Step pulse issue??

[ger21]

I'm seeing the same results with a clean profile.

I'm thinking that the issue may be the AM882, and something with the pulses at 200Khz.
I tried various steps/unit settings, and all exhibit the same behavior when the pulse rate is around 133-135Khz.
When UCCNC is set at 200Khz, when the pulse rate gets higher than ~135Khz, the motor starts to spin slower. It appears to accelerate until it reaches ~135Khz, then starts slowing down. The more you increase the pulsae rate, the slower the motor spins.
It's very noticable with a very slow accel setting.

If I set the steps/unit at 12800, and the velocity at 800, the motor accelerates to about 640, then slows down. When you release the jog key, it actually accelerates back up to ~640, before decelerating to a stop.

The AM882 is rated for 200Khz, so I was assuming that I couldn't use the 400Khz rating.

I decided to try the 400Khz to see if I would get the same results.

What I found was that I could duplicate the problem, but it occurred at higher speeds. I was able to get the pulse rate up to about 198Khz, before the problem became apparent.
At 12800 steps/unit, I was able to run up to 930units/min. Any faster, and it started to slow down.

So, to me, it appears that the AM882 doesn't like the pulse rate from UCCNC in 200Khz mode. Or perhaps there's a noise issue with my test setup?
I have a DMM AC servo that I just got repaired, that I'll try to wire up this weekend and do the same tests. That should tell me if the issue is with the AM882.

If it is an AM882 issue, why would it work better with 400Khz mode, than with 200Khz mode?

Now that I know it works fine in 400Khz mode, that's the mode I'll be using. 400Khz is preferable to me anyway, as I'll be using a mix of the AM882 (Z axis only) and DMM AC servos.

But I will try to check with the servo, to see if it is indeed an issue with the drive.

[cncdrive]

Hi Gerry,

Meanwhile my collegue quickly tested with your profile and he said that he still don't see any problem with the signals.
He will test it further a bit later today, he had to go out now...

He also said that what he thinks is probably happening is that the drive's step/dir optocoupler is too slow.
We checked the datasheet of the drive and it really says that it is rated to 200kHz, but who knows if that is true or not, I mean it's a Chinese drive and the paper has no problem to hold any data they print on it.
So, we think that the issue is probably that the opto distorse the signal and so some pulses can't get through and so is why the slowdown,
because above the cut frequency pulses lost and so the frequency goes lower. E.g. if there is a 200kHz signal and if every second pulse lost then that becomes a 100kHz only signal.
So, we think that the speed is proper until the bandwidth limit of the drive's opto is reached and then steps are lost, so the frequency which gets through quickly becomes lower than it was below the frequency limit.

Could you verify if the position is still correct after this issue happens? I mean to check if a position loss happens or not, because if it does then that would make it clear that our theory is correct.

[ger21]

I won't be able to test today, as it's a holiday and I'll be leaving shortly.

But I would think that it has to be losing position, as UCCNC thinks the motor is spinning at full speed, when it's not.

If you read my previous post, I've basically came to a similar conclusion. But it's strange that the drive appears to work up to the rated 200Khz when UCCNC is set to 400Khz, but the drive only works up to 135Khz when UCCNC is set to 200Khz.

[dezsoe]

Hi, Gerry,

Did you try shielded cable for the signals? You could also check the signal quality at different points (BOB output, AM882 before and after opto) if you have an oscilloscope. As the "dark zone" changes when you switch kernel frequency, it can also be some capacitance (of the cable).

[ger21]

No shielded cable, and no scope.
I won't actually be needing such high pulse rates, so this won't really be an issue to me. And now that I know that it will work at 400Khz, there's no issues.

[dezsoe]

OK, it was just a thought.

[ger21]

Appreciated.

[cncdrive]

Hi Gerry,

My collegue fully finished the measurements he wanted to do.
What conclusion he came to is that the step signals frequency are all correct, so there seems to be no problem on the UCCNC side.

Yes, that is strange what you wrote that the motor does not do this overacceleration and then slowdown if the kernel frequency is 400kHz, but it does on 200kHz.
It would be still interesting to see if when you think it works OK on 400kHz high frequency then there is really no lost steps? I would verify that.
Also it would be interesting to see what opto and what circuit around it is in the drive to see how is it possible that it works like that.

[Gary Campbell]

Gerry...
I've been watching this with some interest. I don't think I have an answers for you, but do have some questions. I also have a UC300eth/UB1 and UC400eth I can test with on my bench. Scope too, if needed. On to the questions:

IS there a particular reason that you are using microstepping that high?
Have you tried 1/8 steps?
What motor are you using and what are the required amps (and drive current setting)?
Have you tried increasing the current?
Were these purchased from the Leadshine website (AmericanMotionTech) or eBay?
Have you tried any of the digital products?

[ger21]

I just have the motor and drive mounted to a test board, so I'm not actually using this on a machine yet.

Is there a particular reason that you are using microstepping that high?

No, just playing around with different settings, trying to see if the motors ran smoother with a higher setting.

Have you tried 1/8 steps?

Not since I noticed this issue. If I set it to 1/8 steps, then I don't know if I can spin it fast enough to get into the range where the issue shows up. At the 135Khz range where the issue shows up, that would be ±5000 rpm (I think).
Note that the motor appears to run fine

What motor are you using and what are the required amps (and drive current setting)?

Automation Technologies 5 amp 570oz Nema 23. Drive is set at 4.75 amps.

Have you tried increasing the current?

No, as the motor seems to be running fine.

Were these purchased from the Leadshine website (AmericanMotionTech) or eBay?

Ebay, straight from China. You can't buy AM882's anywhere else. I believe that they are an older model, similar to the EM806. I tried two different drives, from two different sellers, purchased about a year apart. Results were identical.

Have you tried any of the digital products?

These are digital, with specs similar to the EM806's.

If you have time, I'd be curious to hook up your scope to the UB1, and see if you see a difference, when using my settings, with the kernel speed at both 200Khz and 400Khz. Set the steps/unit to 12800, and the velocity to 800-850.


Note that I won't actually be using these settings when I get this setup on an actual machine. It was just something I stumbled across while playing around. Even if I use 1/16 microstepping, my steps/unit will only be 6400, as I'm using 1/2-8 4 start acme screws.