I have a BlackBox X32 that I have been trying to calibrate over the last year+ so it will accurately cut a circle. I have calibrated multiple times and can never get both the x and y axis to dial in accurately. I have been told my cnc is not compatible with this controller but others with the same machine have success but I cannot achieve my needs. I haven't had a project that has required a perfect circle for quite a while, but now I do. I concocted a test that proves my X it out by at least a mm. When I had calibrated my axis I had calibrated out to 900mm. It appeared to be cutting accurately but obviously it hasn't. I am on the verge of replacing it with something else. Because I need an accurate controller. Any thoughts before I bail would be greatly appreciated.
calibration = once off correction for manufacturing tolerances. Doing it over and over again, is hitting the wrong track This is mechanical. Check grubscrews on shaft couplers, check stop collars, check nutblocks, check machine is tram and true.
I just posted on this issue at Ask Grbl CNC questions here. The controller itself it proven accurate and reliable, Gcode is definitive, a 1" move is a 1" move to within 1 microstep (usualy 0.0004") . Anything else is mechanical.
Okay okay.... I'll concede. However I am convinced there is nothing loose. and it is trammed perfect. But I will tear it apart this weekend and verify everything is tight and trammed. But when I come back on Monday after doing all this work and it's still not cutting a perfect circle what will you tell me then? I'm sorry it just seems like an easy out. Oh it's mechanical and you wash your hands of it. Peter I may not agree with you but that doesn't make me right. I will check Mechanical. David I like your post. Some of that makes sense. I have ball screws so can cyclic error still apply. The science makes sense. Makes more sense then my machine is loose... I built that thing from the ground up. It's tight. It's square. It's trammed. It's perfect. I made sure of that. But I will give Peter the benefit of the doubt because something could have come loose over the year I have been using it. Thanks Guys.
Provided you did get calibration accurate (within couple steps of theoretical value) then it really is still mechanical. Keep looking Sadly a fact i cannot change for you.
What comforts me in that entire statement Peter is your whole hearted belief that the BB can accurately cut a circle. I just have to get my **** together. More to come.
I get a problem like this quite regularly on my machine with one axis cutting the wrong size, despite using ball-screws that should never need calibration for a cnc router level of accuracy (and I don't use a black box). The first time it happened it also drove me nuts and I stripped the machine, ground flats on all the ball-screw shafts to stop grub screws slipping etc. but the problem did not go away. I found the error in the size of circles was more when cutting hard materials like oak or aluminium, and when I cut really soft foam, the error was almost unmeasurable. After much swearing and messing with different toolpaths and trying to calibrate something that should not need calibration, I noticed the bolts holding my waste-board down had been slowly working loose with the machine vibration. As my waste-board is cut to sit snugly left/right between the two Y-axis legs (which have remained well bolted thankfully), the waste-board only moved in the forward-backward direction, giving a very similar appearance to what you are seeing with the apparent scaling on one axis being off. Even though I use nyloc nuts on the bolts to try to counter vibration, the MDF of the waste-board seems to compress a little over time, and the bolts start slipping a little again. I have taken to giving the nuts a tweak every few weeks to counter the problem before it starts and am always surprised that I can get a bit more of a turn on the nuts each time (I will eventually re-drill new holes in my main machine bed to be a super snug fit on the bolts too!). Evan
I am curious as this could be similar to a problem I am having. when you calibrate your axes, you do it at 900mm. but once you have it calibrated at 900mm did you check to see if it would accurately travel 100mm or 1800mm keeping within your table limits of course.
