Hello Guest it is March 29, 2024, 07:42:36 AM

Author Topic: Long encoder cables can you help  (Read 3099 times)

0 Members and 1 Guest are viewing this topic.

Offline comet

*
  •  336 336
    • View Profile
Long encoder cables can you help
« on: February 03, 2018, 09:28:35 AM »
Hi,
  The situation im in is with a retro fit on a customers Wadkin fabricated industrial router.It was fitted with a bosch pmdc system running at 150volts.
I have scrapped all of that system and saved the servos which I have fitted 360ppr chinese single ended encoders on the back (they had tachos on).
  I am using ucnc dg4s 180 drives and there uc 300 motion controller, all of which im happy with.
The ancillary power supplies have there 0 volts commoned to the computer 0 volts and then to machine  ground, but not commoned to the servo main power supply.
The servos have twisted power leads, the encoders have 4core 0.5mm with an overall shield which is grounded only at the drive end.
    Interestingly i can disconnect the grounds etc. and it makes no differance.

The issue :

 Z axis - runs perfectly really smooth encoder/motor leads approx 4 ft long.
 X axis - holds position on start up but if you touch the pulley or give it a command it judders like a crazy thing and creeps forward,uncontrolable and wont stop unless you reset drive. encoder/motor cables approx 12 foot long.
 Y axis - behaves  exactly as X again 10 foot encoder /motor cables.

I have tried swapping drives and motors around with no differance, so I have come to the conclusion its possibly cables, or at least the length of them.
  I have sopken to Balazs at CNC drive and got this answer:
"I think the jumping is caused by electrostatic charge, likely the body of the encoder is statically charged with the motor and maybe even the machine frame"

But im afraid he gave me no solution to it.
  I have read that putting 2 capacitors at the encoder power  end (0.1 +0.001 uf) can help filtering
Also using pull up or pull down resistors on the A B channels can help (will need some clarity here)
Or should I try to totally insulate the encoders from ground.?

I shall be going back on site tommorow to have another go at it, Any thoughts?

Regards
   Tony

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Long encoder cables can you help
« Reply #1 on: February 03, 2018, 11:44:57 AM »
Are these differential encoders - they will have A+/A- & B+/B- outputs?

If so they should be totally isolated anyway AFIK

There are distance limits on wire length but i'm unsure what it is.

The cables MUST be twisted-pair and must be connected so that A+/A- is one twisted pair etc NOT A+/B- etc

Offline comet

*
  •  336 336
    • View Profile
Re: Long encoder cables can you help
« Reply #2 on: February 03, 2018, 11:50:21 AM »
Hi no they are single ended just a+ b+ .
Afik ?

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Long encoder cables can you help
« Reply #3 on: February 03, 2018, 01:31:15 PM »
Hmm, this is possibly useful..
http://www.dynapar.com/uploadedFiles/_Site_Root/Technology/White_Papers/White%20Paper_Encoder%20Wiring%20Made%20Simple.pdf

Also i would not connect the PC ground to ANYTHING outside the PC case, especially machine ground. PC's are usually slightly isolated in the PSU from ground and altering that can cause issues.

Offline comet

*
  •  336 336
    • View Profile
Re: Long encoder cables can you help
« Reply #4 on: February 03, 2018, 02:20:15 PM »
So if you have a single ended encoder how do you use twisted pairs, if you can't twist a&b together? Ultimately I suppose putting on differential encoders plus the associated cabeling is the answer but the expense is prohibitive in this case.

Offline comet

*
  •  336 336
    • View Profile
Re: Long encoder cables can you help
« Reply #5 on: February 03, 2018, 02:22:37 PM »
What makes a cable capacitive? That seems something that you have to be wary of.
Re: Long encoder cables can you help
« Reply #6 on: February 03, 2018, 02:24:16 PM »
Hello Tony;

I think I read your first post carefully... but... are you certain it is not a servo tuning issue?

The following assumes it is NOT a servo tuning issue:

It is possible to get encoders to operate over several meters.  I use CUI encoders (highly recommended), and I also use Differential drivers.  Differential drivers immensely improve the noise immunity of the encoder signal - they almost eliminate any common mode noise.

___
My encoders are easily 4 meters from termination to termination.  I kept my encoder wires in the same cable chain as all power cabling, but pulled them over to the side.  I used stranded Ethernet cable and grounded the cable shield on ONE END ONLY.

There have been replies while i have been typing...
Encoders about 25 USD each
Drivers about 9 USD each

Differential driver CUI-102E-10 "AMT 102 Line Driver Cable"
Incremental Encoder AMT 102-V KIT "Encoder Programmable TTL Radial"

I don't work for or sell these, BTW.

Hope this helps..

Offline comet

*
  •  336 336
    • View Profile
Re: Long encoder cables can you help
« Reply #7 on: February 03, 2018, 03:53:23 PM »
Mad dog, thanks reading your post has made me realize what the problem is ......probably!
I am using a line driver , but I've placed it at the drive end not the encoder end.....doh!!
I'll rewire it tomorrow and see if that improves it.
Cheers all
Tony

Offline comet

*
  •  336 336
    • View Profile
Re: Long encoder cables can you help
« Reply #8 on: February 04, 2018, 11:19:45 AM »
That sorted it , and to my shame it was all there in the manual!
  On the plus side I know more about line drivers than I did.

  Thanks all
Tony
Re: Long encoder cables can you help
« Reply #9 on: February 04, 2018, 12:26:43 PM »
Glad to hear you have it fixed.