Tear down of Cubify cube 3 3D Printer + convert to RepRap

Well I own 2 Cube 3D printers. One was at 1.12 firmware, the other is at 1.14b.

When I put the 1.14b firmware on a thumbdrive, the printer that is already at 1.14b won't accept it.
When I used it on the 1.12 machine, I was able to trick the machine to load whatever I want. The 1.12 firmware machine is the machine I eventually gutted and turned into a reprap compatible machine. I'm going to leave the other machine untouched at 1.14b for now, the thing creates awesome prints.

 

I wonder why such a difference between your "awesome prints under 1.14b" and my very first print (a big failed 3D benchy).... Same machine, same soft, so many different results! I bet yours prints the stuff centered, right?

 

Yeah my prints are centered. I can upload some pics of the stuff I've printed.
Printing in ABS is a nightmare, but PLA comes out amazing. My only complaint is how slow this thing is.

 

The SIM card is readable and writable with Windows. It's just a MicroSD flash drive.
The unit we restored was done by removing all the files and moving them one by one back to the root.
Once it had what it needed... or lost what was corrupt, the firmware update proceeded.

There is one file that has the serial number. Haven't dug into what file that was stored in.
There is an XML file that only has the version written into it. I suspect that this is the handshake expected by the SW version 4.03.

 

A few words on variety; I have 3 printers, one was replaced twice, and have a colleague with one. That's 6 printers so far.
Never buy a white one. They are simply junk. Check the underside to see if the white plastic part holding the head to the rail is cracked.
You can try a warrantee, but likely you will find your machine was clearance and not covered.

One white unit skipped over about 4mm during a long print. Complained, got the rook to print, everything okay... until the reprint where the belt started shredding itself.
Took some convincing, but they replaced it too... but on the 3rd printer, they reset the warrantee date back to the original. Not quite legal, but hey, everything was still within the 90 days and I exercised the heck out of the last white one. Funny thing is, the more printers I got, the older they got. The Each had it's own small differences, notably the build plate. Seems they messed with that for some time. The older ones get large dimples where the 3 magnets are... not so for later ones which pit easier with ABS.

There is also a lot of different levels of build quality. I've resolved myself to check the carrier cover screws for a proper fit. Normally takes a few db out of the noise floor.

Replacement printers had old filament too. One chip unprogrammed and a white cartridge that was too old to get anything to print, even after replacing all the extruder bits.

5 of 6 had 1.11...something up to F and the last white replacement had 1.10...something.

 

And one other favor: What filament are you guys feeding into the stock filament feeder for the Cube 3? It seems to like softer undersized filament. Budget-minder solutions, of course. Any new blend successes out there?

 

Wow thanks for all the info, Tom!

I guess that's why I'm having so many issues, mine is white, but the last thing I would have thought is that printer color would matter! so white ones were assembled at the local zoo while Grey ones were assembled by experts?

may be you can help me out sorting couple issues like you said, noise levels and such?

well first of all, which firmware should we stay in?

secondly, by above comments, I'm fearing Mac software may be buggier, compared to Windows one....

finally, I was thinking that perhaps nema motors would be the main reason to such noise levels, my belts seem fine and tight but if you have any clues on how to improve the noise level I'd than k you as well.

 

It could very well be a second operation startup for white printers. So far this one is behaving.
Main noise contributor is the plastic tray under the Y stage. It rests against solid members and carries all the vibration from the Y servo motor, and amplifies it. Either make sure the entire edge is properly tightened down, or remove it and use some very thin sticky back felt (hobby store) on the aluminum interfaces. The rest is pretty well insulated and properly torqued from what I could tell. I do want to run one in skeleton mode though

The one thing that bugs me, and I hope they fix in any possible future release is the "home" move when purging is complete. Invariably! ...I start another action that pins the motors at both extremes and you have to hope the belt doesn't skip. I'm sure it is designed to tolerate this when new, but give it time!

I'd like to explore the switch thing more. What keeps the Cube from writing to the chip it recognizes next time it sees it?
And what is this right-side cartridge encryption only? Left side acts differently? In what way :|

Still waiting for some filament recommendations... I have some old filament to toss so I have some time on the ticking clock.

 

did u manage to print fine in dual color? got any pics? which firmware do you keep your printers at?

thanks for the tips, will check around the bed and see what I find.

the weird thing on mine besides always printing at the edge of the bed rather than centered (probably macos related) is that the prints do skip a lot of steps during the first 5 to 20 layers, then build fine and straight. which is quite the opposite of what should happen on a lost steps environment. hence I'm assuming something physical is going on there...

 

Dual printing depends on having two nozzles at the exact same level, or at least within about 10um.
That doesn't seem to happened without some very specific intervention in the nozzle. Keeping a paired set of matched extruder tubes could be very useful.
Printing at the edge has got to be some encoder or reference problem.
You could try to re-index the motor/encoder by moving the belt position/relation.
This would be a huge software/hardware oversight. I have not had edge printing issues as yet.
Obviously I am assuming they actually use encoders, not just motor positions.

Printing problems are normally right at the transition from a solid floor to the fill. Invariably, the slightly proud previous layer gets packed back into the nozzle and the backup causes heated filament to swell all the way inside the plastic feeder tube. Eventually hardening and getting stuck. This is normally a very repeatable condition for specific models.

I say this because dual extrusion tends to follow this scenario also. I once has a medallion with 2 colors and only 1 of 10 turned out. The others had mixed colors throughout the print even though only 2 layers of the other color were required.

Overall, yes, it will loose features thinner than the resolution. But you have something much worse going on. Every layer could be explained in all my units until the ~5mm shift of the large print. They said it could be a S/W bug that caused it.

