[Geschlossen] MMU 2.0 video from MRRF  

Re: MMU 2.0 video from MRRF

I saw that. More info please Joe! I like this concept!

Re: MMU 2.0 video from MRRF

More and more it seems like upgrading the existing MMU won't be cheap at all. New board with Trinamics drivers...

Re: MMU 2.0 video from MRRF

This is actually quite revolutionary. I got to say that the mmu2.0 should be cheaper because we don't have to buy that bondtech extruder which is quite expensive for those gears alone. The board, I don't think would cost much like the electronic multiplexer in the initial mmu since we only gonna use 3 motors for the mmu2.0.

I have 4 colour mmu, what I needed to upgrade to mmu2.0 is only the board, 3d printed parts and other stuffs. except that of new stepper motors used in the video. They seem to be a special type of stepper motor where the shaft are already built into the stepper instead of using a bulky coupler.

That hobbed gear array is something as well, I would think that it is going to give us whole new problem where the mk2 units, those hobbed gear are worn due to friction and others.

I also never thought that this unit, they would incorporate that "cutting" mechanism in case of filament jam occurs. Weeks ago I got in mind "would be nice if they would have auto cut the filaments end included..." but they did it anyway.

Re: MMU 2.0 video from MRRF

This is actually quite revolutionary. I got to say that the mmu2.0 should be cheaper because we don't have to buy that bondtech extruder which is quite expensive for those gears alone. The board, I don't think would cost much like the electronic multiplexer in the initial mmu since we only gonna use 3 motors for the mmu2.0.

I am not that sure about the kit being cheaper because it seems to me the multiplexer board on MMU 1.0 based on what soldered on it looks to me cheaper than the new board. You don't have the bondtech but still need 3 motors including one with a shaft built in, filament sensors... In the end, I expect the system to cost nearly the same as the previous one.

For those like me who have MMU 1.0, if the possibility of upgrading exists, it will probably cost a bit there's only the motors that you keep and you need to upgrade to the recent MK2.5/MK3 extruders.

Re: MMU 2.0 video from MRRF

The way I see it, if you have the MMU 1 you bought:
- a now useless board
- 3 probably useless motors (new ones seem to have integrated lead screws)
- 2 or 3 useless bondtech sets, probably one will be reused in the extruder (but still: ouch, 50 bucks a piece)

The new design also needs bearings and hobbed gears (5x). Trinamics are not cheap. It does sum up. Considering the total uselessness of the current MMU for most applications I had in mind, I can only hope there'll be a good discount.
I'm still struggling whether to go back to MK2s single, things which need printing and cannot be done with the MMU installed keep piling up.

Re: MMU 2.0 video from MRRF

I'd like to see it in operation, but I don't think we'll see that until it's unveiled.

Re: MMU 2.0 video from MRRF

More and more it seems like upgrading the existing MMU won't be cheap at all. New board with Trinamics drivers...

Depends, there still was a multiplexer board and the chips could of been the same price. trinamics (depending on model) can be $5-10 each in low volume... high volume could be less. one less stepper is like $10-15 less. no idea the cost of bondtech gears but i guess they're pricey.

However this thing is at least 2x as complex with twice the issues bound to happen.... more moving parts = more issues. There is only a usb interface so how does it work? it seems it wouldn't be connected to the controller board either but a usb port.

Since there was no actual demo at all at MRRF showing it at least can switch filament, it's not even but a prototype..... expecting the thing to ship out anytime soon and actually works given the prusa track record for that is pretty darn low. even the MMU v1 had a demo at least. people will get one, and they'll be the beta testers and 6 months to 1 year later the firmware will finally be bug free enough for a 10 hour print. or it will actually just work.... and i'll have to buy or build one (open source after all...).

at least though it looks cool and might just work out.... but the date it's suppose to start shipping out seems overly optimistic considering the lack of demo.

i was going to do a filament loader that loads filament from a bookshelf sized enclosed rack.... each spool would have a geared loader (without motor) that you load it into and a head moves up/down the entire thing and plugs into it and extrudes it out. i never thought of using it as a MMU type deal, just to make it easier to load filament without pulling them in and out of the damn bins... and it would look cool like those inventory robots.

Re: MMU 2.0 video from MRRF

I hope they do some testing on 120v power. I don't know where it's getting it's power, but if they add it to an already under powered PSU then there's going to be issues.

Re: MMU 2.0 video from MRRF

The way I see it only one motor moves at a time. I didn't hear of any power problems with the current MMU, so it should be alright.
Bondtechs are around 50€ consumer price per set. Concerning the multiplexer, I don't think they'll hang on to that. Josef says in the video there's an Arduino on the new board, that could easily run 3 drivers and an endstop.
Either the USB is only for testing and flashing the firmware, using the same wires for control that are used now. The Arduino would compare the current carriage position with the requested switch state and handle the loading independently.
Or it has feedback, then I guess they'll use the USB port on the printer and a power wire (but this would mean no Octoprint).

