My open build CNC router has had a problem with circles since I built it. When cutting a circular part, at about the 190 degree position, the cutter moves out about 1 mm continues cutting until about the 270-280 degree position, moves out about 1mm again and continues with the cut. It repeats this behavior until the part is cut out - always at the same points. Does anyone have an idea what might be causing this and some way of testing the machine for a perfect circle.
Post a picture to help us visualize it. I see you use Vectric, is that the software you use all the time?
I do use Vectric. In the two pictures above look closely at the upper left quadrant and you will see the issue.
I would say it's backlash. Check that you don't have a loose pulley (set screw not tight) or something like this. It looks to be the X axis, if that the axis that goes from left to right in your pictures.
I checked the allen screws and they were tight. So I drew 3 concentric circles and ran the CNC again. The attached photo shows the results. The arrows indicate the direction of the cut. I watched the bit carefully and at each point where it varied off centre the axis that was controlling the cut would physically move the bit - in this case on the y axis. Close examination will show the uneven thickness between the cuts. Is there any way to calibrate the x-y movement, or is there something else I should be looking at?
Only something mechanically loose can do that. Set $1=255 (lock motors on) then Turn router off, grip the endmill and give it a wiggle... feel what is loose and follow that (tactile approach) to find the cause of what moves when it shouldn't.
I have checked all the parts for tightness and stability, replaced the allen screws on the x axis motors (perhaps I should do the same for the Y motors. I also checked the current to all step motors, the wiring, and the x-y travel. Then I ran a new set of concentric circles with virtually no load and a very shallow cut. Regrettably, I got the exact same results. This is very disappointing and I am not sure how to proceed. Because the errors are all at the top and bottom of the circles I was looking for a problem on the y axis. The same diagrams that I posted above are still valid and the error has been in the same places, top and bottom for all tests. All suggestions are welcome until I solve this problem.
What you show in the pictures if I'm interpreting the "top" and "bottom right" labelling correctly is appearing as Y-axis backlash- the X is moving consistently, the Y is taking a second to catch up. This is especially visible when you look at the offsets of the grab bites based on direction of travel. While the problem is occurring while motion is primarily in the X-axis, don't think that means that it's inherently an X axis problem. On a 1500mm axis, there may be some hysteresis in the actual motion chain itself with sub-optimal components. Narrow belts stretch, narrow screws wind up. My suspicion is that either you need a 15mm belt that's more significantly tensioned at each end (which could cause issues with your motor bearings over time without a separate spindle assembly for the pulley) or you need to switch to a screw at least 12mm in diameter (and lose some rapid speed). Or look at other options like rack and pinion, etc. As I recall, I don't think we generally recommend standard 5mm belts or 8mm screws over 1000mm around here, even with a lot of tension. You could test this by- carefully, but very firmly- clamping the belts in place at the halfway point on the Y axis and re-running the test (away from where you clamped it). If it's drive hysteresis, the motor has far less material to tesnsion and then release at the inflection point of rotation in order to start stretching the other side- during which time it's exerting insufficient force on the gantry, allowing your bit to wander and bite like you can see- so you should see significantly less "backlash". If this isn't a standard problem on other WorkBee 1510s, you may have your wheel eccentrics set too tight.
Thank's Rob, I will carry out that test this morning and provide some feedback. My workbee 1510 came with 5mm belts. I was looking at upgrading to rack and pinion as a possible solution or screw as a second option- where can I get the parts?
My build will be running 1610 ball screws with rotating nuts on the 1500mm axis, so I'm about as far from racks and belts as it's possible to be. The first place I'd look for anything at this point is AliExpress and probably eBay, I've had great results just importing machine components directly from OEMs, their local distributors, and industrial recyclers (for things like Harmonic Drives). We do have some domestic importers and overstock distributors, like Automation Overstock, Automation Direct, Automation Technologies, etc etc. Just a case of looking through them all and seeing what they have. Just remember with rack and pinion you have helix angles, modules, pitches, pressure angles, etc if you decide to try and mix and match from more than one place.
