Denford triac with Mach3

The following is a post on the Mach group. Some tests as indicated may show your problem.
*****************************
An LED array is handy to have, one sits on my desk since 2001 , and I use it ALLOT! , BUT you wont find it usefull in your case. This is because the pulse width is only 3us. The LED will show nothing unless your running at high speed, then it will glow at a % of average voltage. Single pulses cannot be seen, just a display of a varying brightness showing output speed. ( Still quite usefull in diagnostics.) Steve Stallings sells an excellent stepper test board for such things , it even has a small stepper on it to play with in motion, probably a little expensive for most for testing but invaluable for people who diagnose various computers as to driver function.
There is an easier way to know though, the machine coordinate DRO. Purposely, and mentioned many times in threads, that DRO was designed not to reflect where mach3 thinks it is, there are many variables internally for that, the DRO's programming is very simple, it counts output pulses. The code that actually changes a step pulse increments or decrements that counter. The DRO display is ( #ofsteps / steps/unit ).
So you can just reset the motor tuning to tell it that 1 step is one inch. Retune the speeds so it runs and watch the Machine coordinate DRO. It will advance 1 for each step taken. and decrement for each step in the opposite direction. If it counts 1, then you can be sure a step pulse was put out, no question about it. if it decrements 1, then a step was put out with a reversed DIR line, no question about it. Thinking about it, the DIR prechange time sounds like your trouble, picture what happens, the system tries to reverse, it changes the DIR line, waits 3us , then sets the step line, the driver hasnt yet seen the DIR line change so it outputs a reverse step. Why only in one direction? It may be case of the opto on the DIR line taking awhile ( longer than 3us) to turn on, but having a much faster turn-off time. Kinda like those annoying flourescent bulbs replacing incandecent ones, they take awhiel to get glowing very brightly, but when you flick them off, their instantly off, somewhet the reverse of the incandecent which is fast to glow, but fades out when switched off. Be pretty rare, but if your runnign on the edge of that opto's timing, it could in theory happen. Mach3 holds the DIR line on or off permanently until direction changes, it isnt a pulsed output, so this phenomenon can only happen at time of reversal, and is the only way a slow opto on the DIR input of the driver will be seen. ********************************** Richard
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

I had wondered the same thing, and changed the "dir" settings on the motor tuning to 5 (default is 0) but this didn't seem ot have any effect. Can this value be higher, although it says step and dir values should be between 0 and 5/
Regards Kevin
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mon, 13 Apr 2009 04:20:38 -0700 (PDT), Kevin

No idea, as I said I am not a Mach3 guru <G> Did you try the DRO change as above, and monitored the motor for movement step by step. Other option is inverting the signals.
BTW you do have your belts "tight" don't you <G>
It all seems silly to me. If you believe that long movements are accurate but multiple small movements are not, then it points to acceleration, a resonant problem or even a slack coupling.
BTW you have never indicated your error magnitudes between a long and multiple short movements.
Richard
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

Hi Richard, I found the "Sherline Mode" -but it made things much worse, with much bigger errors. I'm not sure what drivers the Sherline uses, but this mode sets the output pulse to 20 microceconds. I think it is loosing steps at the start or end of a move (or both). But I haven't come up with any explanation other then dropping steps due to not driving the motors at full power (losing stepson acceleration). But I haven't found any settings that it works at, even going down to quite low speeds.
Regards Kevin
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Mon, 13 Apr 2009 04:14:41 -0700 (PDT), Kevin

Greetings Kevin, If the motor loses steps at the beginning of motion then it may be that the step rate is ramping up too fast. The motor cannot go from a stop to full speed instantly. For that matter it cannot stop instantly. So the step rate must ramp up and ramp down in order to avoid losing steps. There must be some kind of ramping adjustment in Mach3. The reason I like using servos with the Gecko 320 drives is because the drives accept step and direction pulses just like a stepper driver. The big advantage is that the drive has feedback from the servo via an encoder so that it doesn't lose steps. If the position command is too different from the actual position the drive shuts down. With a stepper system without feedback the drive doesn't know or care about actual steps completed so your parts come out wrong if steps are lost. Cheers, eric
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Kevin wrote:

I'm currently working to produce a driver that uses the Nat Semi chips, so we can run 48V rather than the 36V limit of the Arc Euro drive. In practice, we have been running the 180Ncm motors at 15mm/sec without a problem, and 20mm/sec sees an occasional missed step. The sample 48V driver gives reliable 20mm/sec operation and may well give a little more, but we need to configure Mach3 for a higher clock rate as quarter stepping is running out of steam at 25khz before we can run some more tests ;) That and we need a set of three boards ....
--
Lester Caine - G8HFL
-----------------------------
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Why not join the Denford forum, they are very knowledgeable and helped me with all sorts of problems with my Cyclone lathe. http://www.denfordata.com/bb/index.php?sid 5ca4d87d28b03298e602bfefac9cc2 Bob
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.