A bed conveyor belt for moving bed printers

Its long been a matter of indisputable fact amongst the 3d printer part makers of the world that it would be impossible to put a viable, functional part moving conveyor device on a moving bed printer. “the motor weighs too much, the wires add mass, the conveyor would shift and fling the part off mid print, etc” . People look at the few “functional” part removers, how big and bulkey they are, and they move on to more productive projects.

I’ve always been convinced that this might be possible though. What if you stripped the motor from the equation and you were just dealing with a pair of lightweight rollers and the belt itself? And then you put the motor on a stick somewhere with a limit switch set to activate when the bed rams into it at full Y maximum extension? And you put a spring actuated brake on it of some sort that prevents the belt from rotating when the motor isn’t in contact? Well, in that case you might just have yourself something viable.

I’m not sure if this one is going to work. There are a few significant hurdles to figure out, but I think I have a decent shot at making it work. For rollers there are a huge assortment of printer part rollers of all sort designed to accept 11" wide paper. I picked up a set of Transfer rollers from an HP laserjet 8150 printer. They are lightweight, fairly strong, used in about 80 different HP printer models so they are cheap and plentiful, and most importantly they have a removable gear with a half moon mounting notch on one end.

The Taz 5 and 6 both have unused M3 mounting bolt holes on the bottom of the bed, and nothing is in the way under the corners, so there is a good place to mount a structure. I figure I can get a tensioner on the back belt easy enough. There is just enough room to pass the belt under the heated bed itself, and a modified aluminum bed plate will allow it to be heated evenly on top, move without snags and allow for installation of guide bars as needed. I have a 24v 300RPM motor on the way that will tap off the PSU, with a standard limit switch in Normally open mode somewhere on it. That will live on a stick sticking out the front of the Taz. That stick will also house part of a chute mount, an identical motorless stick on the other side of the Y axis will hold the other end.

The gear will be a straight tooth gear with a very fine point designed to self allign with the motor gear on impact. There will be an "as yet to be determined matching tooth profile gear brake thingy that will either slide in from the rear or the side when the bed moves away from the motor stick thing, and slide away on impact. I need the motor and position before I can figure any of that out though. Initial test fit of the first roller is promising though. The extra mass doesn’t seem to bother the printer at all so far, and this one doesn’t have the belt tensioner on its y axis yet.

These are likely not final parts, but here they are if you want to see what i am thinking so far. The rear parts will be identical but split with a pivot point below and a screw / heat set insert combo to add tension on both corners. There may or may not be a rear belt brake as well.

I don’t know what the conveyor will look like yet. I have some thoughts on those lines, but I’ll worry about that if I get the other bits working. A buildtak conveyor is a theoretical possibility for example.

Ignore the messed up spot on my bed, thats a 9 month old piece of buildtak and it needs swapped out.
convey2.stl (158 KB)
convey1.stl (80.9 KB)

The tensioner is done. The fit seems decent, I may end up tweaking the holes for the shaft a bit later, but it does what it needs to do for the moment. Now we wait for parts to arrive again!
RR_Convey_tension_arm.stl (137 KB)
RR_Convey_tension_base.stl (90 KB)
LR_Convey_tension_arm.stl (78 KB)
LR_Convey_tension_base.stl (90 KB)

This reminds me of the old Makerbot ABP. http://www.makerbot.com/blog/2010/09/13/makerbot-automated-build-platform

It’s similar in some respects, but different enough that it shouldnh’t be affected by their patent. Mine is based off the original conveyor belt patents that have already expired, do not contain the motor on the bed, and use different mounting structures. Plus hopefully this one will actually work. We’ll see though

They patented the abp? It was a piece of crap. :open_mouth: :laughing:

Does the patent even matter if you are not selling it?

Kind of. If their patent covered this unit, which due to the differences in structure and setup and the expired conveyor belt patents I believe they do not, I would not technically be able to license this as open source either. And yes, the ABP was a piece of junk, but some peole did manage to get them working with sufficient modifications. The patent is based off the entire assembly. since this one uses diferent style rotors, has a belt tensioner, uses different gearing, actuates via impact rather than software, has the motor located as a separate assembly, etc. and is more than 70% different, it should be ok even with all of that.

Besides, if it works well, I may sell them too. Quite frankly the patent shouldn’t have been issued anyways since heated conveyor belts have been used since 1975, several for layer cookie assembly (low resolution food 3d printing)

Patents that should not have been issued is a whole other conversation! What a mess.

I received the rear roller today, printing the tensioner pieces now. Buildtak is also able to supply me with a sufficient length for testing purposes. Belt attempt 1 will be Buildtak with an aluminum foil backer to prevent it from sticking. That should get here around the same time the motor does. Motor activation will use a stock taz limit switch sourced from ultimachine.

I have a general layout of the motor stick and the gears, still not sure what I want to do for the brake though. I kind of like the idea with a toothed arm that is held down by a strong spring and is raised when the arm encounters a ramp at the end now since I think that gets me the strongest hold force.

Anyone have any idea what a large taz print weighs so I can calculate what I need to keep from moving under acceleration? Most of my prints are midsized or flat.

