Help Debugging Early Lowrider Cuts?

LowRider CNC Troubleshooting
21

My comment about purchasing from you wasn’t meant to rude, though re-reading it, that’s what it kinda looks like. Sorry about that! I just meant that the details to my hardware is exactly what you sell. And don’t worry, I don’t think you’re a punk, I think you’re a dude who designs kick-ass machines and helps people use them out your own kindness, and for that I am eternally grateful.

I will try exactly the following steps, because you’re right, I’m not keeping to the instructions as listed and that’s on me. I will do the following:

  1. With power off, move gantry to hard-left
  2. With power off, adjust lead screws to square the carriage
  3. Power On
  4. Lift carriage
  5. Move to the right beyond the job area to observe that there are no snares
  6. Move back to the left, to a suitable start position
  7. Lower Z until it touches the material
  8. Power on the Dewalt 611
  9. Load test.gcode
I guess I don't understand the purpose of driving beyond the job, even though I'm totally willing to do it. Is it just an observational thing, or is it actually physically re-aligning something?
22

Are you saying you skipped steps?

The tracks shouldn’t let the gantry move near or far. Maybe they need to be closer?

That can only happen if you’re skipping steps. So that’s not good. It could skip if it’s binding up because it’s racking, or maybe because the leadscrews isn’t straight. Are the z pipes touching the side of the table?

I have no idea how that logo got cut like that. That’s just very confusing to me.

23

Just wanted to make sure neither of us were getting too frustrated, I genuinely want to help.

Driving past the work area makes sure there are no snags or points that might stall it, it could also realign things but hopefully you are already square and that is not the case. If you were to start crooked instead of square this is where you would find out, it would sinch up on one side or the other, which to me seems like what happens to you.

When doing this, personally, I have found power cord snags, clamps I have hit, vac hose facing the wrong way, one time the tension wheel was too tight for some reason (probably dirt under a wheel), All things that would have ruined my job. I run the the lowrider sets in 2 hour batches. Taking an extra minute or two has saved my butt almost every time I have done it, no joke, I use the table as a workbench at times and you never know.

Or if you are not wired in series you might not have enough power.

Or you just have your two side pieces touching the table too much, I try to have only my wheels touching and not use the Z rails as bumpers.

24

Live log as I work:

  1. The assembly is hard-left - both roller assemblies are touching the edge.
  2. Measured both Z-heights are equal (73.74mm from the top of the Z-rail-holder to the plastic below)
  3. Powered On
  4. Raised the carriage by 50mm in the Z axis
  5. Moved the carrige by 850mm in the Y(roller) axis, beyond the workpiece
  6. Moved the carriage back to 0 in the Y(roller) axis, and measured 22mm of difference on the far roller, while the near roller returned to the same place
  7. Moved the carriage to a suitable position, at Y0 (Y0 for the near side, Y22 for the far side), with Z against the foam
  8. Powered on Dewalt 611
  9. Loaded test.gcode
  10. Measured Z-height difference periodically (though I had to move quickly to get the same Z), and things look matching, less than a millimeter difference)
  11. Measured the Y difference periodically, and found the 22mm disagreement each time (as well as I could measure while the thing was moving)
  12. After the job finished, the carriage fell into the foam, so I turned the 611 off
  13. Raised the carriage 50mm Z
  14. Remeasured the Z difference, and found it to be 24mm higher on the far side. The carriage is nearly center
 

The result of all that: It worked! The cuts all look solid, and the dimensions are spot on. Photos are in the album.

So now I’m trying to figure out what the heck is going on that causes this set of steps to work. Is it that moving across the job and back “irons” out all the snags, and gets the disagreement in the carriage before the job? If so, it would mean that my table isn’t square enough for the machine.

None of the Z tubing is touching the table, but on the far side it is visually less parallel than the near side.

Should we consider this resolved? Should I do more to identify the root issue? Puzzled, but at least I’ll sleep well tonight with one good cut :slight_smile:

25

Okay, I do have new firmware I need to release to solve the dropped when the file is finished. I keep forgetting, I flashed mine a while ago and never put it on the site. I made a note, trying to get them out now.

