I really look forward to seeing how yours turns out once you get a chance. 10mm may be overkill but I had the rods and linear bearings left over from a printer build. 2 x 8mm isn’t going to flex over 120mm. I probably wouldn’t go much lower though as there are some lateral forces and the rigidity helps avoid failures.
Rounding. The firmware is only accepting 2 decimal places for E. I feel stupid because I made a mental note this might be an issue, but didn’t spot it when it showed up.
I will update the code tonight to output degrees instead of radians and see what happens.
Ha, nice catch.
Bit of a breakthrough, able to cut 7mm double wall card in a single pass now, so back on track. Turns out very precise blade angle tracking is needed to cut at those depths. Gcodetools has been modified to optionally output degrees in the gcode.
I had hoped for a solution where firmware changes are not required, but sadly that doesn’t seem to be possible. You need the precision of going to degrees (just higher magnitude values, but degrees makes sense). But this triggers the long extrusion prevention. So I had to comment the following:
configuration.h:
#define PREVENT_LENGTHY_EXTRUDE
#define EXTRUDE_MAXLENGTH 200
1 Like
Nice.
After leaving this alone for a week, I had some different ideas. I attached a couple of screen shots of what I came up with so far. Let me know if any of this will not work. I had to add 20mm to the height of the TN_machine_Mount insert for the blade to clear the bottom of the quick disconnect mount & therefore had to lengthen the M8 bolt to 90mm. I can find M8x90mm hex bolt caps screws from http://www.mrmetric.com/M40425 (They also have these in stainless, but are currently out of stock). I changed the 10mm rods to 8mm & made the mount with a plastic bushing for them using the profile of Drylin_Bushing_v2.stl from https://www.thingiverse.com/thing:1871890 . I measure 5.5mm from the bottom 608ZZ bearing to top edged bump of the TN_Blade_holder.stl for the spring adjustment. Is that enough gap? I put the 8mm smooth rod on each side & below below motor to give me more space to work with & those rods are now 100mm instead of 120mm. Will that work ok that way? Is adding 20mm to the length of the M8 bolt going to make it less stable. Seems like moving the M8 rod connections below the motor should give it more stability. I probably need to add kind of a fillet weld where the M8 flat rod connection is on the motor mount to make it more solid. Using the plastic bushings, I might have room to use M10 rod. In the screen shots, I just noticed I need to shift the blade up to the other notch on the blade, but it still fits in the design.
1 Like
Great to see, although I’m not able to understand all the changes so far. I love the plastic bushings - I might try those out, and its great if we can get rid of the linear bearings.
I looked at an arrangement where the guide rods are either side of the A axis, but went with my solution to avoid too much width and too much distance from the gantry. I doubt your changes would introduce any instability. As for the spring gap, it can be as small as you want as long as you get the right spring tension (adjustable better I expect). The actual travel in the Z direction is going to be very small.
I expect the area where we can usefully add stability will be in the lower blade holder assembly. It’s blade tip is too far from the bolt, and creates a lever effect reducing lateral blade stability. I want an Olfa snap off blade design like the TOCK has. Been thinking about somehow cutting a blade channel in the bolt itself. The mechanical stiffness between blade tip and bolt is a key weakness of the current design.
Anyway, as a beginner to mechanical design, it’s good to have your experience involved in the project, so please keep going and let’s find the best design possible. I’ve had some good cutting speeds, but the failures are still there occasionally. It would be great if you can try your design out and feed back. The Thingiverse page is where I’m recording updates to the system mainly, until I can say I’m happy with the whole solution, whereupon I’ll report back here with videos.
EDIT: Quick after thought around the PLA bushings, is there any possibility of printed PLA rods to eliminate the metal ones?
After reading some comments on that 3d printed Drylin bushing, I am not sure how well they will work 3d printed. I also wonder if that is a patented design & I should not use that idea in my remix even though there is that design on thingiverse using it. I have a 7.5mm carbon arrow I bought from Walmart awhile back & might try creating a single bushing to test it with. I have enough room to put 2 bushings on one side & if I used carbon fiber I might use it this way as shown in the attached image.
Mission accomplished - I made a scraplight. It’s late so pics and vids tomorrow.
Latest findings are:
-
Cardboard is the best spoil board for cutting cardboard.
-
Two passes is the most reliable and fastest way for 7mm cardboard.
Cuts are very reliable and fast now. One pass cutting this material reliably will demand more rigidity in the cutting head than my tool allows.
Cool. I printed one of these bushings sized for the 7.5mm carbon fiber arrow I have & it seems to work well. Since it is all vertical movement I don’t think it is a problem wearing the plastic. I found a lot of various designs of those plastic bushings on thingiverse, so it does not seem a problem to use them. I made a customizable openscad design of the bushing to be able to test it for various sizes. After the fact, I found a couple of customizable designs on thingiverse. Just search for “bushing” on thingiverse to find them. Aafter seeing how well this works, A plastic shaft might work if it is thick enough. I am going to make a bushing to use an 8 sided oblong shaft so I can print it on a face & see how strong that is.
