Using Maslow 4 to manually calibrate frame

Troubleshooting
1

I can’t use the firmware’s newer than 0.84 until my new Maslow comes in, and the old calibration routine wasn’t working reliably with my frame.

It looks like I should be able to determine lengths of the frame (according to Maslow) using the retract all function, and converting from belt length to frame position. If I measure the two diagonals and 3 sides, I should be able to calculate all the attachment point locations.

Before I break anything I wanted to confirm that the offset shown after the retract all command is simply the length of belt that was retracted. Is that correct?

While I was working through the math, it seems like the Maslow should be able to solve for all the corner coordinates by finding the point along each diagonal that is closest to one of the other corners (more below). I don’t know how much error there would be, but it could at least give the solver a high quality first guess.

If you’re looking for the bottom left to top right diagonal length, the maslow should be able to start with all belts slacked, then tighten the top left belt, top right belt, bottom left belt, and bottom right belt in that order. Then slack the bottom two belts, and loosen the top left by 0.1 MM (or whatever increment makes sense), and tighten the BL and BR again. Then continue until the calculated diagonal starts increasing again. After you have the diagonal length, the maslow could use a similar process to find the point closest to the top left and bottom right anchors. The top left and bottom right belts would form a right triangle, which lets you solve for all the needed attachment points.

I don’t know if it’s worth pursuing at this point (since those procedures are already working), but the calibration results for my Maslow give cuts that are a few percent off, which complicates the build process for fitted components.

2

I created the onshape doc at Onshape to let you enter the measured dimensions and have it calculate the anchor locations.

I think the problem with your approach is that as you are near the diagonal, the differences in belt lengths are going to get very small, and if your frame isn’t exactly rectangular, figuring out when a diagonal is straight is going to be hard. I think the belt stretch error is going to be larger than the distance changes you are trying to measure.

It’s worth trying it.

3

It worked relatively well, but probably didn’t have the accuracy that’s needed. Some of the anchor points were off by a cm (assuming the calibration values work). The router is too loud to run right now, but I’ll do some test measurements when I get a chance. Here’s what the calibration predicted after putting in the measured/calculated anchor points.

|Point|Before calibration|After calibration:|
|Maslow_tlX: |-24.5|-22.7|
|Maslow_tlY: |2373.9|2382.9|
|Maslow_trX: |2303|2294.8|
|Maslow_trY: |2372.8|2377.3|
|Maslow_blX: |0|0|
|Maslow_blY: |0|0|
|Maslow_brX: |2325.4|2328|
|Maslow_brY: |0|0|

I attached the log file also, if anyone’s curious.
Maslow-serial - calibrate after manual measure.log (24.3 KB)

1 Like
4

I have a very square frame that the Maslow is not able to calibrate. Since I drilled holes in my garage floor, I don’t want to drill any more holes for different sizes. I carefully measured and just entered the coordinates of the mounting points relative to (0,0). It cuts very well but it doesn’t jog diagonally.