Can’t really say for sure what was the issue, just stick with those steps for a while, if not always. After you get a few more large cuts under your belt you might find what the problem was. When it all isn’t so new you might spot it, right now you have to much going on to see some tiny detail.

Try a few more in a few different places on the bed to make sure.

26

What! So you’re skipping 22mm worth of steps on the far roller?

This doesn’t surprise me as much, because it was powered off. If it was off by this much while being driven, then that’s a ton of steps lost. And even if the gantry I’d centered, 1" over 5 feet isn’t that huge of an angle.

Do you agree it’s skipping steps? I wouldn’t cut any wood until I got it to work without skipping steps in foam. That’s my $0.02.

The way I’m understanding it, it has some internal strain that is causing it to skip while moving (in foam) until that stress is low enough to be overcome by the current limit in the motors. If you add more strain by cutting in wood, or a plunge cut in the wrong direction, or anything, it will skip more steps, which will then lead to stair stepping in your work.

27

Is that what that means one axis was 22mm shorter on a 10mm cut job?

Are your Z steppers wired is series or parrallel?

I edited/uploaded 8 new firmware zips…that sucked.

28
What! So you’re skipping 22mm worth of steps on the far roller?
Maybe? Here's my process for measuring the Y distance:
  1. Place the butt end of my calipers against the belt anchor
  2. Scroll the calipers until the butt-measure (I don't know the words for this) hits the side of the MDF
  3. Do the same on the other side and compare
The near-side assembly was measuring 12mm, which is just the thickness of the plastic. The far side was measuring at 34mm. So, a 22mm difference.

I don’t think it’s missing steps, because after moving it across the job and back it was 22mm, and without adjusting anything, it remained at 22mm after the job was done.

My theory is that the table itself is geometrically at fault; that the most square position for the assembly is for it to be at that angle off from the edge. I’ll have to test on the other side of the table and see if that 22mm remains a constant.

I’ve got a friend who shares the shop space with me, who’s much more of a carpenter than I am (He’s the one pictured building the table). I’ll get his guidance in the next couple days about the properties of the table.

Until then, I like Ryan’s idea: Keep running the machine, doing increasingly interesting (and large) things, until it either screws up again, or we figure it out. I’ll need to buy more aluminum if I’m going to do work beyond 4’ anyway.

I’m also not too concerned about Z. I’m thinking it’s got a little less lubricant on the screw. Regardless, once it’s aligned, it stays aligned.

29

How are your steppers wired series or parallel?

If your table axis is 22mm different that is the problem, they should be the same, to a very fine degree, like within 1mm. If you are wired in series this is almost impossible to happen unless you start crooked. If you are in parallel, you need to change to series. This is why I have asked the same question so many times.

 

30

At risk of sounding dumb, I don’t actually know - I’m using the cables provided in the kit. That makes them look like they’re in series, but I don’t really understand how those wires are set up.

31

I thought you were saying that they started at the same distance, you drove Y 850mm (powered on) and then you drove it back -850mm. Then the right side was 22mm from where it started.

32
I edited/uploaded 8 new firmware zips….that sucked.
Hooray! No more spindle bashing!
33
I thought you were saying that they started at the same distance, you drove Y 850mm (powered on) and then you drove it back -850mm. Then the right side was 22mm from where it started.
That is what I'm saying, I think.
  1. At the beginning of the maneuver, both roller assemblies are up against the belt anchors, and are measuring 12mm with my calipers.
  2. Power on the machine, and move +850mm along the rollers
  3. Move -850mm along the rollers, and expect both assemblies to be touching the anchors again
  4. The near-side roller is touching the anchor, the far-side roller is not. After measurement, the far side roller is now 34mm away, instead of the 12mm it had been.
  5. That's 22mm of failure to travel back to the edge, which could also be considered 22mm of disagreement between the near and far rollers.
I'll try to take a video of the process tonight, as I'm having trouble explaining it.
34

If I am understanding you correctly, you have a stepper issue. Take off the belt and make sure the stepper is actually moving. If you are using my wiring harness it is in series and has to be moving. You say the “cables provided in the kit” is that my kit or some other?

