Thank you Wildbill
I am installing a new Harmonic drive into the system. If this resolves the change in tool position due to chatter then I may not require a change in the software. The drive is 100mm in diameter and the output shaft is 25mm. That shouldl be a huge increase in rigidity.
I have a lot of parts to make and then testing.
Steve
Could you record the serial output on a test of calibration and post it? There should be some useful messages from the Calibration piece.
Also, is that unexpected tool message from the Acorn?
It would probably be a good if you could attach the code you're using too so we can be sure we're looking at the same thing. I know it's 010c, but as discussed, some of the constants have changed.
The unexpected error message from Acorn was an operator error on my part. That's resoled now.
here is the serial data and a copy of my 10c program
cnclathe010c.ino (16 KB)
data1 .txt (27.2 KB)
I can't see what the problem is. There's something odd in calibration though. I've made a new version that has extra serial prints in it to help with diagnosis. Please run the same test and post the resulting output.
cnclathe010d.ino (15.9 KB)
here is data from a calibrate attempt with 010d program
data2.txt (87.9 KB)
I found it. There's a bug that must have been there a while that breaks calibration for tool one only. I was using unsigned integer variables when signed was necessary.
New version attached. I left all the debugging prints in it for now, they can go once you can confirm it's working.
cnclathe010e.ino (16 KB)
It seems to be working better. It's off by .006 high but now I can set to higher motor accuracy and a slower speed and retest.
I also got acorn to calibrate tool offset correctly.
I will restart testing in the morning. I have to be grandpa daycare now...
thank you wildbill
Steve
Good news, thanks for the update. I'm keen to see how repeatable it is. Testing should be entertaining.
All testing was going very well and I made 8 parts that were holding accuracy of .001 or less.
The best I have done so far with this machine, just fantastic results.... and then I realized by chance that my used harmonic drive I purchased from Hong Kong has a bad spot in its gearing.
I have ordered a New harmonic drive now and I am at a standstill
This is a copy of the settings that I was running prior to finding a problem. The reverse was set to 10 and the adjustments had settings but I removed them because I was going to try this drive again. More tests proved a weak spot in the gears and I stopped.
Thank You Wildbill
Steve
cnclathe010e.ino (16.5 KB)
You were saying that turret speed impacts index. What happens?
If it's momentum, I would expect the turret to move on a bit when it's supposed to stop because there's no deceleration in the code. That should apply to every tool position so it would get worse with each tool since it stops at each one in case the Acorn wants it. Is that what you observe?
The other possibility I can think of is that the opto detector doesn't react quickly enough and the motor steps on before it triggers. Given that the current stepping is once per millisecond, that seems unlikely. What's the spec on the detector - if I ever knew, I've forgotten.
Here's the latest code. It's actually the code you gave me with the new parameters for the harmonic drive, so apart from some tidying of comments it's unchanged. I'm only sending it so we both have the same version with a new version name.
cnclathe010f.ino (16 KB)
New version:
FRAM is gone
Calibration is gone.
Let's see what the telemetry says.
cnclathe012.ino (10.9 KB)
I got this data for you before your last post
Here is the serial data and my 10e program
I think It keeps rotating because the tool positions were not stored in the code they were in the fram?
cnclathe010e.ino (16.5 KB)
new controller data.txt (923 Bytes)
Looks like it got an error code from the FRAM library - no surprise since the hardware is gone, so as we suspected, the tool position data is all broken.
cnc012 is working normal. It see's tool positions
I'll try to calibrate with settings
everything is working as normal
should i start testing more or wait to see if we can index each rotation?
I suggest that you prove out the new hardware before I change the code. Cut some air for a bit and see whether the curse of trying to video it is still there 
So what you would like is:
When Acorn calls for rotation, note tool position, seek index and then drive directly to the next tool position. Then set Gray code to tell the Acorn the tool is up.
Actually, that's not quite it. When the Acorn calls for new movement, seek index and then go through the tool positions as we do today changing Gray codes until Acorn says stop.
the last one sounds correct