Using Halscope To Track Existing Control Signals (Servo Tuning Aid)

Hello all.

I'm trying to find out, in broad terms, if it is possible to splice in say a Mesa 7i77 board to read the signals from the existing control (for example, the analog voltage and encoder feedback to the servo drives) to help with getting a baseline to tune servos against. The board would be totally passive at that point, serving as a makeshift DAS.

This came to mind while reading through the Brother conversion thread here: forum.linuxcnc.org/30-cnc-machines/32276...29-adventure?start=0. It appears there was some confusion and predictions of what the native Brother controller was doing. If one could capture the signal for a defined move and compare that to the response and control from Linuxcnc, figured it might aid in tuning. This could be taken a step future to track what the machine is doing (for example, with the tool change cycle) to help setup the HAL.

Has anyone done this before?

Thanks,
Joel

A 7I87 (6 channel +-10V 12 bit ADC) would be a better choice
for reading analog voltages

tommylight/PCW,

Thank you both for the response. Knowing that is possible, is there merit to trying to get a baseline control signal for a given move to help with servo tuning, i.e. adjust parameters in Linuxcnc to match the curve as much as possible (as opposed to or at least in addition to the standard servo tuning process)?

I may be oversimplifying this, my thought is to try to capture what the control signal looked like from the old control for say a given short move on an axis and how it handled steady state. My thought with something like the Brother is there were countless hours put into getting the system tuned.

Then with the new (Linuxcnc) control plugged in, you would try to match this same control signal plot while tuning, vs. just tuning as described in your post you referenced. From the original Brother conversion thread, there were some tuning issues, and it seemed that if one could compare the control signal from the original control to what Linuxcnc was outputting, there may be something to glean from there.

All this is with the same servo motors and amps/drives; I do not have experience doing this, so apologies if this is missing a major part. My thought in the end is that the control signal is the final output from the tuning process, so if the control signal could be matched between the new and old controls the performance and behavior should be similar.

tommylight,

Thanks. The reason I'm stuck on the Brother is I'd like to convert a very similar machine, so trying to figure out a way to get ahead of the curve if I go down that route. Your points are well taken and sounds like I may pushing on something that isn't needed, at least not for most applications.

Appreciate the insight.

Thanks,
Joel