Sheet Metal Brake and 3d Printer.

[ddmckee54]

S.Heslop:

As long as you have the 2.54mm pitch JST connector, and that pitch is all that I've seen on my printer or the other "stuff" that I've worked on, then the DuPont connector will fit the pin spacing.  You just don't have anything other than the friction between the sockets and the pins to hold the connector on, so the connector could vibrate loose.

Long term, I'd get the new connectors and eliminate the additional failure point of an adapter.  Short term while you're in the "I just want to see it run" mode, the DuPont connectors will work just fine.

Don

[awemawson]

A blob of hot melt glue works wonders, and if not too extensive can be easily removed

[S. Heslop]

Congratulations!  I'm totally amazed that you got a good print with that few ugly ones!

Same here! It'd have been even better if I remembered to set the steps per mm on the extruder.

A blob of hot melt glue works wonders, and if not too extensive can be easily removed

That's smart. I was wondering what'd happen with the connectors since I want this board mounted upside down.


Still considering what to do for an enclosure.



Something like this is similar to what i've done in the past, and i'd have half the materials. Plywood is appealing since you can easily screw stuff into it. The grey rectangle is the laptop since I think it'd be convenient to run the machine from that.

There's alot of talk about fire hazards so I was hesistant about plywood, but on closer inspections alot of it seems to be the result of crappy machines. It turns out they're disabling safety features like thermal runaway checks to save on program memory in the firmware, so they can use cheaper electronics. But it's still something i'm thinking about.

Also looking at commonly available building materials it seems like cement backer board is the only thing that's truly fire proof, but I can't imagine the dust being healthy for me or the machine, and sticking something like vinyl on it would defeat the point. There's glass too but I don't really want such large panels hanging about, especially on a machine that might get a little warm.

As for insulation - I was in the loft when trying to sort that water tank leak and found there's a few unopened rolls of rockwool up there. I could use staples to trap that behind some woven glass roving, to reduce the combustible stuff. The plywood would probably be enough but i'd also like to try insulate sound.


Oh yeah I also did the two pillars stringing test to check the retraction. Default settings seem fine with no stringing but one of the corners got a bit wobbly. I feel a part cooling fan might help with that.

[WeldingRod]

I used polycarbonate for mine.
I wouldn't insulate; you don't want to get too hot.
I would make at least one panel transparent so you can check up on the motion.

Sent from my SAMSUNG-SM-G891A using Tapatalk

[S. Heslop]

Setting this off on a 3 hour run to see how it copes. When levelling the bed and moving the gantry around though the stepper motors seemed alot stiffer than they were. Are the bearings failing already?

Edit: Ah it turns out it's because they're plugged into the board, and it must be sending the current back into other windings. I was worried something was failing already!

[WeldingRod]

You DID do mesh leveling, right?
3x3 is plenty, but its important!

Sent from my SAMSUNG-SM-G891A using Tapatalk

[S. Heslop]

I've been doing it manually! Honestly i'm not sure how levelling works without a probe, but one of them BLtouches are on the list.

Did a big run today to check out how it handles 'real' features. Was having trouble with layers shifting though. The first 2 attempts were aborted since it was real terrible, and this is after unplugging the Y axis home switch once it'd homed at the start.


The right hand side here is supposed to be a ( shape, but the holes are obviously shifted.


This should be straight.


Here it is assembled. I adjusted the steps per MM a bit more after some measurements of the failed runs and it's all pretty damn precise now. More than i'd expected.


And this is what it's supposed to look like. It's more of a mock-up than anything final. It's also nice to do some actual Rapid Prototyping and see the general size of it, as well as what features clearly aren't going to work.


Anyways the layers shifting. I think that could be from interference but I still got some even when it was unplugged. But that again could be from... weird electronics. I'm no expert but I know there's some stuff about pull-up resistors and stuff, which I believe are built into those mech endstops. I should really try ground the cable and see what happens. It's odd though that the X axis has no problems at all despite being a longer and unshielded cable, although I did twist the wires.

Also I believe i'm still under-extruding. I should measure the steps per MM of the extruder as well and calibrate that.

Oh and the progressive infill setting does a great job at making the top layers flat.

[WeldingRod]

I screw a small dial indicator on my carriage, with an extension.  At each position I go down until the dial stops moving, go up, then down to the last click that moves the dial.  What's happening is the nozzle touches down and then the drive system starts to flex.  It works really well.
Then, you save those parameters, and you are good for months.

Sent from my SAMSUNG-SM-G891A using Tapatalk

[WeldingRod]

On the steps per mm, put a thermostat on the filament, measure the distance from the stat to the printer, then extrude some distance.  Measure, and fix your steps per mm.  Recheck afterwards.

Sent from my SAMSUNG-SM-G891A using Tapatalk

[WeldingRod]

Sorry, stupid spell check doesnt speak mad modder!
Hemostat!  Super useful workshop tool!

