Problem with servo controller (G320)

[faststeve]

I've retrofitted my bridgeport mill with 3 MCG ID33004 servo's with 1000 count encoders. Im using Gecko 320s and Mach2 .

The XY drives fault randomly, mostly under heavy cutting or under acceleration. If i slow it down really slowly it will works better. Also they whine most of the time.
I've messed up a lot of parts so far by drives tripping out and the mill free wheeling, also broken cutters.
Anyone have a similar problems ?

[ger21]

It may be due to the high encoder count. Call Gecko and see what they say.

[Bubba]

I had that problem with one set of encoders that I had (not knowing they were differential) and the problem was due to the fact they drew 120 ma.

Geckos are limited to 50 ma draw on the encoders. Mine worked and then faulted out intermitently like you indicate. As soon as I changed encoders, everything was fine (or you can use an external power supply for the encoder-see documentation for proper hookup)

As a side note, I have also run 1000 line encoders with absolutely no problem from that aspect. Mariss has given out some information on how high you can go before expecting problems and in my configurations, the limit has always been in the 3000 line area YMMV

[faststeve]

The encoders are US digital. Current draw is far below 50ma.

Why do you need a lower encoder count ? Less than 1000 lines puts it into stepper territory.

US digital sells encoders up to 4000 lines . Isnt a high count going to give a finer resolution and a smoother cut?

[HillBilly]

Here is a graph showing 1 encoder line. This actualy works out to be 4 quadrature encoder counts at the drive ( So basically you multiply the lines by 4 to get the true resolution.).

The Gecko has a preset fault of +/-128 counts following error. Going with high count encoders makes this fault happen before it would with lower count encoders ( There is always following error when operating.) Also you may hit the frequency limit, If you are getting your pulses from the printer port, before you get to your maximum motor speed.

A 500 line encoder hooked to a 5 TPI ball screw will give 0.0001" resolution.

500 * 4 = 2,000
2000 * 5 = 10,000
1 / 10,000 = 0.0001

Darek

[faststeve]

Yes, but thats encoder resolution of 0.0001" . The servo error can easily be +- 20 encoder counts at best while cutting so the machine accuracy is only 0.002" .
I want accuracy on the mill.
Fanuc servo's have 64000 counts per turn.

[HillBilly]

You will have the same % of following error to encoder pulses with either encoder. The ability to maintain a small following error is dependant on the drive and motor capability more than the resolution of the encoder.

Darek

[faststeve]

For the same gearing and same motor/drive, won't a finer encoder produce a smaller error ?

Drives still not working reliably. Why is there not a dedicated error trip output. Gecko should use a connector with some more pins. Other drives i've read about have true error output to signal the mach2 to stop when an error happens.

[HillBilly]

Originally Posted by faststeve For the same gearing and same motor/drive, won't a finer encoder produce a smaller error ?

Like I stated before the % will be the same. What will change is the PID tuning to achieve the same response. Higher PID #'s with lower count encoders/ Lower PID #'s with higher count encoders. Where your tuning peaks out with both encoders is where you see the ability of the drive/motor to get to and maintain position.

Darek

[Dan Mauch]

1000 line encoders are nice but you may have a problem that the encoder resolution is to fine in that the Gecko only have a 4 bit(128) count error amp.
If you were to change to a 250 or a 500 line encoder your problem should go away.
Dan

Originally Posted by faststeve I've retrofitted my bridgeport mill with 3 MCG ID33004 servo's with 1000 count encoders. Im using Gecko 320s and Mach2 . The XY drives fault randomly, mostly under heavy cutting or under acceleration. If i slow it down really slowly it will works better. Also they whine most of the time. I've messed up a lot of parts so far by drives tripping out and the mill free wheeling, also broken cutters. Anyone have a similar problems ?

[faststeve]

What will change is the PID tuning to achieve the same response. Higher PID #'s with lower count encoders/ Lower PID #'s with higher count encoders.

That doesn't really make any sense. Have you tried hi-res encoders on other servos before ? Why would Fanuc, Teknic, and others recommend hi-count encoders? For increased smoothness of course !

If you were to change to a 250 or a 500 line encoder your problem should go away. Dan

Thanks, but I think i would change the drive before i change the encoder. I want a very fine resolution.

[DennisCNC]

^^ Mach can only do so much at 45kHz.

Originally Posted by faststeve Fanuc servo's have 64000 counts per turn.

You know how much a Fanuc systems costs don't you?