Great, I really look forward to hearing about your results. So to be clear, you’re going to try out an all PLA linear motion system based around a single shaft? I take it the 8 sides idea is to reduce the requirement to one shaft. If that works well it would be really sweet.
This is what I had in mind & putting like half 8 sided bolt head on the top so it could just be held in with gravity. Or if that does not work, secure each shaft thru the side of the motor mount with a M3 screws. M3 screws would probably be easier. I don’t really want to superglue into the motor mount.
Since the drag knife part is the weakest link, how about changing the design of that part to move the M8 bolt head closer to the knife blade? Looks like you could move it 10 to 20mm closer, but of course would need a longer bolt.
I decided to take a look at the blade holder & associated parts & this is what I came up with. I made the M8 bolt connect 20mm lower which would lengthen that bolt. I could easily change that to a shorter length. I did not add that ridge that you have to match the holder and bolt clamp, but did add a 4th screw to hold those together. Seems like having the bolt head closer to the cutter should make it more rigid. What do you think?
I know the hypercube 3d printer runs on 10mm carbon fiber rods for its x axis using these bushings
I guess he has had them on and working for some time printing 12-14 hours a day with very little wear to the carbon fiber rod. I am getting ready to build the hypercube myself so I will see how well they work soon.
Keep up the great work on this! After following this thread I want to cutout my own cardboard box from a flat sheet just because lol.
Neil
1 Like
Thanks for the progress, yes, shortening blade tip to bearing (fulcrum) distance will definitely help.
The ridge was inherited from remixes of remixes, but it does support the positive location of the blade tip. When your offset is anything from < 1mm to 4mm like a drag knife, this is useful enough, but when you are looking for perfect alignment between centre of A axis and blade tip, I see it as important.
One of the other battles I’ve yet to win is having the blade tip turn 360 with no movement. I went from pineapple to flexible metal coupler back to pineapple, and am awaiting delivery of a rigid alu coupler.
Once I start stress testing corner accuracy with small lettering and intricate work, I expect this to be an issue.
Just discovered this, great job working all of these things out James, I’m planning to make one of these to test it out, but my printer is quite slow, so I’m just wondering if there’s a big change coming up so I don’t end up printing it twice. Also, is it possible to easily switch it back to a normal mpcnc with a 3-d printer attachment? (That is what I currently have and am using to print)
Here is another option for a coupler for you to try while waiting on your aluminum one, but this one is probably not any better than the pineapple. Parametric Shaft Coupler by GeoDave - Thingiverse You can open it in thingiverse customizer to see how it looks.
I have been printing some test prints of the bottom part the blade holder and blade clamp to see how it looks and feels. After the 4th test print, it looks ok as you can see from attached photos. I will put this remix on thingiverse after I cleanup the openscad file some. I made the clamp a little longer to cover the blade edge to the end & give a ledge to bear against the blade. I also made it so I only have to take one screw out to change the blade. I need to adjust my infill orientation, but left it alone for the test prints. These test prints only take 17 minutes to print.
I made the blade holder so you can adjust for 0 to 20mm extra length on the M8 bolt. That way someone can test it if they are using the M8x70mm screw.
I am using Stanley box cutter blades, but in my research in looking for standard dimensions of these blades, seems like they are not all the same for this style blade.
Since the motion the nema17 does not require a lot of power, I was wondering if a smaller motor would work as well for this? It would also cut down the weight. I have been using this pancake nema17 for a year on my titan extruder. Amazon.com
Isn’t that great? Whenever I’m adjusting something, and it’s quick to print, I think, “This is how 3D printing is supposed to be”. Just make a rev, send it to the printer, get a notification that it’s done, go get it, bring it back to your desk and work on the next rev. Whenever things take 4 hours to print, I find that the process stretches out over days, and I make two or more changes at once.
They come in different numbers of slots, and different widths. I printed a holder for one that was for wire strippers, and it seemed like it was hard to figure out which blades I had. Pretty annoying. Although for a project like this, I don’t mind buying specific blades.
The issue here is that during some operations, the blade is turning in the material, and that’s causing torque on the blade, up the shaft, and the coupler is the part that’s doing the flexing? Is this something that the software is failing at, or just the nature of the beast? i.e. is there too much torque being exerted on the blade because of the path it’s following?
Looks awesome, will try for sure, thanks.
It really depends on what you’re cutting. When the blade is 7mm deep, either you lift on every turn no matter how small, or there is some twisting force needed.
Working with arcs has proved to be no problem, but the output of Fusion 360 slicer is such that you have lots of small curve segments with a turn required between each. If you set the threshold so that the blade doesn’t lift for these small turns, you might find you need the higher torque.
On the other hand, if you are just looking for a solution to cut sharp corners in vinyl, any size stepper will do.
The issue I’m describing is with the blade just touching the material, turning it manually should create a tiny point. If the A axis motion is not translated in the correct plane because of axis mis-alignment between stepper shaft and bolt, the point will describe a small circle.
Ah. That makes sense.