G203V and resonance

[stevespo]

My DIY build is documented elsewhere in the forums

http://www.cnczone.com/forums/showthread.php?t=27527

For the past 9 months I had been running my Xylotex 3 axis board, 425 oz/in steppers and 24V/5A power supply at 60 IPM with 5 TPI ballscrews. Super reliable, but it was beginning to feel slow on certain jobs. The biggest problem seemed to be mid-band resonance just above 300 RPM.

So, I took the plunge and got (3) G203Vs during the sale and wired everything up. My goal was to have completely reliable 120 IPM rapids. I had a great talk with Mariss on the phone and he felt that 120 IPM was very attainable, even conservative.

Although I could now generate moves from 120-250 IPM, they were NOT reliable. At some point, the motors would stall/seize/choke and movement would stop or grind to a crawl. My understanding was that the Geckos supressed resonance, so my "Tim the Toolman" assumption was that I needed "more power". In comes the 72V/12A toroidal power supply.

With the new BAPS installed, nothing changed - although the motors do get warmer more quickly. Movements are very reliable at 90 IPM, variable at 120 IPM, and unstable at 180 IPM. One of the problems with testing and diagnosing this is that the behavior seems to change from one day to the next. One day I can run the machine hard at 180 IPM, with no problems. The motors are HOT, but steady and strong. The next day (no changes to hardware or software) the machine will max out at 110 IPM while the motors are still cool. This is all in testing mode, no actual cutting.

I've tried a wide range of acceleration values and step pulses, but no noticable effect on behavior. The yellow light on the G203V comes on somewhere around 120 IPM, so is that really where I am stuck?

The motors seem to run more reliably one axis at a time, and have the most problems doing 45 degree diagonal movements. This is where the X & Y are running at (or close to) the same speed. Circles work well. However, as I mentioned, I see these stalls even when moving either (X/Y) axis at lower speeds (600 RPM = 120 IPM).

I could be wrong, but I do not think there are any binding problems. This is all linear rails and with the ballscrews disconnected, even the gantry glides smoothly with the push of a finger. I can drive the screws with my 9.6V cordless at it's lowest torque setting, and it flies at 200 IPM. I do not think there are low-torque or power issues. At 120 IPM, I can't even slow down the movement by leaning hard (200lbs) against the moving axis.

After reading every message I can find related to stalling, I have rearranged my wiring, so I have isolated the step/dir and 4 wire 18G motor wires. I don't feel like it is electrical interference. So (in my mind) that is leading me back to resonance.

My next step is to beef up my X axis motor mount (the Y and Z are firmly bolted to the 1/2" aluminum frame). The X should be firmer, but I have had this problem on all 3 axis. Step B is to try some type of damper on each axis and see if that helps. After that I'm at a loss - new/bigger motors? Maybe toss the steppers and just go to servos? I don't know.

So, before I spend another 3 days (and $$$) banging my head against my router, any words of wisdom? Have I diagnosed this correctly or am I on the wrong track? Is this attainable with my current setup?

Thanks!

Steve

[stevespo]

I should mention a few more things.

The 425 oz/in motors are wired bipolar parallel and rated around 4V/2.8A. I'm running them near/at the 2.8A level with a 33K ohm 1/2 watt resistor. I have not installed the 470 uF/100V cap as that was recommended for amperage over 3A.

The Geckos run very cool, and I even have a small fan blowing on them just in case. I have 18GA wire to the motors, and 12GA from the PSU to the star-configuration at the G203V.

In re-reading my lengthy post - is it possible there is some issue with the parallel port as pulses/sec increase? I am only running an 800 MHz (very clean XP) computer, generating the max 45K/sec pulses with Mach3. Maybe processing power is also an issue here?

So many variables! Thanks for looking.

Steve

[ger21]

I'd guess processing power is very likely the problem. Art recommends a 1Ghz for 25Khz, and 2 Ghz for 35 or 45 Khz. Open your task manager while running a program and check the CPU usage.

[stevespo]

Gerry, thanks for the pointer. Without swapping in a faster computer, I won't know for certain, but I believe you're right. That machine worked so well at low speeds that it fell off my radar until I really started to isolate the problem and narrow things down. With the Xylotex, 90 IPM blended speeds were the best I could do, maybe resonance or corner velocity were non-issues. Perhaps it was CPU all along...

(Mach3 @ 25K pulses/sec)
At idle, CPU is at 8%
At 30 IPM, CPU is at ~80%
At 60 IPM, CPU is at ~90%
At 80 IPM, CPU is at 100%

So, anything over 80 IPM is probably asking for trouble. That's pretty consistent with my testing. Then again, today I dialed my pulse rate back to 25K and I ran my 30 minute "abuse test" at 120 IPM twice (back to back). Not a single hiccup. So better results than yesterday, but maybe unrelated to the config change I made.

I now realize that this thread probably has nothing to do with my new Gecko drivers. I posted here because I felt the problem was mid-band resonance and wanted some confirmation as to whether that could be a problem or not (given that the G203V is designed to nullify the resonance). It would still be handy to know that answer.

Also, before I put the money into a new motherboard/cpu/memory, does Mach3 running the G100 (GRex) via ethernet have the same CPU requirements? In other words, could I use my slower computer with a G100 and get better results? Basically, I'd like a new computer AND a G100, but my budget was shot long ago!

Thanks again, any other ideas are appreciated.

Steve

[ger21]

Since Mach3 now runs at up to 100 Khz, I'd go for the new CPU/Motherboard. Art has said that unless you need specific features that the G100 offers (and Mach3's G100 plugin supports), you should stick to the parallel port.

There's a new USB board being developed that may be available in the next few months, with a rumored target price of ~$150, and capable of the same 4Mhz step rate as the G100 and I think it will have lots of inputs and outputs as well. Supposedly lots of interesting Mach3 developments may occur in the next 6 months or so.

[stevespo]

Ok, new CPU/Motherboard/memory/etc will get ordered and installed in the near future. I was hoping to eventually use my home office 1.6GHz P4, but even that might be too slow for what I'm after. Worth a test, I suppose.

It doesn't quite meet the 2GHz recommendations, but does anyone know if this type of computer can sustain 180-210 IPM rapids, or a pulse frequency of 35,000 pulses/second?

Steve