April 16, 2017 at 9:17 am #31576
I do see it happening, so far no issues though.
This is a tough one for a rigid coupler because of the high chance of one side lowering before the other. Damage is sure to happen. This needs some ability to bend, so either a flex/hinge mount at the nut or trap the spring coupler a little better with two bearings. I am not sure.
Or the new anti backlash nuts for these rods might add enough friction to help hold it up by itself then we can switch to rigid couplers.
I am messing with it for the mpcnc’s update.April 16, 2017 at 10:09 am #31586
As soon as I start my machine, the motors are powered, so as long as they start at the same height, and they never skip steps, they will stay “level”. I am nervous to give it a big piece of plywood to see if it’s an issue. Maybe there’s something I can do to force them to me right when I start it. Like twist the screws up to remove any slack before powering up.
I got much better results with the mpcnc (pre 525) when I pulled the x/y tight to square it up while plugging it in.
I should also test drilling to 3mm or so in several places in the spoil board and see what the variance is. This table is much flatter than my mpcnc was.April 16, 2017 at 10:27 am #31590
I see what you are saying, like a preload from the start. I have never looked at that, but makes sense. The stuff I have done was a cut all the way through so if the edges were at a half a degree angle it would not have mattered either. That could be checked before you start with a jog to the other side and see if it is still at the surface (or whatever your zero point is). Check the z zero nearest you, jog to the furthest side of your material check for the z zero, adjust by turning the leadscrew if necessary.
Interesting, I have not looked at this, but it could be a bit of an issue.April 22, 2017 at 7:47 pm #31995
Do you guys use limit switches with the lowrider?April 22, 2017 at 7:59 pm #31997April 23, 2017 at 3:29 pm #32038
So, I’m having a problem. Sorry for the crummy picture.
I was skipping steps the first time I tried this gcode, and I upped the stepper output and tried again in roughly the same place. The first cut only made it 2.5mm deep, so that’s why there are all those mess ups on the top.
The really issue that I can’t figure out, is the leftmost side in this image. It has those wavy cuts. It was cutting a pocket, then going around the outside, then cutting the next depth pocket. So on that left side, it had already done the pocket and it was moving from close to far every pass. It’s hard to tell from the perspective, but the waves were getting bigger not smaller.
Any ideas on what’s causing it? The other side (counter clockwise) was perfect. The left side of the right side of the pocket also had the waves, but to a smaller amount.
Bit was 1/8″ downcut, 2x. Not brand new.
Linear speed was 15mm/s. Depth of cut was 2.5mm/pass. Material is plywood.
You can see the cart in the image, but in case it’s not obvious, toward and away from the camera (I call that y) is on the roller wheels. Left and right (I call that x) is on the long rails.
This is a plate for my table saw, and it’s possible it wouldn’t have fit with a perfect cut, but it seems like it’s not removing enough material, not that it’s removing too much.
Attachments:April 23, 2017 at 3:33 pm #32042
Extreme closeup (same image, just less easier to focus on what I’m talking about). The waves on the left of this gap.
Attachments:April 23, 2017 at 4:07 pm #32047
I can think of anything obvious, maybe a bent bit?April 23, 2017 at 4:37 pm #32048
Huh, actually, it’s not cw vs CCW. It’s doing that wavyness when going from close to far, but not far to close… This makes no sense.April 23, 2017 at 4:57 pm #32056
Is there any way something is hitting the belt as it’s cutting? That could explain the pattern.April 23, 2017 at 5:29 pm #32062
I measured the period, and it’s the same as the period from the pocketing. I think the pocket made the pattern and the clean up passes were either in the wrong place, or the machine was flexing and not cutting smoothly on that part.
So that explains the pattern.
Hmmm. I could do smaller depth cuts. I could replace the bit. I could change the pocketing scheme in EstlCAM. I could just recut it without the initial screw up. Lots of variables.April 23, 2017 at 5:34 pm #32063
Here is the .e10 and .gcode, if anyone is interested. I think I’ll peek at the gcode a little, to see if it’s somehow a CAM issue not lining up the pocketing with the trimming. Then I’ll probably just cut it again and see if it’s reproducible.
Attachments:April 23, 2017 at 6:36 pm #32067
Huh. I made a 10mmx100mm box, pocket, linear 15mm/s, 2.5mm/pass. I was expecting to see the same waves. I didn’t, but I thought it was ok when I noticed the right side wasn’t lined up later to layer… So I put in a new bit. Similar, but it was the same behavior, but a little less.
