Poor/no response in EMC control vs. PnCConf

I am getting good response on axis motion in PnCConf when I run the Open Loop Servo test. The following settings give good performance:

DAC Fast: 5.0; very speedy
DAC Slow: 1.0; would call it fast creep speed.
Offset: 0

The config page has settings;

I set P=4 to start with and left all other PID values 0.
I set acceleration around 30 and max speed at 30 in/min.
Max DAC output I set at 6 (expecting it to match the speedy in the Open loop)
Dac scale at 0

When I transfer to EMC, I try to move the axis under manual control, and all I get is a very slow creep. I tried moving settings in PncConf up and down a little, but I have been unable to get any significant speed out of the axis in EMC control.

I have spent some time scaling the encoder readings to get very close accuracy using a dial indicator; so I am pretty sure the encoder scale is not causing the problem.

I'm running 90VDC thru the Mesa board, with plenty of amperage available, which should be easy to get performance. (and I think that is vindicated by the Open Loop run)

I also attempted to run 'Test/Tune Axis' in PnCConf to see if it cropped up there, or I could try modifying settings under a loop, but what I get is the message "Servo Tuning is not Finished" but there is nothing left for me to finish! I tried putting small values in the rest of PID columns and tried it again to no avail. Am I missing some step here, or is this a bug?

And a note for the EMC developers;
It would be really great if the Help button had context included with the description. You say what it is, but without a scale or context I have no idea what value to put in, even to start. i.e. - 'for this setting, match your input voltage' etc.

Thanks!
Nick

Also, I set the following error up to 0.5" in order to allow testing without continually getting shut down becuase of the slow response.

Hey Nick

Glad you got your hardware problems behind you.
The open loop test is for doing basic testing of the hardware in particular to make sure the encoder and motor direction match.
The open loop test does not use PID control and thats what you need to tune now.
The numbers you use in the open loop test DO NOT relate to the PID settings.
Unfortunately currently the tuning test in PNCconf does not work with servos thats what the (poorly worded) message meant.
If you start your config with AXIS look under Machine for 'calibration'
This is where you can easily change PID settings and then try them in AXIS.
PID tuning reference:
www.linuxcnc.org/docs/2.4/html/motion_pid_theory.html#r1_1_3
I will let other direct you more on PID tuning
And opening the following error (as you did) is standard for beginning servo tuning.

Yes context with description in help would be better.
I have put more effort in coding pncconf to work properly and keep up with changes then the help.
I'll keep that in mind.

A DAC SCALE of 0 makes no sense (assuming this is actually pwmgen scale)

Please read the hostmot2 manual page to understand what this does
type:
man hostmot2
in a terminal window
to display the hostmot2 manual

I would set the PWM scale to 1 and the PID max output to 1

Peter
If you set the scale to 1 and max output to 1, what does the output represent ?

The manual suggests that the PID output should represent voltage.
then if the servo drive is +-10 volt then max output should be 10 shouldn't it?
I understand that Nick is using a MESA Hbridge so maybe the output represents something else.
I find the manual not as clear as I would like.

1 represents full scale output from the H-Bridge (supply voltage X 1.0 )
so this roughly translates to full motor velocity

its convenient for a +-10v device to have a 10V scale
if you want the set output in volts but I think for simplicity
choosing 1 as full scale for all devices has the advantage
of keeping the PID numbers all scaled the same on
different systems ( well you would need a scale of 25.4 for
metric systems to keep the PID values the same)

and I forgot that actually the PID maximum should
be somewhat less than 1 for the 7I29, perhaps 0.975
for the high side gate drive bootstrap.

Thanks for the feedback! I will make these adjustments and give it a shot again!

Peter - do you mean 'DAC Max = 0.975', and DAC scale = 1.0? I don't have it in front of me, but I don't remember a PID max/PWM scale?

Should I be expecting 'P' adjustment numbers in the 1-20 range based on the scale you suggest?
And what range to expect/experiment for I/D?

Nick

No, theres a PID max and a PWM scale
PID max should be set such that you never get 100 % PWM

Nick in PNCconf it is
DAC output scale = 1 (That ends up being PWM scale)
DAC max output = .975 (ends up being PID max output)

Yes, thanks! That is what I figured and plugged in - looks like that took care of it!

Glad it was not something more excessive.

I started PID tuning as well (using homing function to test it)

Is there some Gcode string which works well for PID tuning run?

I got it to home to 1.0003 or so using P=10 and I=1 on Z axis. (X axis limit/homing switches have mech. issues right now)

What range of PID values should I expect?

Thanks everone! Making great progress, a little more learning and I shoudl be ready to take some test cuts soon!

Nick