Cut 2 curvy during calibration

I built a new, slightly smaller and lighter sled for my machine. Installed it, and ran through the calibration the first time. During the test cut, everything stopped between the 2nd cut and 3rd cut. No errors, messages or anything.

I shut everything down, the computer, maslow, everything. Powered everything back on, ran the calibration again, but now, the 2nd cut is “curvy”. (I stop it after the 2nd cut, so I don’t know what cut 3-4 look like) It looks like it’s lowering the Z a little too soon, and once it starts cutting, the curves.

Here’s the part that really worries me… I smell a faint “electrical” burn smell… At first I thought it might be the Z axis motor, since I kept hearing a squeaky noise when adjusting the Z - initially, I thought it was just the screw drive/rails, which I lubed (seemed to help, but didn’t go away)

But now I’m worried it’s the shield. Looks like the one I have (two processors with stick-on heatsinks) has a tendency to burn up. I’ve yet to get on a ladder and put my nose up to the controller.

Does this sound like it could be the controller?

can you get us a picture of the cut?

David Lang

1 Like

I went through, checking everything was plugged in securely. I opened up the Z motor to try to figure out what the squeaky noise was, and it looked like one of the “magnetic sensors” (if that’s even the correct term for it) was rubbing against the circular magnet on the motor. I pulled it back just a hair so it wouldn’t rub anymore. I went through the calibration process again, this time to get a photo/video of what it was doing… and guess what? It cut perfectly… I’m about to go back to the garage and cut out a couple projects, see how it goes - fingers crossed.

1 Like

Nice sleuthing, I will keep my fingers crossed also :crossed_fingers:

Okay, so, it happened again… this time, I took pics.

I was cutting a job that took up (nearly) the entire sheet of plywood. After maybe an hour of cutting (about half way through) USB lost connection… this happens occasionally (not nearly as often since I re-routed the router power so it doesn’t run along the Z axis cable and put ferrite rings on each end of every data/power cable) - So, I start the calibration again.

First, third and fourth cuts are perfect (and I mean the physical cut order, not how they’re numbered on the calibration screen) Second cut was curvy, again.

The first attempt of calibrating, the 2nd cut was really bad. So, I shut everything down, again. Restarted, wiped the memory, restored defaults and started from “scratch”. This time, I took a photo and video of the 2nd cut, which was better than the first attempt, but still off.

It looks like the Z is lowering before it reaches the beginning of the cut, and I think this might have something to do with it. (though, not sure why) Every other cut, the sled moves into position, stops, then lowers Z.

I completed the calibration, and ran through it one more time - skipping everything until the cut test. This time, everything cut perfectly.

I create a new job to cut the missing parts of the original job, and start cutting. It completes the cut, but the corner cuts in the far bottom left corner and mid-way up the right side was way off. There were 3 other shapes, of approximate size that it cut perfectly - like, dead-on.

Here is the cut in the bottom left corner. As you can see, it didn’t follow the same path between the first and second pass:

And here is the cut on the right edge of the board, about half way up:

12’ top bar, approx 2’ above the work area. I’ve cut in these areas before, just fine… only thing I’m wondering is, why the z axis starts to move before the sled reaches its position - this is new, haven’t observed that before. And maybe a separate but possible issue - this sheet of plywood was exceptionally rough (Lowes) and really scared up the bottom of my sled - going to have to sand it down smooth again - could that have anything to do with it - and if so, why just those two particular corners (both passes) and not the other 18 corners on the same job?

Anyway… if anyone has suggestions, ideas, I’m all ears. Thanks!

Do you ring carriages bump into your z-axis brackets? Its not clear in the video, but those big blue things on the side of the router seem to protude rather close to the ring. If they interfere then the cutting will be off.

The video certainly does show that… that’s odd. What version of ground control and firmware are you running?

Even with a wide beam, its still tough to cut along the extremities… there’s just not a lot of force to pull the motors into position. This, coupled with a rough stock, might be the cause.

Really weird that the z-axis lowers before finishing the cut. Check that ‘buffering gcode’ is not enabled in settings and see if that helps (probably under advanced settings).

Yes, the z axis clears the ring by about 1/8" of an inch. It moves freely without coming into contact with the ring.