Okay... So I spent an hour checking couplers, stop collars, connections, grub screws and a partridge in a pear tree AND I am trammed within .002 of inch. EVERYTHING is tight and where it should be. I even marked the coupler and the shaft on both ends and nothing moves. We are tight. So then I ran my x axis to right. Jogged it 10mm and then jogged it 10mm to right and made my mark. I jogged it 700mm to the right and made my other mark. Measured it... It is perfect. Right on the 700mm tic. So I checked my y axis. Did the same process and ran it out 700mm... On the nats ***... What the hell is going on???? Why will this thing not cut a perfect circle? I mean you see the pix. It's off like 2 or even 3mm's. What I haven't done is cut another circle but I'm going to bet the results will be the same. I'm at a complete loss right now. It has to be in my ball screws or something.
you are aware that a diode laser dot is not square, right ? (-: ok, not 2 or 3mm out, but still, not square, the diode is actually a rectangle https://www.edmundoptics.com/knowledge-center/application-notes/optics/laser-beam-shaping-overview/ so, just in case, lets have a look at your Gcode, please upload it, there is a slim chance your drawing is not round.
I'm puzzled by the jpeg on the left (above). If the disk is sitting in the aperture from where it was milled, (and it appears to be inline with the grain), then I can understand the gaps left-ish and right-ish from the where the tool has cut, but the "tight" fit top and bottom doesn't make sense.
Do a test again..circle diamond square with caliper measurements. Pics of the machine too, lets see if something looks sus. If you hold the tool and give it a wiggle is it solidly firm or some give/flex
It looks like you have cut the test pieces from oak which is hard; try machining something super soft like foam as if it is a flex problem somewhere, then the cutting forces on foam are so much lower, there is likely to be less flexing. Also repeat the test in oak, but with the board grain turned through 90 degrees to see if the same axis has the error in it; if there is flex in the Z-axis, the grain of hardwoods like oak can cause deflection differently depending on whether you are cutting across or with the grain (a test on plywood is usually more consistent as grain has less impact). Looking at how your circle is tight with the grain top/bottom, but about right when the grain is placed sideways, it is almost as if the bit cut a wider slot when running with the grain than when cutting across, which could be your Z-axis flexing in different ways depending on how the machine is vibrating. Also try cutting the circles close to one corner of the bed, such as near the home position, and then repeat it right in the centre of the workspace to see if there is any difference (cut in the oak or plywood to make sure there is plenty of cutting force). The corner/centre test is a good one to catch if the Z assembly is applying a torque and twisting/flexing the cross gantry (usually the X), as gantries are usually less susceptible to torsional twist at the ends. The other common flex-point is the vertical end supports, such as in an XYYZ machine, where the two ends of the X gantry drop down onto the Y actuators. As others have said, try the simple test first where you grab hold of the collet and put some force on it by hand to see if you can move it, but repeat the test in the centre of the machine and corners too to see if there is any difference. The tests Peter suggested with diamonds and squares are also really good to try as they can also reveal problems with other issues like strange 'wobbles' in the sides due to resonant flex as the machine accelerates and decelerates too. Evan
LOL... Yes David I'm well aware that it's a rectangle. I hadn't considered my CAD could be off. I'm going to check my circles in Vectric when I get home.
Peter I had been focusing on the x and y axis and not fussing with the z. I just went out and literally grabbed the gantry and yanked on it. Yes there is some slight movement in the z. Can't tell if it's enough to be 3mm or 4mm but there is movement. Won't be able to tear it down till this weekend if the honey do list isn't too big. Will keep you guys posted.
Tore gantry apart and found most bearing block screws to be anywhere from a quarter turn to almost completely out kind of loose. Tightened everything and cut a new test. I think I need a good tramming and I should be good to go... I suspect... (Peter)... When building my machine the instructions had told me to leave my bearing block screws loose one quarter turn until I got them on the rails. Once I put them on the rails I missed the step that had me tighten everything. So ever since then the vibrations have just made matters worse and making it impossible to cut a perfect circle. So Peter I owe you an apology for my snarking comment "But when I come back on Monday after doing all this work and it's still not cutting a perfect circle what will you tell me then?" You were right and I was wrong and I will gladly eat crow. Here's a pic of the new test. It's so tight I had to turn it slightly to get it in. I'm guessing because it now needs to be re-trammed.
Just glad you found it. That sounds like quite a hunt! Well done No apologies needed! Frustrating one that was for sure, deep in not obvious i get it