Sent from my SAMSUNG-SM-G891A using Tapatalk

[S. Heslop]

I've calibrated the extruder... extrusion better now. Still having trouble with shifted layers on the Y though. Set the accelearation and jerk extremely low and problems continued. Still running tests.

Shielding the endstop cable too and seeing if that makes a difference but I doubt that's the issue. I'd expect a different look to shifted layers than just a slight... wobble. It makes sense to me that being on the Y axis it'll be due to its weight. Everything moves freely, belts seem tensioned fine, pulleys aren't slipping. But even with accelaration and jerk way down plus a slow printing speed all it did was make the quality way worse. The shifting is also random but in small amounts, so it's hard to tell if any changes are a real improvement or not. Current is as high as I dare for the built in controllers and motors (1.4 amps).

My last guess is that the poor quality infill, a problem i've not fixed (this current test with the better extrusion calibration might work better, I was massively under-extruding it seems), does raise up enough that I hear it clicking as the head catches them occasionally, which could be what's losing steps.

[S. Heslop]

Slowed to an absolute crawl and it's still leaning. Added a capacitor accross the ground and signal of the endstop switch and that hasn't made a difference. Even probed it with the oscilloscope and i'm not seeing any obvious noise. They don't always lean in the same direction.

Maybe I could try a program different to Cura.

[WeldingRod]

Please post photos!  I've made lots of bad parts!
Pick something simple, like a tall rectangular block to start with.

Sent from my SAMSUNG-SM-G891A using Tapatalk

[BillTodd]

Watching carefully here, as im also embarking on a 3d print exploration...

re: layer drift,  might be worth checking tightness of drive pullies - mine came loose after a few hours work.

Also, are you running the steppers too fast?   your axis might be losing steps on fast moves.

Bill

[S. Heslop]

Watching carefully here, as im also embarking on a 3d print exploration...

re: layer drift,  might be worth checking tightness of drive pullies - mine came loose after a few hours work.

Also, are you running the steppers too fast?   your axis might be losing steps on fast moves.

Bill

I've really gone through the whole checklist of things. Next paragraph is a big block of stuff I can remember trying.

Made a new simpler model, grounded the cable shields, put capacitors between the endstop signal and ground, probed about with the oscilloscope to see if I can spot any obvious interference, ran it super slow, ran it super fast and every speed in between, set the jerks low and high, accelaration too, got stuff tuned in so curling infill wasn't bonking into the head, made sure every nut and screw is tight along with putting lines on the shafts and pulleys to see if they slip, adjusted the belts, switched around the X and Y connectors to see if a controller is faulty, ran an indicator up the Z axis rails to make sure the screws werent pulling it about, set motor currents high and low and everything in between, made sure things moved freely and easily, tried other slicer programs... probably more i'm forgetting but you get the point.

I just completed an absurdly slow run on a new test object, and it's still wobbly.


The one on the left is the new one and the others are randomly selected from my now vast collection. The X axis direction is perfect.

So a pattern i'm noticing that isn't super clear is that it starts to shift in a sort of bulge outwards right before the top. Wonder if it's related to the progressive infill setting.

I also set up a dial indicator on the Y axis gantry, so when I jog it back and forth it should reset to 0. And it did for a few turns sending it back and forth 100mm, checking the indicator every few moves back and forth. It stayed at 0 for a while then moved a couple graduations away, then a couple more. And then back to 0. A full step is about half the dial - it's one of those lever type indicators that are way too sensitive and I don't trust the actual numbers, but it's all i've got.

So i'm not sure if that's from losing steps or if it could just be from how the belt sits on the pulley or something. Can you lose microsteps? Also I doubt jogging it back and forth a bunch of times is enough to really test if or how it's losing steps.

I've ordered a new stepper motor cable for the Y axis. Perhaps there's some intermittant short or something and at £1 it's probably worth a check.

Also I'll probably post this problem on a 3d printing specific forum and see what happens. I just want to make sure I get the uh obvious stuff out of the way first.

[BillTodd]

Can you lose microsteps?

Micro steps are only a sort of 'fudge' between steps they make the motor run smoother but have no defined position.

That said I do not think that is your problem.

It does appear to be worse (or only ) related to one axis .

[edit] One thought (looking at pics) : the head is mounted by two bolts (height adjust?), could it be wobbling between the two ??

[S. Heslop]

Can you lose microsteps?

Micro steps are only a sort of 'fudge' between steps they make the motor run smoother but have no defined position.

That said I do not think that is your problem.

It does appear to be worse (or only ) related to one axis .

[edit] One thought (looking at pics) : the head is mounted by two bolts (height adjust?), could it be wobbling between the two ??

