Wow! Thank you for testing, I actually never looked at the effect on the servo thread, I was more concentrated on velocity and acceleration changes, it is an interesting testing method, there maybe further possible optimizations, possibly some c++ tricks I'm not aware of, but in current state is the best I could come up with, and it is IMO respectably decent, nice to get some validation...

I have edited the original pid.c to include a minoutput pin, and now it works like a charm. There is no windup and zero overshoot on the way down, just like on the way up.

To install it, open a terminal in the folder containing the modified pid.c and run:

Then set the minoutput pin to the minimum ADC value your setup can handle. In my case, this value is zero, since I am using a software pwmgen to generate a 0–10 V signal:

This is the piece of code I have customized:

video showing it functioning in sim and how to test:

load tool 99 probe tool, move probe tip somewhere clear in the work envelope, use inside corners probe page and select a single move probe routine, trigger the probe by clicking the probe graphics probe tip. try jogging to see if sim repsonds correctly.

 

if you still have probe basic installed, would you mind running a test in the simulation? I am running a sim test here and cannot reproduce the error using the on screen probing buttons. i can probe a point in space and immediately jog afterwards using both the keyboard and on screen buttons. so its hard for me to discover why your install is not responding the same. but if you test the sim and the problem does not exist then it would be something in your configuration most likely.

when you say locks out, were you referring to keyboard, on screen buttons, or both?

Yes, the PWM value (from the PID component) should increase as the position
error increases.

Did you check the 5V power?

Are CR23 and CR27 (Yellow LEDs near the center of the card) both fully illuminated?
(these LEDS indicate the analog +16/-12V power supply operation)

The reason I suggest at least 20% extra terminals, is that you will certainly add a pendant, and most likely a 4th axis, but once you lay out your cabinet, it will be full (they always are) so no expension built in means a PITA to add expansion later. Ask me how I know
Cheers,
Mark.

Furthermore, the Mesa disconnects after approximately 5 minutes with these values. The 1A Ff fuse doesn't blow, but the wire deforms. I suspect that due to increased contact resistance, the voltage at the Mesa is no longer sufficient.

When I move the axis further using the LinuxCNC interface, the value keeps increasing, as if it were the value of the distance traveled... for example, -77... Is that supposed to happen? I'm also only getting 0.35V at the other PWM outputs, both with and without the power amplifier connected.

I will definitly check that, when the machine is back together. At the moment i cant power on the machine because most of the wires are disconnected.

There is some good information in there, thanks.

my1987toyota wrote:

Ok it is officially January 1 2026 in my area and fusion 360 is still working on my Windows 10 machine.
Will it stay that way through out 2026? Don't know. 

Post from Autodesk forum:Microsoft will end mainstream support for Windows 10 on October 14, 2025. As outlined in our official Autodesk announcement, Fusion will still run on Windows 10 after this date, but:

• Windows 10 will no longer be a supported or validated platform for future Fusion releases.
• Autodesk will not guarantee compatibility, bug fixes, or product support for Fusion on Windows 10 after October 14, 2025.
• Fusion will not support Windows 10 LTS builds beyond this date, as Fusion follows Microsoft’s General Availability Channel.

So F360 will run on W10 for now, but my interpretation is that if there are issues due to F360 updates Autodesk isn't going to fix them.  Eventually F360 will stop working completely with W10, just like it did with W7 & W8 a few years ago.

As I mentioned early in this thread, I've moved to Win11 LTSC in the last month and am happy with it.  No AI, no telemetry, no spyware, no junk, no suprise 'feature' updates.  Can load it with zero MS add-on's.  So far it's been clean and fast.

Not an apologist for MS or Windows, but for those who must remain on Windows like myself, the Enterprise LTSC version seems to be the least objectionable option.

36 bit IN TRUE hm2_7i97.0.pwmgen.00.enable <== x-enable
36 float IN -2.5 hm2_7i97.0.pwmgen.00.value <== x-output
36 s32 RW 1 hm2_7i97.0.pwmgen.00.output-type
36 float RW 10 hm2_7i97.0.pwmgen.00.scale

With these settings you would expect about -2.5V output of analog out 0

Possible reasons to not get about -2.5V

Looking at the wrong pins

output overloaded

5V 7I97T supply low

Damaged 7I97T