On Tuesday 24 October 2017 08:47:02 John Kasunich wrote:

> On Tue, Oct 24, 2017, at 08:14 AM, andy pugh wrote:
> > On 24 October 2017 at 12:10, Gene Heskett <[email protected]> 
wrote:
> > > Correct, except encoder.nn.velocity is fed thru a last 4 edges
> > > averager to remove some of the noise in the encoders output. That
> > > unforch also affects the spindle to Z phaseing in the G33.1
> > > backout.
> >
> > You might want to filter the velociity, but I see no point in
> > filtering the position, which is what G33.1 uses
> >
> > You could also try using the DPLL timer to take jitter out of your
> > encoder sampling. That's a pretty simple setting with the right
> > firmware (one that ends in "D" for DPLL)
>
> I'm pretty sure Gene's "4 edge average" is dealing with mechanical
> errors in his homebrew spindle encoder rather then sampling jitter.
> The 4 edges mean that it averages over a full quadrature cycle and
> thus deviations from proper 90 degree quadrature phasing are
> masked.
>
>   John Kasunich
>   [email protected]

Thats been my take on it too, John. The encoder wheel is mounted near the 
top of the spindle, on the sleeve the oem encoder for the oem crap 
controllers rpm indicator, and there is nothing there in that design to 
really make it truly concentric, plus at 268 edges, those slots are 
plenty fine, so a 1 thou error from backlash in the machine while 
carving the disk can easily be a 5 to 10 degree quad error. Note helped 
by the OD of the wheel as it has to clear the rest of the stuff under 
the drawbolt. 2.3" IIRC. I note that the average output still contains 
some of the 4 "pole" pattern, just reduced to 1/4 of its former peak to 
peak amplitude. Before the filter, the peak to peak is 50% of the 
average dc output, and that level of noise in the feedback hammers the 
motor around pretty bad since the Pico pwm-servo driver is a full 4 
quadrant driver, taking up the backlash in the gears quite noisily.  
With the filter, its still noticeable at free spinning but quiets quite 
a bit under cutting loads. Add in the very poor quality of the bearings 
in the backgear, and you have a recipe for recommending the wearing of 
30 db rated shooting ear muffs.

Developing the clock for the fifo is done by looking for a diff across 
the first sample/hold, actually a mux2, and when there is a difference 
indicating an edge has passed, then the whole 4 stage fifo is clocked 
one step in the same servo-thread cycle.  That of course is another 
source of timing error because the clocking is done at servo-thread 
intervals.

All rigged for a gain of 1 thru the whole thing.

Thanks John K.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to