Firmware and gc are both 1.26. I thought about re-flashing the firmware to see if that would make any difference. shrug

I’m thinking the corner cuts that were off in the example above probably had to do with the roughness of the plywood. Like I said, it really scarred up the bottom of my sled (I use layers of enamel spray paint to give it a “smooth as glass” surface) I’ve made a couple more cuts since then and they were fine, so I dunno.

I noticed too, sometimes when I adjust the Z prior to a cut, the initial Z movement causes the right motor to move, just a bit. Wondering if there is some sort of crossover (for lack of a better word)? If I remember correctly, I think the right motor and the Z share the same processor? Anyway, I ordered the new shield yesterday, the one with individual processors for each motor. It’ll be here Saturday, hopefully I’ll have time to get it hooked up and run a few test cuts.

1 Like

I think this is a software thing, moving one motor (z-for example) will power up all three causing them to position hold which can make them twitch a little bit

Edit: Although I am curious to see if switching shields changes the behavior

Got the new shield Friday. Tinkered with it this weekend (had to solder the aux pins on because I do use the auto spindle on/off and touch z) I was nervous about that because I’m pretty terrible when it comes to soldering. Didn’t have any spare pins around so I took the ones off an old motherboard (floppy disk controller) Got everything hooked up - re-flashed the firmware (doubt it was necessary but I was going to do it anyway as a troubleshooting step) - went through the calibration, and everything seemed to go well.

A couple things I noticed, first of all, the issue I observed with the old shield, where it would move Z prior to getting in position to make the 2nd calibration cut isn’t happening anymore - it moves to position, moves Z, then moves to cut - like it should.

The motors sound different… not sure how to explain it, but they sound “angrier” … lol. More pronounced sound. I believe the new shield is able to provide more power to the motors, so maybe that’s it.

When I did the step to measure the distance of the motors (pull tight and measure) after it pulled tight, and released, there was still a bit of tension on the chain - with the old shield, there was quite a bit of slack, easier to take off the sprocket.

I cut my first job with it last night and was really nervous. I never cut a job that big (not all at one time, I usually break them up into smaller jobs) and was expecting something to go wrong - but it didn’t. Cut the whole thing perfectly - I’m very happy. Still have a couple more sheets to cut that I’ll try to knock out this afternoon. Fingers crossed it still continues to behave. I think it will.

When I’m done with this tree, I’m going to go back and try to re-cut the job with the corner cuts in the far bottom left of the sheet of plywood and see if it cuts all wobbly, or not.


That is fantastic news! Thanks for the update. I’ll be keeping my fingers crossed as well, but it sounds like that shield is a huge improvement across the board (pun intended)

I’ve had nothing but good luck with the new shield. I estimate I’ve made 4-5 hours of cuts since installing the new shied… zero problems. I’ve made several cuts down in the extreme bottom left, and bottom right corners of the plywood - so far, they’ve all been dang near perfect (definitely good enough for my needs)

Don’t know if the shield has anything to do with it, but I’ve also had no USB communication lost errors. I usually get them from time to time, but so far, not with this shield.

1 Like

In your post from Dec 6th with the first video you describe the z axis moving when it wasn’t at the point it should have moved. Is the fact that things are working better now partly because it is cutting mdf now and not plywood? I was cutting some plywood Saturday and saw a similar thing with the z axis and it was the sled rocking away from the board a little because the long chain went slack and so the bit pulled out of the cut groove. The picture of the lower left plywood in that same post looks similar to what I saw. I think it matters which direction the router is moving in the corner. If moving vertically into the corner or sideways into the corner. I find my system has issues going into the corner from the side, but no problem vertically and it will lift sometimes down there so the z distance isn’t right and things get weird. I had a cut generated from a couple different layers that overlapped. One layer went down into the corner from the right and the others approached from above. the first one was messed up and the other cuts had no problem, but the corner was jacked from the first cut. If I want the cut perfect I need to “help” it and keep the chains tight in the corners -a small thing (defeating the purpose of a cnc) but effective to not waste wood in the corners. I also wonder if my bit is dull or it is feeding too fast. Your fix was the z axis motor “adjustment”, the shield board and a power cable routing change?

this is too much frictionbetween the sled and the wood being cut. sand the
bottom of your sled really smooth and was it and it will slide easier and be les
likely to have this sort of problem

kDavid Lang