Duramax Custom Probe Not Reporting Runnout Correctly

[[Er...]]

Hi Forum,

Story:
I've been measuring a part shaped like an octagon (custom angles) on Duramax with Vast XXT sensor.
The part has ports on each side calling out runnout. The majority of the ports are measured with a star probe which is also used for alignment. The rest is measured using 2 custom different probes, Probe1 with 3 stylus 120° apart, and Probe2 one stylus 40°
With star probe the runnount callout is measuring in tolerance (less than .0010"), with custom probe1 and 2 the measurement is between .0020"-.0040"
Callout is using runnout datum as cylinder and measured as circle. I have removed outter tangential since the results comes wayy worst.
I have verified the position on bench and the part bore to bore is within .0002". Form out bores are less than .0005"

Question:
1-How can I improve the program with custom probes to report the correct value since my error comes when I switch probes?
2- Can another alignment be added to help inspect the part, or other strategy?

Thank you for your Help!!!

[[Ri...]]

First thing I would do is re-qualify all the probes involved.
I had a similar instance this morning where a program using 9 probe builds would report a s hole location off by .005" in one axis.
I re-qualified that probe. Same thing turns out for an unknown reason the primary probe that does the majority of measurements was the culprit.
I re-qualified and now all is back to normal.
I have no idea how a single stylus became corrupted.

[[Ni...]]

Just like Rick mentioned.
Requalify all the probes you are using.
Verify the sigma is below .001mm(.00003937")
Make sure they are tight and the ruby is not damaged.
Make sure it is not shanking during the calibration and while scanning the part.
Look at the form of said circles for both the star probe and the custom probes. They should be the same.
Look at the model of green arrows and red arrows. Magnify them. Do you see a bump in the path in one of them? Does one have more red arrows than the other?

These are all the things I automatically do to troubleshoot something like this. I am sure there is more, but I can't RECALL. get it? lol.

[[Er...]]

Thank you guys for the input!!!
I tried calibrating the probes, and it did improve the runnout a lot.
I'm still struggling with the data as I'm measuring the 2 set of ports in same orientation but 1.00" apart on Z with same strategy and the top comes out ok but bottom is out of tolerance. I have measured the part in a z- position and all measurement come within .0005"
I'm assuming at this point the error is due to the probe, and not strategy.