On Sunday, June 03, 2012 08:11:48 PM andy pugh did opine:

> On 3 June 2012 16:39, gene heskett <[email protected]> wrote:
> >  But its duty cycle was
> > about 57 on to 43 off,
> 
> I don't think duty cycle is all that important, unless it drifts far
> enough to break the quadrature.
> 
> If you use analogue-output sensors with comparators to square them up,
> then you can change the threshold voltage to set the duty cycle.

Yes, but because the encoders velocity pin outputs a new sample/hold value 
for every edge that goes by, its velocity output is almost a random value, 
which at best jumps up and down by 10-40% of its average value on a sample 
by sample basis, so both the duty cycle of the signals, and the accuracy of 
the quadrature are in the real world, very important. A very minor tweak to 
the led current can make a huge difference in the peak to peak noise in the 
encoders velocity output.  And that setting does not necessarily coincide 
perfectly with a 50% duty cycle as I found today after I tore up my circuit 
board and incorporated a 1k trimpot in series with each led so that I could 
set it for a 50% duty cycle.

>From the docs on the encoder module:

encoder.<chan>.velocity (float, Out) - Velocity in scaled units per second. 
encoder uses an algorithm that greatly reduces quantization noise as 
compared to simply differentiating the position output. When the magnitude 
of the true velocity is below min-velocity-estimate, the velocity output is 
0

"greatly reduces quantization noise" everytime I read that, I'm like "in 
what universe does this happen?"  It sure isn't happening in this one.

Next I suppose is to get the disk to run a bit truer, it currently has 
about a 7 thou runout, probably because I fit the hole to the spindle a bit 
snuggly, so the side which happens to be sitting over the thread groove is 
possibly being pulled into the valley of the thread at that position by the 
thread peak 180 degrees away.

I can put a dial on the edge, position the spindle to the high point and 
loosen the lock nut, but I can't tap it with a screwdriver handle and move 
it to reduce that runout. I should give it a good enough whack to mark the 
thread on the inside of the hole & sand the mark away, and might yet.  
There is also about 10 thou of wobble runout, but short of selective 
sanding on the bearing nuts to make the faces of the nuts run truer, or 
adding a layer or 4 of alu foil in the right place (I've done that a couple 
other places in this POJ), it simply isn't going to run any truer.

In any event, it is now running noticeably  better in mode 0 than the 39 
cycle disk ever did in mode 1.  Some pid tuning yet to do of course, but 
its fairly stable.  It should cut some pretty high quality threads.  Those 
I've already cut were better than I usually get with std dies if I keep the 
back end of a cutoff tool properly sharpened.

Cheers, Gene
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
My web page: <http://coyoteden.dyndns-free.com:85/gene>
One way to stop a runaway horse is to bet on him.

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to