tab that were incorrect. Once those were corrected the step-over problem I could not figure out for months permanently disappeared.

I don't remember which settings they were now, but I'll see if I can find the info I posted in my Work Table CNC Machine log.

CarveOne

[CarveOne]

Here is what I posted in #1924 or so:

"I found the reason the A axis Dir terminal was not going to ~5vdc. Mach3 Config/Output pins settings were wrong. Outputs #2 thru #6 were disabled. Outputs #7 and #8 were enabled. Don't know why they were enabled. I disabled #7 and #8, saved, exited Mach3, loaded Mach3, and checked the Dir voltage again and now it is 4.85 vdc when it was 0.67vdc. A axis matches X axis voltages now."

There have been no problems with step-over due to missed steps from erroneous DIR voltages since that post.

CarveOne

[hemsworthlad]

Mariss asked me to check the DIR signals at the two X axis G203V drive modules (X and A) to verify that the voltages were 0v (< 0.3vdc) in one direction and 5v (> 4vdc) in the other direction. They were not meeting those conditions and the voltages were abnormal for TTL logic. I found that there were a couple of check boxes in the Mach3 Output Pins tab that were incorrect. Once those were corrected the step-over problem I could not figure out for months permanently disappeared.

I don't remember which settings they were now, but I'll see if I can find the info I posted in my Work Table CNC Machine log.

CarveOne

Ok C1 seems like the only bit I remembered correct was Mariss helped out and the problem was fixed. . Either way I knew you were a very happy relieved.:banana: . . Lol

[JC50]

Since I'm retired, I guess I have nothing better to do than to rebuild and try again. I'm thinking of going with a smaller machine and using 8020 aluminum. Since it's been 2 years since I built my first one, it's almost like starting all over again. So I guess it's time for the research to begin.

Again, thanks for the help. I'm sure I'll be asking more question, so stay tuned.

[CarveOne]

Ok C1 seems like the only bit I remembered correct was Mariss helped out and the problem was fixed. . Either way I knew you were a very happy relieved.:banana: . . Lol

Seriously, that situation almost killed my interest in CNC machines. Somehow, Active Low was enabled and I don't think that is a default setting in a new installation of Mach3.

CarveOne