[TravisR100]

This is my first CNC project so I'm not sure if this is a common problem or just my fault somehow. Here is my setup:

A small (dyna 2400) mill. I have 269 oz. in. stepper motors driven by a xylotex 4 axis driver. The driver is set to 1/8 microstepping. I have a single start 2.5mm pitch lead screw.

Here is the problem. I go to Mach 3 and go to the MDI screen and manually jog an axis. Specifically the x axis at this point. I have it set to step jogging and have the step size as .001". If I try to jog it one unit (.001") the stepper will attempt to turn and then it "cogs" back to where it was. It is similar in continuous jog mode. I'll make a jog of several inches and when it comes to a stop it "cogs" either forward or back.

It seems like it's not holding the microstep detents and is cogging back to one of the actual steps. Does this sound like what's happening?

In my motor tuning I have it set to 640 steps per. 200 actual steps * 8 for the microstepping * .4 for the threads per mm. I have the native units set up as mm as well.

If any of you need more information to answer please ask!

Is this a normal problem? Should I just abandon microstepping altogether?

[TravisR100]

Also, yes, the motors are directly driving the screw. No, I have not yet checked to see if it exhibits the same behavior if the motor is disconnected from the screw.

[Crevice Reamer]

Micro stepping is only a type of electronic "Hocus Pokus." A stepper motor has no detents for micro stepping. The only detents are for the 200 FULL steps. The motor will always stop on the nearest full step.

I wouldn't worry about it. Since 2.5mm is a full revolution, 2.5/200 steps= only .0125mm deviation. (0.0004921 inch.) Actually LESS than that because the jog stopped somewhere in between.

CR.

[TravisR100]

I was under the impression that with microstepping that the motor was actually held between the detents by the driver. If it always stopped on one of the 200 detents even with microstepping then it wouldn't really be microstepping, would it? Also, this "cog" I'm experiencing makes a difference of up to .001" inches when it does it. I have a dial indicator on the axis. I jog to some arbitrary point, for example I stop right on 0 on the indicator, and then it will "cog" to something between .0002 and .001 on the indicator.

If I put it in step jog mode and set the step size to .001 and click the button to jog the motor tries to turn just a fraction (presumably enough to move the axis by .001) but then cogs back to where it was. So each time I'm clicking the jog the DRO is indicating a move of .001 yet the actual axis isn't moving at all since the motor is "cogging" back to where it was after each move.

[ger21]

The motor will always stop on the nearest full step.

That is incorrect. I've read varying views on this subject over the years. I've read more than one post from Mariss which stated that with Gecko drives, you can reliable move 1 microstep, or 1/10 of a full step.

According to this, each 1/8 step will give you ~20% of your full torque. And if you move 4 1/8 steps, you have 70% of the torque available to get to that position.

When you say you are trying to jog .001", is the DRO set to .0254? (for mm)
That's actually about 16 steps (2 Full Steps), so you should end up just about exactly where you want to be.
What if you try increasing the pulse width? What do you have them set at?

And last, why is the machine set up in mm if you want to move in inches?

[TravisR100]

I set the machines native units in mm because that was what the mach 3 documentation seemed to suggest if that was the unit of the screw. When testing I set the DRO's to read in inches as per the mach setup manual. I've since gone back and set the native units to inches and simply converted everything. So, for my 2.5mm pitch screws I have the "steps per" set to 16256. 16256 = 10.16 revolutions per unit (inch) * 200 motor steps * 8 for the microstepping. Velocity is 50 ipm. Accel is 4. Step pulse is 1.

[TravisR100]

[QUOTE]That is incorrect. I've read varying views on this subject over the years. I've read more than one post from Mariss which stated that with Gecko drives, you can reliable move 1 microstep, or 1/10 of a full step.
[QUOTE]

Either your post is contradictory to what you previously stated or I'm misunderstanding. You said that the motor will always stop on one of the 200 detents. If that's the case then how could you move 1/10 of a full step?

[Crevice Reamer]

Travis: Somehow you are confusing what I said with what Gerry said.

Please check that you actually HAVE the Xylotex set to 1/8 microstep. Also please check that you have the step input in Mach3 set to Active High.

Also check that the Xylotex is set to give your motors the full 2.5A current setting.

CR.