The Gap is wider than 10mm, so I think it’s skipping steps at some point. Weird that it’s skipping a small amount, exactly the same on each layer. I am guessing it’s during the plunge…
I am thinking both these problems fall into “Jeff needs to learn to CAM”.April 23, 2017 at 6:43 pm #32076
Or your sides aren’t perpendicular? Something like this // or maybe a little of \\ that?April 23, 2017 at 7:12 pm #32079April 23, 2017 at 7:17 pm #32083
Here’s a video if you like. This is the filler bit cut. I can’t see the trouble here.April 23, 2017 at 7:25 pm #32085
That was with fresh gcode, or did you just cut out that part. That is hard to explain unless it is a CAM glitch?April 23, 2017 at 7:32 pm #32087
Your gcode looks okay, I usually tell it to put a speed on every line but I don’t think that is an issue.
I am pretty confident you have an angle on your machine somewhere. From your camera angle something looks to be angled high on the left. If you cross section it is the bottom of the cut flat, or angled?April 23, 2017 at 8:19 pm #32095
The stairs on the right of the cut is new gcode.
The machine and bottom of the cuts are level. I’m not great with a camera, I guess.April 23, 2017 at 8:23 pm #32096
Colorado’s just a 20 hour drive. You could be here by breakfast. Err. Dinner.
Thanks for looking at the gcode. Here is the box cutout code. Dead simple though.
The big cam mistake I’m making I think is plunging. I wasn’t cutting 2.5mm/pass before, and I don’t think I have that plunge angle right. Plus I’m using a downcut bit. So if it skips a step after plunging every time, then that would explain the stairs. Skipped steps have to be whole steps, right? You can’t skip a 1/32 step, so that might explain the consistency.April 29, 2017 at 7:25 am #32373
So, when EstlCAM uses the default pocketing (linear), it doesn’t respect the plunge angle, so it just drives straight down. With my downcut bit, and the 2.5mm cut depth, I think that’s causing a lot of stress. I’m changing it to parallel, and I can see in the preview that it uses a spiral cutting pattern to drive down, which I think will work better. I’m guessing this will also fix the wavy pattern, but I haven’t figured out why yet 🙂 .April 29, 2017 at 8:38 am #32378
Huh, this is making even less sense.
I switched to parallel pocketing, and I accidentally did 1mm/step, and it still is skipping steps on the right side…April 29, 2017 at 8:41 am #32379
Something fishy going on. Is the driver power low for that axis?April 29, 2017 at 9:10 am #32380
Just trying to push the head around, it’s really not giving up easily, I don’t think it’s the stepper giving up.
I’m going to try to cut in foam next, but I have two new theories. Just theories, but if there’s a good way someone can think of to test them, I’m all ears.
1) As the Z is moving down/up, maybe it’s also moving left/right. I think the bit is perpendicular to the spoil board, but if both the left and right sides were skewed to one side, then each layer in Z would be offset in X.
2) Maybe the wheels have some hysteresis. Possibly moving more left after a clockwise rotation. Something like that. The right side is flush against the table now, and I’ve seen that a couple of times, but it just now clicked that it might be the issue.
I’m going to test this same thing in foam, to eliminate all my speeds and strain. If it still has the steps then it’s something like one of these, or it’s maybe something in grbl, but I haven’t had issues before with it.April 29, 2017 at 9:11 am #32381April 29, 2017 at 9:12 am #32382
The slope of that pipe on the right doesn’t match the slope of the steps at all, so I’m thinking it’s not likely #1.April 29, 2017 at 9:31 am #32383
Those z rails really need to be perpendicular to your work surface. That could cause your steps.April 29, 2017 at 9:38 am #32384
I think it’s #2, and I think it’s some goofiness in the carriages.
I measured the squareness of the pipes to the z carriage part, and they are very square, but the carriages themselves aren’t doing so well. All these pictures were taken without rotating the camera.
You can see that the back wheel mounts are to the right, and the front ones are to the left.
I am also sure the table (under the spoil board) is square, and before turning on the motors, I’m making sure that each carriage is flush to the front of that table.
I think I just need to nudge them a little in the right direction. I’m not sure why they are tracking differently front to back, but my guess is that this is just happening over a lot of movements (and probably exaggerated with some missed steps, or mistakes I made earlier). Maybe I should make some kind of bracket to adjust the square before I start the machine…April 29, 2017 at 9:40 am #32385
Those z rails really need to be perpendicular to your work surface. That could cause your steps.
It looks like I’m arguing, because of the order of the posts, but I saw this after I responded. I agree with you. I think the goofy tracking is what caused them to me misaligned.April 29, 2017 at 9:41 am #32386
I was able to nudge them back to aligned. The bottom wheels touch the bottom of the table, but I can still turn them. I’m not sure if that’s a good or a bad thing. Maybe they should be tighter?
I also have been driving “off” the front, where the thing is only supported by the top wheels. That might be a problem. I can fill that area out with more material, but I was hoping it wasn’t necessary.
You must be logged in to reply to this topic.