On Wednesday 27 February 2019 12:16:17 Jon Elson wrote:

> On 02/27/2019 07:50 AM, Gene Heskett wrote:
> >   But repeated measurements require I add about
> > 2 counts to the scale value for every repeated measurement done that
> > way, and I've added around 20 counts trying to catch up. In a 16750
> > count, 20 is not that much.
>
> This is not a MATH problem.  I'm assuming open loop steppers
> here, hope that is what you have.

yes.

> If you get a fairly consistent accumulating error every back
> and forth movement, it is likely something to do with timing
> of the step pulses relative to the direction pin.  Make sure
> the space between step pulses is in the right sense, so that
> the direction never changes when the step pulse is "true".
> That is the one that will surely foul up many step drivers,
> especially if they are microstepping.

These are, by 8 I believe.

> Also, check all 
> details of the step timing, pause between step pulses, etc.
> If you don't have detailed specs for your drives, then try
> doubling the step pulse, step space and setup time on the
> direction and see if that helps the problem.

Ok.

> Jon
>
I haven't checked with a scope to see if the pulses are right side up, 
meaning the opto's are only driven for step time, with a large off time 
in between pulses but will. With the 7i76, thats very easily reversible 
as it has both polarities of drive available. 

The hole patterns its cutting are round and it takes 12 passes thru the 
loop to cut one hole, stopping and digging 8 thou deeper per pass around 
the circle. I'd thought of that, which is why I checked them for round 
although the pattern in the spoil board might be questionable, its 
cherry ATM & tends toward leaving a little fuzz. It doesn't cut as clean 
as mahogany.

But I'll double check with a scope yet this afternoon. The drivers are 
the original 2M542's, step and space are 1750ns, setup and holds at 
4000ns. Thats some longer in each case than what I've found for that 
driver.  All 4 are running with that timing. Getting a solid 5 volts on 
the + and the -signal for on time from the 7i76 on its - terminal.

Its crowded up there though as the garage door motor impinges on the 
space on the overhead shelf where everything but the coolant tank 
resides.

I should also comment that a g0x0.000000 always ends up at 0.000xx when I 
do that, then grab the moving caliper and slide it back against the side 
of the tool, with the xx often less than 10.

Far more accurately than this renishaw clone can do. I find in the spec 
sheet its a .005" device.  Thats so discouraging I haven't even mounted 
it. I'd assume one measures the ball, and figures in half the ball 
diameter plus a fudge of 2 to 4 as its true trip point. Piece of junk 
IOW. Not at all concentric as it recommends locking the spindle to use 
it. Moronic waste of money on my part.

Thanks Jon. I'll go fix the missus some lunch and get to it.

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>



_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to