PID issues or calibration help

John,

Would you be willing to do a skype session with me. It's like a phone call but with video chat so you can see whatever you want to look at. I can show you whatever you need to see on my screen as it's happening or I can even switch to my phone to show you the physical machine itself. It really sounds like you know your way around these servo drivers so that may help to make things a lot faster with troubleshooting if you're willing to help me out. Even just a few minutes on skype and then once we identify something I can let you go until I get that part tuned. Then we come back to skype after I'm ready for the next step.

If you're not comfortable with that I can work through here. It'll just take 20x longer.

My driver boards are identified by stickers on the back as
30060-33

So hopefully this is what you were looking for? Both X and Y have that same identifier.

Ok, I'll Google those in the morning. I think that they are real close to the generic Baldor, Servo Dynamics, Westamp boards that are common. At least the instructions for tuning all of them are basically the same. In the mean time check all the mechanics and resolver feedback.

John

John,

I tried google but found nothing useful on those boards. Did you turn anything up or do you have a copy (even a pic from your phone) of those pages you spoke of. I am turning up very little on the matter.

Ok, I didn't find anything for your particular board. What I did find was photos of the boards. Most of these analog velocity mode boards from that time period are very similar. Most of the differences are in the output section and in the default gains for the amps. So I have a couple of tuning guides from that era, Baldor and Servo Dynamics that should be real close. If you go by the adjustment name and not the reference designater you should be good to go with the doc you presently have.

I still think after reading your previous posts that your problem is mechanical or in reading the resolvers. Poor tuning of the amps is unlikely to result in asymmetrical physical response. Checking the amps with the motors disconnected and driving them with an external signal (battery) will confirm if they are close to working correctly.

John

OK I think I have an idea and I'd like some help in figuring it out. I found the documentation for my 7i49 board and how the resolvers are wired up. There's 6 pairs of wires coming off of my resolvers. A driver coil pair (excitation coil) and then a sin pair and a cos pair. I am wondering if one of these pairs might be reversed.

I would like to view the 3 signals on my HALSCOPE as I slowly rotate the hand controls. The one video I watched was great for showing me what happens as you slowly rotate the resolver a full rotation.

In looking at the HALMETER and HALSCOPE I don't see signals that would let me watch the wire pairs directly on my 7i49 board. Can someone help me determine what I could watch to emulate the results that this guy from the YouTube video was able to see? Or is the only way to do this with an external o-scope?

Ok, I've never used one of these but you should be able to read the position of the resolver. Look at the Docs.

linuxcnc.org/docs/devel/html/man/man9/hostmot2.9.html

Rotate resolver by hand and you should see the angle smooth increment or decrement depending on direction of rotation.

John

OK Guys I did some resolver testing and more tuning on my WestAMP drivers. They seem to all be in good working order. I even adjusted my TACH settings with a 2v battery signal and I'm happy with the response. I'm very happy with the way the servos operate, sound and behave they are so smooth and solid instant start, instant stop.

But when I hook things into my 7i49 board then everything goes to shit. I'm now starting to realize my issues are all my 7i49 board or my config. It might be my HAL settings or it might be my PWM gen settings. I need help here. My hardware appears to be running as well as I could ever hope.

I have also verified that my signal input wire and my signal common are correct. It further appears that my sin, cos and drv wiring are all correct on my 7i49 board. So this leaves my 7i49 configuration to look at now as the cause for my pain.

Observations a positive signal on the pins for X produces movement of the table to the left and a positive signal on the Y pins produces movement of the table forward (toward me) when I stand in front of the machine.

Does anyone know signal I can monitor to see what the HAL meter/scope view as the signal coming out of AOUT0 and AOUT1 on my 7i49 board? I would like to plot this on HALSCOPE in relation to my f-error and so forth.

So I'm back to one of my very first questions. Is my driver that I copied in the very beginning of this thread what's wrong? Do I have an old out of date SVRM_6 driver? Also if people could look at my HAL configuration. I think that could also be the cause of my pain here.

You can look at the PID output pin to see what LinuxCNC is sending to the analog output.

7I49 FPGA firmware has not changed for more than ~10 years so I doubt if there are any firmware issues.

Can you measure the resolver sine/cosine voltages (with a voltmeter)? you need to rotate the resolver by hand to get the point of peak output

You can check the 7I49 analog outputs by disabling the drives, and unlinking the PWM pins from the PID component in hal. You can then hand set the analog outputs via setp
of the PWM pins.

I did they are only a bit more than 1 volt. The driver signal is about 2v. Reading literature driver should be about 10 shouldn't it?