As much as what Rob is saying is true, keep in mind Rob builds tanks (; not hobby CNCs (; So I wouldn't run out and replace parts in what you are seeing. The repeatible slip when changing direction still indicates clear signs of something loose: If its a belt drive, the most likely being grub screws on the pulleys - or grub screws not on the D-flat of the motor shaft. Under force it slips (but looks tight when you test by hand) - it was a old popular issue since the Ox days, the visual symptom matches the old issue too closely. A drop of locktite (non permanent) inside the pulley, put it back on the shaft, put new grubscrews will probably sort it out. Recheck belt tension (approx 12kg pull with a travel scale) - make sure both sides have same tension
Hahahah. True. But they do have V-Slot in 'em. I actually agree with this. I think 5mm belts on a 1500mm axis are a bad idea in general, but purely based on what you're seeing here specifically, I wouldn't instantly replace them. Figure out exactly where the problem lays, work out the physical capabilities of the machine, THEN upgrade based on empirical knowledge. I meant to mention this before too. Don't assume that just because the motor shafts turn when you wiggle the gantry that the connection between them and the pulleys is super strong. They always look more stable than they are, and the grub screws alone have a lot of wiggle room against that flat- in an ideal world, they'd come with keyways.
I conducted the suggested test and it did not make any noticeable improvement. I then replaced the allen screws on the Y-axis and that improved the circle cutting but it is still not perfect - just less bad. I did notice that there is up to 3mm of play if you push the gantry on the left or right side and I suspect that is the problem. A set of lead screws or rack and pinion would solve that problem so I am going to look at that solution. The question is - where can I get the parts and what changes do I need to make. The preferred solution would be rack and pinion but I am not sure that I can retrofit my workbee 1510 with that solution. Any suggestions?
Finding whats loose will be easier, and might not be related to the drive system (loose wheels etc) - so could very well still be the cause with a new drive system if you dont know what it is
Odd. If it was definitely the belts it should have noticeably improved the results. Like I said, literally half as much belt to stretch/release. Hard to figure out another way to show it since grbl doesn't have backlash compensation. Some parallel sine waves machined in X might show you something, jiggling the Y axis back and forth. Did you superglue/blue Loctite the pulley onto the motor shaft? Improvement from the set screws implies that more of the issue was there than originally thought. That's called racking and it's... Not good. I assume you mean with the motors turned on and locked up? You'll always get a little bit with any gantry with the machine off, it's how well your motion chains control it that matters. That's gonna take some learning and research on your part. Critical to understand what you're doing and why before modifying a machine. In terms of buying, of the places I mentioned earlier, Automation Overstock is probably the most likely to have rack and pinion, but I don't know if it'll be in a usable size. AliExpress has rapidly become my go-to this year. Since a fixed belt system like the Workbee is already halfway to being a rack and pinion machine anyway, that's probably your best bet. The motor's in about the right spot, just figure out how much clearance you need to get a rack in there instead and if it needs shimming to height or not. That not being sure is exactly why I say go and learn, though. What's blindingly obvious to me isn't going to be to you, but you're the one doing the work and solving the problems as they come up. Learn all about rack and pinion, how it works, clearances, backlash, etc etc. By the time you're done, you'll see what needs to be done. Lots of it on plasma machines, might be good to look into as well. In the meantime, tighten up everything on the machine that can be- Loctite the motor shafts, more belt tension, etc- and play with it.
I am on to complete overhaul of the machine today. I will apply all of the suggestions and see if I can correct the issues. I will report back - success or failure!
I did a complete overhaul of the machine. Locktite added to the critical parts. Ran a new circles test and they came out much better - except this time they were slightly oval. I am going to re-adjust the y-axis carriage rollers and see if that makes a difference.
Oval would be recalibrate steps per mm. Tighter belts changed length a little, or old calibration was off because of the backslash
I have re-calibrate the x and y axis and I believe that I have achieved an acceptable circle.Is there a way of increasing the height of the gantry. Mine always sets off the alarm. I need a bit more space between the work piece and table.I did have the bottom supports in a horizontal position but I changed that when I rebuilt the system.