I'll remove my MMU this weekend (what a massive disappointment overall) and sit this one out.

Re: MMU 2.0 video from MRRF

It was stated that the MMU 2.0 upgrade for 2.5/3 would be the same cost ($59) as the MMU 1.0 upgrade. I'm not sure if this was only for pre-orders or not. I can't find the Facebook post right now.

It was also stated that the MMU 2.0 was less expensive to produce than the MMU 1.0, but the upgrade to 2.0 was more expensive to produce than the old style upgrade (with just sensor boards).

Re: MMU 2.0 video from MRRF

It looks like, for every material change, at least a foot of filament will have to be withdrawn and then another foot of filament will have to be inserted. Won't that increase print time dramatically? On top of the purge time?

Also, when returning to the same material, that length of filament that was already withdrawn will then be re-inserted. That means the first foot of filament after each change is going to be inserted twice, which means three passes over a gripper pulley. Suppose you just use a small amount of a particular filament for a layer. Some amount of that filament could be withdrawn and re-inserted multiple times. In fact, if there is a blockage, there is a knife that will recut the end for you to try again. That's another situation where the entire length will be fully withdrawn and then re-inserted. What will the effect of running a portion of filament back and forth numerous times over a gripper pulley going to be? Seems like it might get pretty chewed up.

Actually, a foot of tubing doesn't seem nearly enough. The vertical range of the Y axis is nearly a foot. The tubing needs to reach every extreme between the material feeder and the extruder with enough slack for bending and flexing. Seems like more than a foot is necessary.

What happens when the filament ends in the tube? How do you get it out? Seems like the only way would be to manually extrude it.

Re: MMU 2.0 video from MRRF

I think they'll use an unmodified MK2.5/3 extruder (except addition of the PTFE transfer tube), which means they can use the filament sensor. That way you can live with lots of slip in the multi unit (just push until something arrives at the sensor) -> less pressure, no grinding.
The bit that goes through the gear twice won't play much of a role because you're very close to the hotend when it happens and there shouldn't be a narrow spot where it can catch.
You're probably right about the retract paths, but this motion would be fast. Also I think you can get away with a lot less purging, which should more than compensate.

Re: MMU 2.0 video from MRRF

Also I think you can get away with a lot less purging, which should more than compensate.

I don't see that at all. The volume of filament remaining in the hot end will be the about same with both systems.

And then there's the problem of the different gearing. The MMU2 uses a different pulley which will impress on the filament; there may be a situation where the filament will move in an attempt to match the impressions with the Bondtech gears.

The filament could well be driven through the MMU gears many times (not just 3 or 4), depending on the model extrusion and purge required - just think of the situation where you are extruding a few millimetres with a black filament (purge = 25mm); the path between MMU and extruder will be maybe 400mm, so that could well be 14 times through the upper gears.

Yes, I do have some concerns about MMU2 but I do believe it will work better than the first version. My biggest concern is that it will be released too early. MMU1 underwent 6 months of beta testing and still didn't work for many users.

Peter

Re: MMU 2.0 video from MRRF

Agreed. Frankly I think this announcement is just putting a bucket of water on a roaring fire.

Re: MMU 2.0 video from MRRF

There will have to be two filament sensors; one at each end of the PTFE transfer tube. They have to detect the case where the filament happens to end inside the tube. There is literally no way to extract it at that point except to purge it through the extruder as waste. The only other option that comes to mind is to have a runout sensor for each filament, before the roller assembly, so that the filament could still be pulled back.

Maybe they can design a system where the filament won't get chewed up after running the same length back and forth a dozen times, but I would not dismiss the time involved.

The engineering tradeoffs for MM are difficult, for sure.

Re: MMU 2.0 video from MRRF

I wonder with that many inputs if it will have any of the "stuck filament" errors that people and myself have experienced.

Re: MMU 2.0 video from MRRF

There will have to be two filament sensors; one at each end of the PTFE transfer tube. They have to detect the case where the filament happens to end inside the tube. There is literally no way to extract it at that point except to purge it through the extruder as waste. The only other option that comes to mind is to have a runout sensor for each filament, before the roller assembly, so that the filament could still be pulled back.

Maybe they can design a system where the filament won't get chewed up after running the same length back and forth a dozen times, but I would not dismiss the time involved.

The engineering tradeoffs for MM are difficult, for sure.

well, we know how good those sensors are so they need to fix false readings first and make it reliable or ditch the laser sensor and use a proper encoder based one.

Re: MMU 2.0 video from MRRF

Maybe they just use lever switches on the top end ? If I understand the workings, there's no precision needed and timing is not critical. It only needs to know if filament moves into the tube or not.

Re: MMU 2.0 video from MRRF

How does this new MMU 2.0 get powered? We have been told it can be easily swapped from printer to printer - and since it can both be used for MK2 (12V) and MK3 (24V) - and due to that I don't belive som 5V can deliver power to the steppers - I looks like it will have it's own AC/DC power adapter?