If it is my kit you are in series and both motors have to turn, or both will stall, you should not see a distance discrepancy unless it is physically being held or getting stuck. this should be very obvious. you can see and hear it happen.

If it is from somewhere else it could just be skipping steps.

What control board, what drivers, what are they set at. I ask these questions again to try and speed up this process. If you bought it from me what board do you have, if not what board do you have and what is it set at.

We really could speed this up with the answers to these questions that I keep asking.

35
“cables provided in the kit” is that my kit or some other?
Sorry, your kit, so series.
What control board, what drivers, what are they set at. I ask these questions again to try and speed up this process. If you bought it from me what board do you have, if not what board do you have and what is it set at.
The RAMPS1.4 board, as purchased from you, with whatever FW + drivers + etc you had installed on it. I purchased it July 24th, if that helps.

I’ll poke around when I get home and keep you posted.

36

Yeah, this is key. When I don’t have my hearing protection in, and I am not running my router, I can easily tell when it skips steps. I think it sounds like crunching gears. You might also be able to feel it if you put your hand on the cart.

Thinking aloud (well, in text). When it does skip steps, in series, I would think that the two sides could skip by a different amount. If the far side was jamming, the current needed exceeded the limit, then both steppers would turn freely at the same time, and the one with the jam would move more.

37

We have liftoff!

Sorry for the hiatus from the forums. I wanted to take some time to really investigate, because it seemed less like one big issue, and more like many small ones. And that turned out to be true! Here are the things I found that contributed to the machine working better. Hopefully this helps someone else in need:

Table Squareness

Certain corners of the table were square, but not all of them. The table was actually a rhombus by a handful of millimeters. We took some cargo straps and tightened them until the shape was all square (disassembling as needed), then installed some cross-supports under the table to keep it that shape.

Rail height

The 2x4’s that we used for Rails ended up curving a bit, which meant a slight angle on the edges. We forced them straight and re-screwed them to hold at a consistent height.

Set Screws!

This one is probably the killer. I was demonstrating some of the issues to a friend, and he nudged the roller while the motors were on. I expected it to resist, but it rolled freely, which was a red-flag. Turns out the set-screw on the far-side roller’s X motor had backed itself off over time. I think what’s been happening is that a lowering friction with the set screw meant that the motor would turn 1 revolution, but only turn the pulley part of that. I tried to tighten it up, and found that that specific set-screw didn’t like to be tightened into that pulley. I swapped the set-screw with the Y-motor’s on the carriage, and both pulleys are now satisfied.

Corner Plastic

Originally, I didn’t know how to assemble the Corner-Z and Corner-X pieces together, and as a result, I unevenly overtightened some of the corners. This resulted in plastic warps and cracks where the screws went through. Those cracks loosened the pressure on the tubing, making it more sloppy. One of the zip-tie holes for the X-belt also gave out. Additionally, the warps were blocking some parts of the assembly from lowering all the way. I re-printed the Z and X corners in 100% infill, and re-assembled, making sure to very slowly tighten all screws together. The parts are now unflinching, and I feel a lot more confident that they won’t loosen.

 

Belt Anchors

The belt anchors also started to crack, largely due to the decking scews I used to put them in. I replaced them metal hooks from Home Depot. I’ll have to add something to keep the belt from shifting around, but these new hooks are super sturdy.

 

Tightening Everything

Earlier in the thread, I talked about some play that my carriage had. Jeff suggested backing off the MDF screws and tightening the bolts that hold the bearings. This worked great for the carriage, so when we disassembled, I did the same for the Z bearings. Now there’s no play anywhere.

 

After all that, I ran the test cut, and the skull again, and it looks awesome! There’s a little section that has a step, but it’s about 0.1mm, and I’m happy with that for now.

38

Sweet!

Luckily for me, most everything I do on it is through cuts so deck twist and un-levelness are not an issue, I just cut all the way through plus some. Any 2.5D stuff I do tends to be small enough not to matter or my table is actually pretty flat (I doubt that).

Cool, can;t wait to see some pics or videos, the LowRider gallery is pretty sparse.

