All of my cuts seem to be off by a very consistent 1mm on the X and Y axes. I re-initialized the machine with Carbide Motion (I usually use CNC.js as my gcode sender) and I’m getting the same result.
I’ve tested using some very basic squares generated by FreeCAD, and I’ve also tested with hand-sent g-code to the machine.
The amount it’s off doesn’t seem to change with the length of the movement, so that would make me think it’s not the steps/mm calibration.
Anyone have any idea what might be going on here or what I might want to look at?
This is way lower level than toolpath issues. I’m actually sending gcode to the machine directly, and when I measure the result, the machine is off by 1mm.
I’ve had this issue in the past and resolved it with re-calibration. I know you said you don’t think that is the problem, but if it were me, I’d rule it out for sure sooner than later.
With an offset that consistent which doesn’t change with the distance of movement, I would check for slop in the mechanics. Seems like something is allowing 1mm of play so the gantry moves distance minus 1mm.
I’m not seeing any slop in the actual machine. The lead screws and the couplers all seem tight.
It’s also consistent on both the X and Y axis… which I would think would be very unlikely to be slop in the mechanism.
I’m going to dig into the router and the bit a little tomorrow. I don’t think there’s anything weird going on with my endmill or anything like runout in the router… but something like that could also cause this kind of a problem.
I also have a v-bit coming in tomorrow that might be useful for troubleshooting this.
My Carbide Compact Router has a SIGNIFICANT runout in the shaft. It’s enough runout that it’s visible to the naked eye just turning the bit. The 6.35mm endmill is cutting a 7.3mm slot… so that accounts for my constant 1mm of undersize on everything.
Yeah it’s either bad bearings or your collet is slightly crooked. Clean the router side with some iso alcohol and q tips to see if there’s any dust or grime in the collet pocket. After you know that’s clean, you could try a new collet if it’s still wobbling.
I’ve been through 4 carbide routers so far. They’re pretty good about replacing them if you’re within the 1 year mark. I finally swapped over to the Makita router and it’s been solid so far. Also uses the same collets I already had so that’s nice. I think I’ve read somewhere about the bearings being higher quality on the Makita branded ones vs the Carbide one. No idea if that’s true. We will see how long it lasts over the carbide router.
I’d like to switch to the spindle but I can’t wrap my head around spending $300 on the diy kit, let alone $750 on the carbide plug and play kit when I can buy a router for $100 and it does the job just fine for cutting wood and carbon fiber. Accurate within .15mm with the router. I think the spindle could get me slightly better results but not enough to justify the purchase for me.
Cleaned out the collet pocket and it’s still off. I haven’t put an indicator on it or anything, but I think the actual shaft on the router might be off (or its the front bearing).
It’s less than a year old and it’s never been used hard or crashed or anything. I have a ticket open with support, so hopefully I can just get it swapped out.
The whole “off by 1mm” thing didn’t make any sense… at least I’m glad to have found the issue.
I was thinking about trying to upgrade to the spindle but I saw something on the spindle page about incompatibility with GFCI outlets. My shop has GFCI breakers, but the outlets themselves are regular conventional plugs. I am not an electrician, but I would imagine this would have the same issue? Does anyone know what the deal is with GFCI and the VFD?
I would have an electrician come install a non-GFCI circuit for a VFD… but that starts turning into a whole thing. I just want to put my machine back to work.
GFCI breakers provide the same proctection to regular outlets that GFCI outlets provide to themselves & outlets downstream from them. In your case, you’d need to replace a breaker in your shop with a non-GFCI breaker to use with a VFD.
If it were my shop, I’d make all the breakers in the shop standard breakers - and 20A breakers at that even if I had to run new surface-mounted cable to handle the current. Breakers aren’t really hard to change out if you know what you’re doing. But better to get an electrician to come in & do it for you if it’s just not your thing.
The shop is about a year and a half old. It’s a metal building and the panel and all the wiring is super accessible. It’s all metal clad romex. There’s even an empty 20A breaker in the panel (non-GFCI) to be used as a spare. So it probably wouldn’t be very hard (or very expensive) to get a circuit dedicated to the VFD.
Pretty sure they started out with standard breakers and had to swap most of them out for the GFCI kind in order to pass code.
Wonder if Carbide ever has black friday deals on VFDs? Maybe for Christmas this year.
Again, sounds like the number of steps per mm is off…not a big deal if this is the problem.
If this is the problem, you will notice the kind of issue you are describing. A constant / repeatable distance off.
I won’t bring this up again, not trying to offend anyone.
The calibration of the machine itself is dead on. The problem was significant runout in the router spindle. The cutting width on a 6.35mm endmill was actually cutting more like 7.3mm, which resulted in a 1mm undersize on both axes when cutting a square on the machine.
If the calibration of the machine were off, it would result in a proportional amount of error, and the error would be larger with a larger traveled distance. I was getting a constant error regardless of the distance.
