[Solved] GC 1.11 strange results during calibration

Troubleshooting
21

This is actually a feature in 1.10 :grinning: Although it is expressing itself in a strange way :roll_eyes:

The extend chains step used to automatically center the sled when you pressed the “next” button. This could cause unexpected behavior if the machine had a measurement wrong (like a negative distance between the motors) so we switched to having a dedicated button in that step which will center the sled…I didn’t expect that if you don’t click it it would happen automatically during the adjust the z-axis, but it makes sense.

TLDR: It sounds like everything is working right, but I should keep trying to make it less confusing.

1 Like
22

I still wonder why the chains turned CCW and not CW when he first tried it. @bar, did you try to replicate it by wiping out the ini file and the eeprom (i.e., starting new)?

1 Like
23

I did and I couldn’t get it to happen. My hunch was that the chain over the top/bottom setting wasn’t getting pushed to the machine at the right time, but even with a wiped EEPROM I couldn’t find steps to reproduce the issue

1 Like
24

Do you thinking calling this after making the selections in calibration would help ensure settings do get pushed?

Clock.schedule_once(self.data.pushSettings, 6)

This seems to manually push the settings (it’s from the wipeEEPROM routine).

1 Like
26

It would, but there is an automatic system which pushes the settings every time one gets changed which is working as far as I can tell so we shouldn’t need to

1 Like
27

I should have actually posted in this thread about the issues I had. After reading this I agree, it’s as if GC thinks that I am in top feed but the shield thinks I am in bottom feed, as if a setting didn’t push.

@madgrizzle, you said if you have a top feed machine, after you measure the motor distances, you don’t reset the left chain. That means the motor should move CCW? Is this new? I have always reset the chain and it has always rotated CW.

28

After you measure the motor distance, do you remove the chain and reset it on the left sprocket? This might be the problem. I haven’t used the new calibration recently, but in the old calibration you retract the chain (the left motor turns CCW) using buttons to get it back to the correct position since you fed out a lot of chain. If you are manually pulling the chain off and putting the first tooth on the sprocket, it would do as you describe (drop the chain). @bar, did the new calibration eliminate the rewind of the chain?

Edit: Reviewing the code suggests that in top feed configuration, you don’t ever remove the chain from the left sprocket once you initially set it in order to measure the distance between motors.

2 Likes
29

If that is the case, then my left motor has never moved CCW after measuring the distance between the motors. I measure the distance and then hit Next, the left motor feeds out some slack CW and then takes me to the next screen.

Then I reset the left and right chains and hit, extend chain, for each side. Both motors feed the chain out towards the sled.

It works just the way that the automatic chain length calibration works. Is that not correct?

30

So I’m confused. Is this what you do during calibration?

  1. Choose Kinematics (you pick triangular, I assume)
  2. Vertical Distance Guess (you measure and put something in)
  3. You set both sprockets to 12 o’clock
  4. You put left chain over left sprocket with first link at 12 o’clock and measure distance. After measuring and hitting next, chain slackens a little
  5. You select top-feed
  6. You review measurements
  7. GC “computes” calibration steps
  8. You enter rotation radius guess
  9. You “measure out” chains.
    a) You do this by removing the left chain from the right sprocket, but you also reset it onto the left sprocket (i.e., put first link on sprocket)
    b) You press “Adjust Left Chain” and left motor turns CCW to retract the chain and it falls, correct?

If so, the problem is that you shouldn’t reset the left chain onto the left sprocket in 9a. Leave it where it is on the left sprocket and the chain will “rewind” (i.e., left sprocket turns CCW) to the correct position.

2 Likes
31

That’s correct, those are my steps and I appreciate the quick response. I will no longer remove the chain from the left sprocket. I guess my confusion is more, when did this step change? My left motor, until 1.10 has never spun CCW it has always spooled the chain out, never retracted.

I have probably done that portion of the calibration routine well over 60 times since GC 0.96.

32

I believe in 0.96, there was still a step where you retracted the chain after measuring the motor distance to set the vertical distance between the motors and the top of the workpiece . In order to retract the chain, it had to turn the left motor CCW.

2 Likes
33

I thought that step was always prior to the motor distance measurements.

I am incredibly confused now.

1 Like
34

In early versions of calibration, the first step is to measure the motor distance. After you do that, you retract the left chain until it hangs where the first link is at the top of the workpiece. This calculated the vertical distance between motors and frame. In the recent calibration process, you just enter a value early on (manual measurement) and the Maslow-measured step has been removed.

After you measure the distance between motors, in top-feed, the left chain still has to retract. But when you are in a bottom feed configuration, you have to reset the left chain since you measure the distance between motors in the traditional top feed configuration (you can’t measure distance between motors in bottom feed because the chain would fall off the bottom of the sprocket… gravity is a killer).

2 Likes
35

I must be remembering incorrectly, it’s all blurring together. My apologies.

Still doesn’t explain the odd zaxis behavior which I just assumed was a symptom caused by the same code change.

1 Like
36

Yeah, I don’t know anything about the Z-Axis issue. It is very odd.

2 Likes