39

This thread has kinda become my personal troubleshooting journal, so I figured for posterity I’d update it with the latest findings:

I discovered another weak point in my Lowrider - Wires. I made my own extension cable to run the far-side Z and X motors, and I made them poorly. Specifically, I made them using what I had around, which was an ethernet cable and some bread-board wires. The CAT5 did fine, but unfortunately, the bread-board connectors were way too thin to be running the load. After a few hours of what looked like a very good job, the machine stopped moving, and started producing rumbling noises.

I powered the machine down, and upon inspection found that some of the wires had simply melted out of their connector housings! This is because the bread-board pins were crimped very poorly, with only a few strands making proper contact with the connector. In retrospect, I decided that I am not as clever as I sometimes think I am…

To solve the issue, I bought a crimp kit and the requisite connectors. I crimped new pins to some thick speaker wire. Now the wires run cool even during longer jobs.

 

But I do have a new topic to figure out, if anyone sees this and wants to help. I was cutting out a hyrule shield, and I’m getting what looks like skipped steps on the Y axis. Here’s a photo:

The Cut: https://photos.app.goo.gl/R99VqGsyti6OVMgo2

The Source Image: https://orig15.deviantart.net/c75d/f/2011/258/3/9/hylian_shield_vector_by_reptiletc-d49y46o.png

One good place to see what I’m talking about is a horizontal line stretching from the “beak” of the bird to the top of the left wing. Originally, there was a horizontal stretch of missed material there, though it was thin enough that it broke away when I touched it. Additionally, you can see that it over-cut on the left side of this horizontal stretch. It seems like it did that cut operation while the carriage was too low in the Y.

The Gcode was generated with Estlcam, with origin being the bottom tip of the shield. All the pocketing operation was done “linear” so it travels left/right on X during each cut.

What I find really odd is that the design is only messed up for that one moment. I don’t see a permanent shift, which is what I’d expect if the belt slipped.

Here are my theories:

  • Some of these could be fixed when the machine does its finishing contour
  • The belt could be stretching briefly
  • ???
 
40

This thread has kinda become my personal troubleshooting journal, so I figured for posterity I’d update it with the latest findings:

I discovered another weak point in my Lowrider - Wires. I made my own extension cable to run the far-side Z and X motors, and I made them poorly. Specifically, I made them using what I had around, which was an ethernet cable and some bread-board wires. The CAT5 did fine, but unfortunately, the bread-board connectors were way too thin to be running the load. After a few hours of what looked like a very good job, the machine stopped moving, and started producing rumbling noises.

I powered the machine down, and upon inspection found that some of the wires had simply melted out of their connector housings! This is because the bread-board pins were crimped very poorly, with only a few strands making proper contact with the connector. In retrospect, I decided that I am not as clever as I sometimes think I am…

To solve the issue, I bought a crimp kit and the requisite connectors. I crimped new pins to some thick speaker wire. Now the wires run cool even during longer jobs.

 

But I do have a new topic to figure out, if anyone sees this and wants to help. I was cutting out a hyrule shield, and I’m getting what looks like skipped steps on the Y axis. Here’s a photo:

The Cut: https://photos.app.goo.gl/R99VqGsyti6OVMgo2

The Source Image: https://orig15.deviantart.net/c75d/f/2011/258/3/9/hylian_shield_vector_by_reptiletc-d49y46o.png

One good place to see what I’m talking about is a horizontal line stretching from the “beak” of the bird to the top of the left wing. Originally, there was a horizontal stretch of missed material there, though it was thin enough that it broke away when I touched it. Additionally, you can see that it over-cut on the left side of this horizontal stretch. It seems like it did that cut operation while the carriage was too low in the Y.

The Gcode was generated with Estlcam, with origin being the bottom tip of the shield. All the pocketing operation was done “linear” so it travels left-to-right on X during each cut.

What I find really odd is that the design is only messed up for that one moment. I don’t see a permanent shift, which is what I’d expect if the belt slipped.

Here are my theories:

  • Some of these could be fixed when the machine does its finishing contour
  • The belt could be stretching briefly
  • ???