strange thought about pololu steppers

i was thinking today ( ya i know its a bad habit ) but on my newest stepper board for my pololu stepper drivers . i made it so that i can change all my micro stepping, but is it ok to change them when they are in use.? or should i power them down when i do it.?

but wow were the motors allot quieter when i enabled the micro stepping. that i did notice right off the bat.

if i had the room on the board i would have added some led's so that i could tell when they were getting a step signal.

Edited 1 time(s). Last edit at 10/19/2010 08:19PM by dissidence.

dissidence,
According to the functional block diagram, there is no latching or memory of the microstepping inputs, so the stepper driver should instantly change the microstepping when you change one of the inputs. No need to power down or reset at all. This means that if your main control board can control the 3 microstepping input lines, you can dynamically switch from 16th steps for fine control and as you said, noise free operation, al the way up to full steps for speed, then back down to microsteps again.

And yes, LEDs always give a warm fuzzy when checking a motor driver!

Mike

that is good, i was just not sure if it would hurt it or not. but that is good.

The only thought would be not to change the stepping WHILE stepping. Best bet would be change it when nothing is going on.

if you plan to change while it's running, be aware that you will probably lose your exact position.. not much of an issue for experimenting with the best setting, but an issue if you want to do it automatically during printing

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

i am not planning on doing that but it would be a neat feature to have available though, you would have to home each time you changed if you did it that way,

i also just found out about the optical encoding stripes on some dell printers that got me thinking again. too bad i cant find them yet

I think the problem is that the motor will snap back to one of the "detents" of the stepper. I think that that won't be a problem as long as you only switch at a full step position. the software would need to keep track of all the stepping changes, as well.

Yes, if the software only kept track of full steps, you should never lose a full step. However, if you 8 micros steps into a step change at 16 microstepping level and suddenly jumped all the way up to full stepping, it would not be well defined whether it would roll back to the previous whole step, or roll forward to the next full step.

If the software kept track of steps and microsteps using a 'fixed point' integer arithmetic, namely that the bottom 4 bits (lowest hex digit) was 16th microsteps, and the remaining higher bits are full steps, it would be safe to switch from 16ths to 8ths when the lowest bit is zero, 16ths or 8ths to quarters when the bottom 2 bits are zero, 16, 8, or 4 down to half steps when the bottom three bits are zero, and any substepping level back up to full steps when all four substepping bits are zero. There would be no risk of loss of accuracy going the other direction, from big steps down to little steps. So the easiest way to do that automatically in code is to set a flag to 'down shift' to coarser stepping, and the next time all four microstepping bits are zero make the change then. In terms of system speed, you would never see the delay, and that way it would guarantee no loss of accuracy just due to changing the step size.


Your other comment, using an inkjet printer optical encoding strip, I have just discovered that the cheap, Harbor Freight digital calipers have pins for outputting the current digital position, and also work off the printed millimeter scale. So you can unscrew the encoder body, mount it on anything else the has a millimeter scale printed on it, and get 0.01mm accuracy from the digital position signal. There are websites that talk about 'hacking' the chinese digital calipers and what the interface signals look like. And it should be easy to cut an aluminum track that will fit and print (or print on paper/transparency and past onto the metal) a decent scale and make the optical encoder work for ANY length linear axis. The digital calipers get their accuracy from having 48 tiny optical sensors in 12 banks of 4 each that span 20mm. Since 12 does not divide into 20, some of the sensors are always off, some are always on, and some are right on the edge. By tracking the rise and fall of the ones on the edge, it can get an accurate measurement down to 100th of the line spacing. It saves on trying to get a laser or inkjet printer that can print 2500 dots per inch!

Mike

i think i will look more into this in the future, i think if we could find some way of doing it, i think that we could make this whole project allot more accurate and reliable

One thing it would make a lot faster is homing an axis. Shouldn't make that much difference if it was only used in homing an axis (you don't care much about losing part of a step there), and would prove an ideal test case show the functionality off.

Just make sure it ALWAYS uses this for homing the axis, otherwise it could result in discrepancies.

BTW: If you want to be super accurate homing the axis, step it in 1x steps till the opto triggers, then forward the axis by 1 or 2 steps to clear the endstop, and then step backwards in 1/16th steps till it's activated again.

in my next board i might just try to make it so that i can change the micro stepping with my arduino, there are a few things here worth thinking about.

... implement some jumpers, then you can set them by hand (fixed/predefined) or plug a cable to appropriate output pins of the arduino ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

i am using jumpers 4 2x3 pin headers for it, i need to order more jumpers for it though, i am running a little short.

so ya in that regard i could just make a cable to my arduino if i wanted to go that route.