More to come!

Hmm… weight of large TAZ print. Could try to use a percentage of the build volume in the online tool. It has presets for ABS… just put in the dimensions of the Rectangular Prism.

Using 290x275x200mm resulted in 17kg… or 37lbs. Seems high… but you get the idea. :slight_smile:

thats probably max volume max density, I doubt many people would be printing something that large. I think i’ll assume a 10-15lb max bed load for now and see if the brake can hold that if I can figure out a safe way to attach it to the bed in the first place to reasonably simulate a print. I suppose even then a weight is going to be more dense than a print would be so it would give me more of a point load and less of a true indicator of how a print would handle. Maybe several smaller weights taped to the conveyor?

A couple minor updates and pictures. Here is what the assembled rear idler tensioner units look like To add belt tension, you turn the M3 bolt in each corner, which pushes against the “anvil” section and pushes the roller back. It’s not mounted yet because without a belt to hold it in place, it will be floppy. I also got the other roller and the new aluminum bed. I’m already using an aluminum bed on my main printer, this one will end up on the second Taz I built for the openbuilds documents, and will be modified to include a rounded edge for better belt travel and belt guides if needed… The belt itself isn’t here yet though.

The plan right now is to rig the not yet installed second machine bed and a 24v source up for testing. The main objectives will be

  1. can the motor I picked turn the belt
  2. can it turn the belt with a large print worth of weight on it
  3. is the belt itself capable of being turned or do I need a different material
  4. Does everything line up?
  5. etc.

Thanks to a user over in the reddit thread about this project, I was able to order a set of corner mounts to replace the stock springs since I need the bed slightly higher and slightly stiffer, but still minimally adjustable. Auto leveling probably isn’t a bad idea for use with this particular project, but i’m ignoring that for now.

This is all still seeming to me to be feasable so far. I haven’t ran into anything that looks like it won’t work yet. The buildtak conveyor is probably the biggest unknown. it may be too stiff to work in the manner i am hoping for without a bigger roller. thats also the part of the project I’m actually least worried about though too. If nothing else a PET conveyer over a mylar or titanium foil sheet will do the trick probably if that doesn’t work.

It’s getting there! The motor and the belt parts are the next critical bits to move this forward, more after they arrive!

Could try PEI film adhered to mylar or maybe tyvek if its suitable to the heat.

On the print weight, I think seeing how much you can typically print from a 5# roll that is more likely max to try for.

Huh, I didn’t know PEI came in film forms. That would work well too. Seems inexpensive enough too. I’ll get some on the way. Thanks for the idea!

Yeah, especially if you have an automated process and are anticipating running multiple prints in a row, which is what the conveyor belt is for. I definitely need to be careful on the balance between amount of holding force the brake can exert, the amount of force it takes to disengage it, and how much it weighs.

Finally got the motor. It looks like it should work, uses M3 mounting bolts, etc. I will NOT reccommend the “US Seller” who shipped me the #$%^&ing thing from China though. The buildtak conveyor belt also showed up, but I am starting to think it will be too stiff to make the bend radius. I’m thinking a base of laser printer mylar (which can take the heat) and a PEI film on top now. Not sure yet. I know I can make something stick, my main concern remains the mechanicals at this point. The new aluminum bed plate is also here for the conveyor mod. I still need to buy another bed heater to stick on it, but we should be able to test most of it regardless.

With the motor and the limit switch in hand, I can finally start designing the gears and mounts. I expect to have some progress on this by this weekend.

how is the print supposed to come off? is the rolling of it going to peel it off? the force required for that could be pretty high, especially with big prints. This may cause your gears to skip.

or are you thinking of a long roll of PEI that just continually unwinds. you could add a cutter blade and just Print-unroll-cutoff-print…etc.
Or better yet, have a magazine of pre-cut sheets that get fed onto the build plate, and removed when the print is done, then you could use the stiffer buildtak

My thought is that its going to move the part over to the edge of the Aluminum bed, pull the film downwards at a 32 degree angle away from the part slowly at whatever temperature I let it cool to, moving the part forward and removing sections about 0.10mm at a time, then rolling over the roller and knocking the part off the rest of the way if the first angle doesn’t do the trick. The motor has a gear box so it will have a fair amount of Torque, and the gears will be fairly beefy so I am hopeful it won’t take too much force to remove the parts. I won’t know until I try one though.

I do like the print magazine idea. The tricky part would be securing it between prints, but aside from that you could use a 2d printer guts for the sheet transfer

Here’s a side view of the part detatch theory

there would have to be some spring clamps of some sort to secure the plates down. but I know those black metal clips have been used for years to hold stuff on the platten, so it should not be a problem to design something. It can dis-engage when the bed moves to engage the transfer mechanism. If you did it right, you would not even need a motor, the Y axis movement and some clever clamping mechanism would do the whole transfer.

the removal force for prints varies VERY MUCH depending on geometry, material, 1st layer thickness, etc. You will have a hard time making your system work for all instances. I suggest you test the removal force on some samples before you get too deep into it.