On Jan 30, 2012, at 3:56 PM, Viesturs Lācis wrote:
> 2012/1/30 Peter C. Wallace <p...@mesanet.com>:
>> 
>> I am suggesting this test because  if there is a noise problem, different
>> encoders may not help.
> 
> Of course, I totally agree, but I just did not see a reason for noise by now.
> 
>> A possible noise source is the PWM signals. This can
>> couple into the encoder if the motor frame ground is not returned to system
>> ground
> 
> Motor GND comes from transformator + rectifier bridge, so it is
> isolated from encoder power supply.
> Motor GND is connected to motor frame, so basically - to machine frame.
> I have not yet connected machine frame to earth ground, but I am
> planning on doing so for safety purposes.
> So it means that in that case mains supply and motor supply will not
> be isolated any more.
> 
> Does Your suggestion imply doing this kind of connection or should I
> connect motor GND and encoder GND?


Do you have differential drivers for the encoders or are you running them 
single ended?  For reliability (your two year warranty) I would certainly run 
them with differential drivers.  Also, I am with Peter, I think that you aren't 
going to magically fix the problem with new encoders.  The issue that Jon found 
is very subtle and (may - or may not) help to get the last little bit of error 
out of the system but it isn't going to help with gross movement.  Sounds like 
something else is at fault currently.  

It is so easy to pick up noise even when you think you have taken all 
precautions.  It took me weeks to find out I had two simultaneous issues where 
solving one at time didn't look like it helped, finally, solving them both and 
my issue went away.  I had re-route wires that I thought were out of the way 
even further out of the way of some Gecko stepper drivers AND add a decoupling 
capacitor to my single input to a spindle motor driver.  I would put the 
encoder signals on an oscilloscope and see what they look like to the 
controller.  You may discover a lot of crap that shouldn't be there...

Or, it could be noise introduced into your signal driving the motors (rather 
than the encoders).  That was an issue we ran into on the Plasma gantry table 
and had to design some custom filters to deal with it.

-Tom




------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to