Servo turning the wrong direction during calibration

I had my machine configured in a bottom feed design and recently switched to top feed. When I attempted to recalibrate I had two issues. First after taking the distance between motors I did not see an option to rotate the left servo back to the vertical position before extending first the left and then the right servo. The other issue I’m experiencing is both servos rotate the same direction when attempting to extend. The left servo appears to be turning the correct direction (ccw). This results in the left chain being considerably longer then the right, and the sled positioned to the upper right of the work area. Any help would be greatly appreciated.

I don’t understand why someone would go that route, usually it’s opposite.
So perhaps clarifying the confusing terminology is step one.

If this was clear, then checking the settings in GC would be the next step before calibration.
If the problem persists, wiping the Eeprom and renaming groundcontrol.ini to start completely fresh has been reported to solve similar issues.

1 Like

@Gero. Thank you for your reply. I am going to work on it more tonight and I will try your suggestions. Can you explain your reasoning for wanting to do the opposite and go to a bottom feed design? I switched because of constant issues with chains wrapping around sprockets, and weak bungees. To me it doesn’t make any sense to fight gravity rather then use it to our advantage, but maybe I am missing some key advantage to bottom feed. I am new to Maslow and any suggestions what others have done for a reliable machine are appreciated!

1 Like

I changed to bottom feed because:

  • replacing bungees with weight. That gives a constant weight on the slack chain. Bungees are weakest when needed and strong when not needed.
  • getting the chains away from the sides gave me the chance to slide my sheet left/right to utilize the more accurate centre of the frame if the project does not need a full sheet.
  • the chains get less chips/dust at the top

My at that time 0$ approach that work perfect is here:

1 Like

With the top feed you are fighting gravity in a different way that’s harder to
compensate for, the gravity is pulling the slack chain to be vertical, while the
sprockets are 15 degrees from vertical. This misalignment is FAR more than
chains are designed to handle and at the very least results in wear as you force
the chain onto the sprocket with some sort of chain guide, and is the primary
cause of the chain skipping.

David Lang

@gero @dlang Thank you for the info!


@Zach_Holt, I switched to the Chain off Top method for the same reasons you mentioned. I understand why some people prefer the other method, but this seemed much cleaner to me. I immediately encountered the chain skipping problems brought up by @dlang - I solved this by adding a wooden spacer below the sprocket to force the vertically hanging chain into the correct angle of the sprocket, as well as using the plastic rollers included (not sure if part of main kit or z-axis kit) to force the chain under and around the sprocket. I have not had any skipping issues since. I also used some scrap pieces of 4x4s as counter weights (also on that $0 budget) - see pics below for my setup.

I DID have issues with the calibration during the portion that feeds the chain out to a known length. I got past this by:

  1. I performed the “Calibrate” function under “Actions” for the first section only - essentially to measure the distance between motors with the chain. I verified this using a tape measure just to make sure the default tooth number settings, etc matched my hardware. I then used the skip button to bypass the rest of the calibration and save the settings to the EPROM.

  2. I set the chain length by using the “Set Chain Length - Automatic” function under “Actions”. This essentially seems to be the same portion skipped during the full calibration, but it took into account the feed direction (set in the initial full calibration) while for some reason the full calibration did not.

  3. I marked the center lines of my backboard (similar to the Ground Control display) and verified that after attaching the sled with the automatically measured chains, it did indeed rest exactly in the center. I painted the chain link that rested on the top tooth of each sprocket so I could easily remove and reattach the chains if need be.

  4. I then went back and performed the full calibration, skipping all of the steps until the test cut. I just wanted to get the test cut measurements in there as I understand they account for chain sag.

  5. I cut out the sled using the github hosted files. I had not attached my z-axis yet so had to manually adjust the router bit depth at each cut. It was a pain but it cut out beautifully.

  6. After attaching everything to my final sled, I attached the z-axis control and calibrated it manually using the “Z-Axis” button on the main Ground Control screen, by the manual directional controls (I had to reverse direction by adding a negative sign to z-axis pitch for my particular router).

That’s it! I have only made 3 cuts since then, but they have all been perfect! I’m using Estlcam and still tweaking the tool settings to speed up the cuts, but accuracy hasn’t been an issue. Let me know if you have any more questions :slight_smile: