Has anyone used a Quadrature Divider Circuit instead of a G902 Pulse Multiplier ?

The Problem:  1000 line Quadrature Encoder ( 4000 pulses per rev ), 2:1 on the motor pulley, then 5 turns per inch on the lead screw = 40,000 pulses per inch or .000025 inches per pules.

I do not need that kind of granularity and the MAX speed is 100IPM.  So I tried G902 Pulse Multiplier set at the 1:2, works ok.  I now get up 200IPM but the acceleration must be set very slow to keep the G320 from Faulting.  This make since as the G320 will fault after +/- 128 counts of position error and by adding the G902 set to 1:2, I have just cut this in half to 64 counts.  I am also concerned with steps loss after reading the post below:

http://www.machsupport.com/forum/index.php/topic,2749.msg21457.html#msg21457

Possible Solution:  Quadrature Divider Circuit of 2:1 or 4:1.  Would this allow for both faster IPM and Faster Acceleration ?  Has anyone built one and tested it ?  I found these circuits on the Gecko Yahoo Forum, however I can not find any real data of success or failure using something like it.





Thanks,
Joel

Hi,
 I have never had any problems on a metal cutting machine (lower feeds,less chance of tripping up)
using up to 5 times pulse multiplier on a g340 (902 fitted).I would not use 10 times on anything other than a point to point machine.
When you have the modifier in 1x mode it is in effect a g320.
  The Idea of the encoder divider circuit sounds fine I did try and make one once but it didnt really work.
The only thing I would say is that it will make the motors noisey on stand still as they will have further to "hunt"
Plus the system will only be accurate to 4 encoder counts,those four counts could under the right circumstances add up or decrease and lead to position inaccuracies.In my case the 3d mold making can take up to 8 hours to machine.
   To sum up if your not using very high CV feed speeds,and you cant change your encoder, then the puse multiplyer is great.
I always buy g340's over g320's just to have this benifit,once you are aware of the limitations the "problem" can be worked around.
Tony

Tony,

Thanks for the responce.  Good point about "The only thing I would say is that it will make the motors noisey on stand still as they will have further to "hunt""

Did you build a circuit like one of these, or something else ? 

Also, my current resolution is .000025in/pulse.  So a 4:1 circuit would put it at .0001in/pulse, which is still better than the machine was built to back 1986 (.00013in/pulse)

All thought and comments are welcome, the design phase is sometimes the best part....

Thanks,
Joel

Joel,
 yes practically the same,I don't know why it was a bit unstable.
I ended up biting the bullet and using 902's.
 I have always built my systems to have a machine resolution of about 1/10th
of the expected work accuracy.
  Are you sure your power supply can provide enough umph? is it the correct voltage for the motors?
if so your best bet would probably be to mount some 200 count US digital encoders on your motors.