Oh yeah that's one of the things I checked. It seems fairly ridgid but I should probably put the indicator on it to see for sure. Right now though i'm making that exact same test piece again but rotated 90 degrees to see if it's more to do with the geometry or some slicer settings. I already did a test like this on the more complicated part, but i'm feeling the taller part makes it alot easier to see what's happening. The pattern being seemingly random too also makes it hard to really gleam anything meaningful since one run might randomly look fairly good if it's relatively short.

[BillTodd]

 Tall part.. is it rocking ???

[S. Heslop]

Oh yeah WeldingRod, what speeds do you generally run your stuff at? I thought I had it for a bit with even lower speeds since this part looked very even from the front. But it turns out it's just a relatively even lean in the Y.

The part rotated 90 degrees turned out well so I was wondering if maybe it was the slightly faster travel speed as it scoots accross the whole part long ways. So I tried turning the speeds down even more. But i'm also starting to think i'm running on bad info as to what counts as a 'reasonable speed' (40mm/s). Running it at 30mm/s now but I've had the accelaration riddiculously slow previously where it probably never broke 20mm/s. Hesitant to quote the numbers in case it's real obvious that's my problem, but dang it feels slow.

Also the change of direction in the infill felt somewhat violent so tomorrow i'll try running it even slower on the infill. It's getting a bit late now and I don't want to trap myself into over 2 hours of waiting for the machine.

[S. Heslop]

Tall part.. is it rocking ???

I checked that too. It's only 20mm tall but with a brim and good adhesion from what I can tell. But the lines are appearing in the long direction. It's 115mm long and 15 wide.

[WeldingRod]

My Simplify3d speed settings for 1.2mm nozzle:
Default printing speed 2500 mm/min
Outline underspeed 50%
Solid infill underspeed 30%
Support structure underspeed 50%
X/y axis movement speed 4800 mm/min
Z axis speed 1000mm/min

For 0.4mm nozzle (which I haven't run in a long time):
Default printing speed 3600 mm/min
Outline underspeed 50%
Solid infill underspeed 80%
Support structure underspeed 80%
X/y axis movement speed 4800 mm/min
Z axis speed 1000mm/min

I will have to get acceleration and jerk off the machine tonight.

Sent from my SAMSUNG-SM-G891A using Tapatalk

[WeldingRod]

Can you try printing a single wall square tube, say 30mm on a side, at least 50mm tall?
You may want to try both vase mode and regular.
This takes out some of the factors like infill...

Sent from my SAMSUNG-SM-G891A using Tapatalk

[S. Heslop]

Switched the X and Y again to try it once more. And with this taller part it's clearly a problem with the driver since it's now wobbly in the X direction. I guess I should see if they'll exchange the board. And failing that i'll have a look at external driver boards.

Edit: I'm a little peeved but also relieved to know what it is now. I should've caught it earlier but I ruled it out after a test. Because the layers are randomly shifted that one test I did must've had the planets align and most of the layers turn out fairly in-line. This time it's way more blatant.

They accepted the return so i'll send it off tomorrow. But woof what a hassle. Second controller to go bad even if the first one was my fault.

[hanermo]

Excellent thread.

Does anyone know of great uses for 3d printed stuff ?
Commercial ?

[S. Heslop]

Excellent thread.

Does anyone know of great uses for 3d printed stuff ?
Commercial ?

I still think a milling machine will take you further than a 3d printer will for actual engineered parts. Or even with wood you could produce stronger parts quicker. And for artistic stuff the FDM style of 3d printer is kinda ugly and requires alot of finishing, where more traditional sculpting might work better in most cases. But as far as I know though these machines have been around since the 70s for rapid prototyping in commercial use. But I think other methods (SLS?) are more popular for that now.

I might've mentioned it already, but what I want the machine for is making more of those goofy fursuit masks. I've made a few in the more traditional way - sculpting clay/ plasticine and making a silicone mould from that. But storing them takes up alot of space, you're limited to the base shape you started with, and it's hard to really visualise the final look since the fur and other stuff will increase the size. Plus the urethane is hard to adhere to and attaching 'hard points' and features to them requires alot of fettling to make them fit. It feels very bodged and I was hesitant to sell the one I did because of it.

So for this one very specific job I think 3d printers are probably the ideal machine. Without being locked to a mould it'll be easier to iterate the 3d model to fix shortcomings, or to sculpt it to various shapes. Or even scale it to fit different heads. And it should be easier to design more complex and elaborate features.


Er all that said though, i'm sure i'll find other uses for this machine once it's running. But I've gotta be careful to not get carried away with it. Like I was thinking earlier about clips to hold the power supply under the machine, and how I'd best 3d print them. Before realising it'd be a 5 minute plywood job on the bandsaw and drill press. One of the first times I ever saw one of these machines being used for something, in a video, had the guy 3d printing a square stick.