Kirk Wallace wrote:
> 
> 
> I used the existing cable which has four shielded twisted pairs. I have
> one pair for each A, B, I and power. 
> 
Are the shields grounded at both ends?
> 
>>also what about an inductor or capacitor at the encoder in its power
>>line to help smooth any noise introduced there?
> 
I wouldn't.  Possibly adding a capacitor across +5 to ground 
might help.
> 
> I have a short run between my encoder and a differential driver. I am
> wondering about whether I should have some sort of filters on the high
> impedance inputs of the driver. The thing is that with 50k pulses at
> 3,000 I'm not sure how much of a filter I could use.
> 

Is the differential driver module grounded to anything, like 
through mounting screws?  Check if the grounds of the encoder 
are grounded to the machine frame.  Since you have had the 
encoder all apart anyway, make sure it is NOT grounded to the 
machine.  If ungrounding it is really hard, then run a huge 
copper braid cable from the machine to the computer/controller 
grounds (not a bad idea anyway).

If you have a scope (not Halscope) look at the index signal all 
the way from the encoder itself, out of the diff driver, into 
the diff receiver and finally out of the diff receiver, to see 
where the noise is getting in.  If it is coming directly out of 
the encoder, possibly a small cap across the power input to the 
encoder will help.  Otherwise, maybe a better ground between the 
encoder and diff driver is needed so it doesn't see common-mode 
noise.

 From previous messages, I think you have 5000 counts/rev, not 50K.
3000 RPM is 50 Rev/sec, x 5K encoder pulses is 250000 counts a 
second.  The quadrature frequency would be 1/4 that, or 62.5 
KHz.  So, a filter with about a 8 us cutoff would be pretty 
good.  The digital filtering I used on the A and B channels 
wasn't applicable to the Z channel since it doesn't have a 
predictable state progression, so it has less filtering, and is 
sampled every microsecond.  But, assuming your Z pulse is at 
least one quadrature count wide, then an 8 us filter might do a 
lot of good.

In your next mesaage, you said :
When I was getting index pulses without an encoder disk installed, I
decided to disconnect the index connections to see when the noise
stopped. The noise stopped as soon as I disconnected the encoder 
sensor.
So either the noise is coming from the sensor, or having the sensor
connected promotes the noise. I am going to have to give this some
thought in order to plan the next move.

This may suggest problems with the encoder, but could be 
common-mode noise on the encoder ground or noisy power to the 
encoder.  A simple R/C filter might be all that is needed. 
especially if the noise pulses are just a 100 ns wide or 
something, the filter should give perfect results.

Jon

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to