All my printers are running 1.14B right now. The only reason for me to change that would be a encryption chip bypass solution.

 

OK i have a gray unit coming in, we'll see if anything differs. Also i'll ask a friend to process a stl on his pc with windows and cubify and send the file to me instead. is a USB key required, right? for him to send the file to me via email attach for example.

 

A cube3 formatted file can be loaded any way you like. Direct from a USB drive or from the Cube s/w (add from file).

I attached the 3D Systems Rook file in PLA. Ready to print.

 

thanks, which Windows are you using out of curiosity?

 

Let's see if this works...
When you get your new printer, look here before anything else:

Cube3 - Printhead carrier cracked by Tom Dirriwachter posted Nov 13, 2016 at 10:12 AM

 

I will. however, loosing steps has been "consistent" on the Y axis all the prints I've tried,but just for the first 10-40 layers or so.

 

I'm still on 7. Only issues I have is that the 4.03 software is really hard on the NVidia card.
Already blew out one GPU and get constant kernel warnings when deleting files from a massive library with the replacement.
Communications is also a hit and miss endeavor. Sometimes the printer is found, and other times not. Real touchy about network glitches.
Otherwise, WIFI is working well for transferring files.

After the shift error, the following prints self-centered again.
Error:

Cube3 - Failed print; offset by Tom Dirriwachter posted Nov 13, 2016 at 10:12 AM

Centered again:

Cube3 - shift error not permanent by Tom Dirriwachter posted Nov 13, 2016 at 10:26 AM

 

what is belt fraying? sounds like the exact issue I'm having.... How did u solve it? just full replacement?

since we are on our own now, anybody dismantling their units should really try to document everything! hehe

 

The belt should not shed many particles at all. In my case, the left and right upper shelves were massively dusted in the unit that was failing. This tells me that something got locked or jammed where the system lost its reference. Again, I am not sure yet how they do positioning between the software and hardware. They can count stepper motor steps without an encoder. Any failed move due to excess torque would not be noticed by the software.
They replaced this unit after much convincing.

The belt should be standard stock once you confirm the pitch and number of teeth. Try to match the duty of the belt too. It may even be open-ended, which would be great.
Anything with a 39xxxx part number is 3D Systems... everything else should be searchable or source-able.

I do know for a fact that the bearing rails are -not- Japanese made IKO. I was told they lost that contract. I trust IKO bearings.

 

The rails thing.... on whites only or everywhere?

 

No joy, your rook file prints at the edge as well

 

Hmmmmm I just re-reflashed once again the printer to 1.14B and started a new print. It seems now its printing centered....But what I actually think that happens is that the bed this time was all moved backward before starting the print, hence the printer "centers" it. This apparently wont happen if the bed was all at the front position. Can anyone check theirs?

 

OK now Im speechless!

From the producers of....

THE MOST... (pic 1)
GRUESOME... (pic 2)
BENCHY PRINT! (pic 3)

now coming to a theater near you....

THE ROOK!









Looks like the night and day, two different printers.... but it's the SAME one!


The only two things I did:

1) Tried to re-reflash the 1.14B firmware with the SAME 1.14B firmware of the SAME file of the SAME USB I used initially on the printer after recieving it. It proceeded as usual.

2) Tried the rook file uploaded by you created on Windows instead of a Mac.


YOu know the rest of the story now!


Just wow.

I really thought my printer came with severe hardware deffects out of the box....but.....wow!



So next torture test:

1) mind uploading a BIcolor rook file made on Windows Cubify? I'm assuming you used latest version on the monocolor Rook, right?

2) anyone with Windows AND Simplify3D installed mind creating a sliced file, converted to .cubepro and renamed to .cube using this?



BTW Would Cube 3 owners be interested in a wikia of CUbe3 hacks and findings, containing all possible known hardware and software files, links, tips, tricks, hacks and suck?

 

My biggest issue (other than proprietary filament) thus far is with supports. I've tried using the support generator in Cubify with mixed results. In big overhangs I still get a lot of very loose filament, the first few layers aren't adhered well. I've tried generating supports in meshmixer as well, and I clearly have no idea what I'm doing there as the transition from support to part is a gloppy mess and is impossible to remove. I've got some learning to do there.

My most pressing concern right now though is going to be getting some bulk filament running through. I think I'm going to have to try the cartridge swap and printing with 0% chips trick. We're planning on doing a few gifts and I need to be able to run colors I don't currently have in cartridges.

I looked at my white cube and It looks like my printhead carrier is also cracked. When I get to taking this thing apart I'll have to see how easy it would be to draw up a replacement and either print it or possibly machine it out of some aluminum on my CNC.

Haven't decided yet on the lobotomy and conversion to reprap for mine. I don't feel like I have time this week to do that.

 

Did you have to disassemble anything to check for the crack?

I dont think mine is, but want to be 100% sure.

ALso there's a swap chip method based on a switch which relies on 2 used chips wired to the switch like this:

















Basically you need a 3% chip and a 0% chip and swap them using the switch when printer warns you about low filament.

 

Nice. And you just ran the wires back to the contacts on the outside under the cartridge?

Any trick to removing the chips? I haven't tried yet, worried about breaking them though. Do they pop off pretty easily?

As for the crack, I just tipped the cube back a bit and looked behind the print head. I could see a line radiating out from one of the cap screws, so it looks like it got over torqued.

 

I'll check my head tomorrow... pics and method are not mine but I guess it works, there are a couple youtube vips proving it.

 

Where'd you find those pics? I've seen the YouTube video, but hadn't seen those pics before. Curious what other research I've missed. The Wikia is a good idea, btw.

 

I got them emailed by one of those youtubers

 

Wiki a is up! at cube3.wikia.com

 

The chips are just glued in place, they pop off with a little elbow grease.