[TravisR100]

Yep, I did confuse the posts. Didn't see there was more than 1 person responding. The Xylotex is definitely set to 1/8 stepping (both jumpers removed). I confirmed this in the Xylotex docs and on their app note "microstepping truth table."

The Xylotex docs say that for 2.5a the ref voltage on the board should measure 3.6v. It also says just below that, "Full current (2.5a) can be achieved with voltages as low as 3.42v on Vref. When attempting to deliver 2.5a/phase start with a Vref voltage of 3.42v." A quick check with my meter shows 3.43v on Vref.

You say to check that I have the step input to active high. I assume that it is since my "step low active" checkbox in ports & pins has a red x.

Crevice, so you contend that the motor will always stop on one of the 200 detents? That just doesn't seem right. If that's the case then what does microstepping do? Again, I thought that microstepping was the ability of the driver to hold the motor between two of those 200 detents.

[Crevice Reamer]

Crevice, so you contend that the motor will always stop on one of the 200 detents? That just doesn't seem right. If that's the case then what does microstepping do? Again, I thought that microstepping was the ability of the driver to hold the motor between two of those 200 detents.

Apparently I was wrong in this assumption and was confusing microstep stopping point with motor stable resting point.

How much backlash do you have in this axis?

CR.

[TravisR100]

The backlash is about .0015". Although I'm not sure what backlash would have to do with it. All these moves are same direction moves. In other words when the axis stops from the move (where I'm getting the cogging) it is in the direction such that the threads would be loaded or engaged.

[ger21]

You say it backs up when it stops. It must be moving quite a few steps in the opposite direction, if you can see it???

[TravisR100]

I'm thinking that it is indeed moving quite a few steps. It's generally between .0002 and .001" that the table moves when it cogs.

[TravisR100]

Just an update. I ordered a new Xylotex board to see if this would fix the problem since the z axis was blown on the previous board anyway. I figured since the z axis was blown maybe something else could be wrong with it. After putting the new board in tonight I still have the same cogging issue. This is really puzzling.

[TravisR100]

Here's another update on this issue. If I haven't stated previously, I'm using a Xylotex 4 axis drive board, a CNC4PC C23 BOB, and a smoothstepper. In finally decided to pull out the little PC scope I got and figure out how to look at the pulses. Attached is a capture. This is captured when jogging the axis .001". As you can see there are 16 pulses. This should be correct since my mach steps per inch set to 16256. What happens is that the motor turns to make the .001" jog then immediately cogs back to its previous position (or sometimes forward). My initial thought was that it always cogged backward because the the table was too stiff. It tried to make the move, couldn't, and the force caused it to cog back. But this isn't the case. The table does indeed make the move when the jog is commanded. Sometimes it stops after the move, sometimes it cogs forward a "notch" and sometimes it cogs back.

From the trace you can see there is no pulse going to the motor from the BOB after the commanded jog.

[ger21]

Is the xylotex plugged directly into the Smoothstepper, or is the breakout board in between them? If your going through the breakout board, try removing it and going direct to the Xylotex. My guess is there's noise somewhere. I've read that the SS is very susceptible to noise on the USB cable.

[TravisR100]

The BOB is between the smoothstepper and the xylotex. I'm measuring the scope traces from the BOB however where it goes directly to the xylotex. So I'm seeing the exact pulses as they are being input into the xylotex.

[TravisR100]

For anyone who's been following this thread, I THINK I may have found the problem. My stepper motors are wired through a panel connector on the back of my cabinet. All of the wiring is good, no issues there. The cable I used for the steppers has 4 conductors plus a shield and drain wire. I used some amp type 5 conductor connectors. On the cable from the steppers I hooked the drain wire up to one of the five pins on the connector. Inside the panel I hooked up that drain wire to a chassis ground. At the motor end neither the shield or the drain wire is connected. As per previous questions I've asked, this is how it should be connected.

After pulling my hair out on this problem I decided to disconnect the drain wire from the chassis ground inside the cabinet. Voila. It APPEARS that this solved the problem.

[Al_The_Man]

To hazard a guess at what really is going on there, It would have to be known as to what parts of the system are grounded, how they are actually grounded and how the ground itself is set up.
It is very unusual to get interference from a drain wire that is grounded in the proper manner, which yours seems to be.
Al.