Can't get Abs Pos to count positive with CW rotation

[kevincnc]

I'll say this one last time. Either your encoder is causing the rpm to be reported wrong; your encoder is good and your VFD can't hold a constant speed; or your axis drive doesn't have enough power to keep up with the rpm fluctuations.

BTW the spindle speed on the screen is stable within 2 RPM.

[frijoli]

250ipm
.25 accell seems really quick. I would reduce the accell drastically, and then reduce the max speed drastically. It won't hurt anything. See what results you get.
The real answer to your question isn't that simple; it depends on a lot of factors that we don't have. What you think you see and what results you get, don't align until you chip away at the variables.

[kevincnc]

Well I'm glad it seems fast, but I'd actually like to speed it with brushless servos. I checked and it's currently 225 IPM max and .3 sec Accel/Decel. I can try slowing it down, but here's what I've done so far-
1) I hooked up another 1024 PPR encoder that I had (replacing the 2000 PPR) as shown below with exactly the same results (working at 300 RPM, not usually at 600.) I haven't seen anything that says what to hook the shield to (I assume the shell,) but it doesn't affect anything either way.

2) I noticed that when it failed (always starting with the Z moving positive,) it wasn't faulting the servo drive. It was still getting the "327 Fault: job cancelled" and the "438 Spindle slave position error". I had forgotten to set the Spindle Encoder Counts in the Wizzard to 4096. After doing that, it faulted the z-drive again. I don't know why it didn't show the actual error 9031 DRIVE FAULT from the drive faulting, but probably since it had already faulted with the 327. I physically disconnected the DRIVE OK input for now to make sure it wasn't triggering the 327, and it's not.
3) Since the wrong encoder setting caused tapping motion to slow down proportionally, I thought I'd try 16000 for the Spindle Encoder counts. I know that would drastically change what the related setting should be, but it doesn't seem to matter. Now a tapping cycle will still run at 300 RPM, just at way too fine a pitch, which makes sense. At 600 RPM, now it gets through an incorrect cycle where it starts in the positive Z, but it doesn't fault at all. This may be meaningless, but I took a video. Maybe it's doing what it tries to do before faulting when the encoder resolution is set correctly.


The clicking that you hear when the spindle is accelerating and decelerating is just the back gear making noise.
So,
1) I think it's pretty unlikely that the same thing is wrong with both encoders
2) I don't think it's noise on the encoder judging from the consistent motion
3) I'm pretty sure my axis is not under-powered

I guess I'll set everything back how it should be and keep trying to find a parameter that's set wrong, but really not sure what to do next. #240 is set at .1 and 241 is 20.

I can't get the report to attach, I'll keep trying.

Kevin

[frijoli]

Well I'm glad it seems fast, but I'd actually like to speed it with brushless servos. I checked and it's currently 225 IPM max and .3 sec Accel/Decel. I can try slowing it down, but here's what I've done so far-
1) I hooked up another 1024 PPR encoder that I had (replacing the 2000 PPR) as shown below with exactly the same results (working at 300 RPM, not usually at 600.) I haven't seen anything that says what to hook the shield to (I assume the shell,) but it doesn't affect anything either way.
1024 PPR_HY VFD.jpg
2) I noticed that when it failed (always starting with the Z moving positive,) it wasn't faulting the servo drive. It was still getting the "327 Fault: job cancelled" and the "438 Spindle slave position error". I had forgotten to set the Spindle Encoder Counts in the Wizzard to 4096. After doing that, it faulted the z-drive again. I don't know why it didn't show the actual error 9031 DRIVE FAULT from the drive faulting, but probably since it had already faulted with the 327. I physically disconnected the DRIVE OK input for now to make sure it wasn't triggering the 327, and it's not.
3) Since the wrong encoder setting caused tapping motion to slow down proportionally, I thought I'd try 16000 for the Spindle Encoder counts. I know that would drastically change what the related setting should be, but it doesn't seem to matter. Now a tapping cycle will still run at 300 RPM, just at way too fine a pitch, which makes sense. At 600 RPM, now it gets through an incorrect cycle where it starts in the positive Z, but it doesn't fault at all. This may be meaningless, but I took a video. Maybe it's doing what it tries to do before faulting when the encoder resolution is set correctly.


The clicking that you hear when the spindle is accelerating and decelerating is just the back gear making noise.
So,
1) I think it's pretty unlikely that the same thing is wrong with both encoders
2) I don't think it's noise on the encoder judging from the consistent motion
3) I'm pretty sure my axis is not under-powered

I guess I'll set everything back how it should be and keep trying to find a parameter that's set wrong, but really not sure what to do next. #240 is set at .1 and 241 is 20.

I can't get the report to attach, I'll keep trying.

Kevin

The testing needs to be analytical. Make big changes, guess what should have happened against what does happen and then guess logically what the next change should be. BUT change one thing at a a time. It's a machine and humans always defeat them.

[kevincnc]

Thanks Clay, definitely one at a time. I've done lots of troubleshooting, and don't remember a machine beating me yet

[martyscncgarage]

You say back gear, is the spindle in back gear?
If so take it out and direct drive it and retest....

[cncsnw]

Try significantly reducing Parameters 68 and 69. Maybe try 100RPM (P68) for 0.5 seconds (P69).

I think that the anticipated deceleration distance (based on programmed RPM, P37, P68, P69 and P82) might be exceeding the depth of the thread, so that CNC12 somehow thinks that spindle deceleration needs to begin to the right of where the thread even begins.

[kevincnc]

You say back gear, is the spindle in back gear?
If so take it out and direct drive it and retest....

Sorry no it's not in back gear, there's just a little slop in the pin that locks to gear to the step pulley on the spindle.

Try significantly reducing Parameters 68 and 69. Maybe try 100RPM (P68) for 0.5 seconds (P69).

I think that the anticipated deceleration distance (based on programmed RPM, P37, P68, P69 and P82) might be exceeding the depth of the thread, so that CNC12 somehow thinks that spindle deceleration needs to begin to the right of where the thread even begins.

I'l try it tonight, thanks.

I have been designing automated machinery for 30 years, and I can't say I fully understand how all the tapping parameters interact. One thing that seems clear in the Tapping Setup Bulletin and the manual is that parameter 240 is in inches, not turns, and 241 is degrees.

I really don't understand how it could work sometimes at 600 but usually not. It seems to me that the calculations done in the controller right when (or before) it sees the index pulse should be exactly the same, and start moving consistently in either + or - Z.

I forgot to mention that the setting P323 to 16 had no noticeable effect.

Kevin

[kevincnc]

Try significantly reducing Parameters 68 and 69. Maybe try 100RPM (P68) for 0.5 seconds (P69).

That worked, thanks again. With the suggested starting values, I'm not sure I would have tried that combination. Now I'm able to see what affect all of the others have. Looking at the parameter graph, I don't see how it could even work at slow RPM like it did. I'm sure there's a reason that 640 RPM and 1.75 sec works on other machines for 68 & 69 though- maybe I'll find that there's a reason to increase them.

With 68 and 69 set at almost zero, the spindle seems to slow down to the the max speed it can be (before starting to advance the tap) to still decelerate in time to be the right depth. I don't see anything that makes it look like it has the ability to do that, so that's a little baffling. Without a brake on the VFD, it seems like It will go about 900 RPM max without slowing down before starting the threading on a 5/8" deep 1/4-20". I haven't had time to mess with it much yet though, and the parameters still don't all make sense to me but I'm sure they will eventually. I still have to see what affect the parameters have at different depths and thread pitch also.

Synchronizing to the spindle speed doesn't seem to be a problem.

Thanks everyone for your help!

Kevin