MMU_CALIBRATE_ENCODER and MMU_CALIBRATE_GATES return different results on the same gate

[iamnotahippy]

Steps to reproduce:

1. Home ERCF
2. Run MMU_CALIBRATE_ENCODER on gate 0. Result is ~400 steps
3. Run MMU_CALIBRATE_GATES GATE=5. Result is ~415 steps
4. Run MMU_CALIBRATE_ENCODER. HH runs encoder calibration on gate 5. Result is ~400 steps.
5. Zeroed calibration values in variables.cfg. Repeat procedure. Same result.

See attached mmu.log

mmu.log.txt

[moggieuk]

Hmmm

22:51:02     TRACE: Setting gear motor rotation distance: 23.427965 (ratio: 1.036667)
22:51:02     TRACE: Running macro: _MMU_ACTION_CHANGED ACTION='Idle' OLD_ACTION='Selecting'
22:51:02   DEBUG: Setting servo to up (filament released) position at angle: 19
22:51:03 Tool T5 enabled
22:51:03     TRACE: Setting gear motor rotation distance: 22.599316 (ratio: 1.000000)

This is entirely expected!.... think about it:

1. You calibrate encoder of gate 0 (the gate you calibrated the gear motor movement on). This establishes the relationship of gear movement to encoder movement.
2. You calibrate gate 5 which finds that it has a very slightly larger BMG gear and sets the gear ratio (relative to gate 0) to 1.036667.
3. You move to gate 5 and ask to calibrate the encoder again. This resets the gear ratio to 1.0 and then correctly determines that there is more encoder movement. This step is invalid though. You must calibrate the encoder once on the same gate (gate 0) that you calibrated the gear.

I.e gate 0 is the designated reference gate.

If instead on step 3 you ran something like MMU_TEST_MOVE MOVE=400 and looked at the encoder reading you would see it, within margin for error, the same 415 counts..

[iamnotahippy]

So the fact that I get the exact same 415 counts on every gate means that my BMG gears are uniform and that the gear ratio is uniformly off by 3.6%. OK, I can wrap my head around that. Thanks for the clarification!