Please Help - How to Configure a 5 Axis Head-Head

made me reread my instructions and realize that the statement below is actually totally false.

Thats okay, I understood what you were getting at. I was never using the touch plate.  I am using an indicator under the lolli to measure z error.  This easier to read than using the touch plate.  I still jogged z back to zero on the indicator and got the position from the DRO.  This just makes it so I don't make a mathematical error. Maybe the touch probe would be more precise idk but.

Seeing that your 'Qz' values are very similar and that the results from your screenshot are very small

It was designed such that the A and B axis intersect, and I am not an awful machinist, so I would expect the values to be very small (though keep in mind my machine is in inches not mm)

Seems quite plausible. Backlash is a big headache and 'backlash compensation' is an ill fitting crutch at best to 'fix' it.

Damn.  Was hoping you'd have some stroke of insight for how I could address this haha.  I def have backlash in my machine and in compensation in INI but doesn't seem to be fixing the issues when way tipped over. 

I have a thought though.  If the error is only happening upon the reverse of direction (ie the angle and z height are accurate until direction changes), I can use the machine in 3+2 mode where I tip over and do my machining.  Then all I have to do is insure the machine returns to angle 0 before making another tip over for a shallower operation. I obviously couldn't use the machine in full 5 axis (at least not at steep angles) but this isn't the end of the world and will save me the hefty cost of a Fusion 360 machining extension license. I'll need to use trigonometry and an indicator to ensure the angle is indeed correct when the machine is way tipped over though. 

Yes. What the script does is.

Okay, I have a question, is SigmaX giving a range it thinks the the value is inside of? I assume there would a confidence interval with this? For example: SigmaX is saying that theres a 90% chance the the given offset value is within 41 thou each side of that value?  In other words, is SigmaX outputting a length value? 

I ask because it seems like the confidence on both 'z pivot' and 'z offset' are both relatively low (high valued number).  Like almost as much as it is suggesting I change the value by.  I believe this high amount of uncertainty has to be coming from the bad ball bearings in the Z axis.  I am in the processing of machining a fly cutter to redeck the fixture plate now (I can't reach to the ends of my fixture plate with my old tooling because of the increased size of the new 5 axis head).  So hopefully will be getting the new z axis on soon. 

The other, highly labor intensive option, is to machine a test piece of whatever I am making, measure where the error is introduced due to the backlash, then go modify to CAM to deal with the error.  

Or remove the backlash.  I have one idea for how this may be done but I am not sure that it is possible.  I will test this idea on the 6th axis as I put that together. 

Here's the build video! -

Congratulations, even if it didn't quite work out all the way. Still an impressive achievement.
Thanks for sharing.

Aciera - absolutely! Still a very useful machine. 

Tommy - thats awesome 

Thanks again aciera and thank you everyone so much!