Prusia i3 Mk3 Multi Material @ MRRF 2018  

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

Doesn’t sound like hard sw (says the guy not writing it). All this occurs “offline”, ie not printing. Timing is not critical. You could fully handshake each step and still work out fine.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

Is there a reason we have not yet seen this thing running on an actual machine yet? hmm?

Ummm, because it's not reliable yet?

Given that they're not going to start shipping them for months, my assumption is that it's still under heavy development and there's still kinks to work out.

MMU will start shipping next month.

I hope you're right, but I'm going to keep my expectations tempered. Prusa is famous for many things, but meeting deadlines is not one of them.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

I'm trying to figure out why everyone seems to think retraction is needed.

Wouldn't it just make more sense to cut the filament when the appropriate length between the cutter and what's left to be extruded is reached in the gcode.

1) Cut filament
2) continue printing until the sensor (near the extruder says it's empty)
3) select new filament and feed that until that sensor reads it has filament
4) turn the revolver to the slip all position so that the feed is only happening from the extruder bondtech
5) print purge block or infill as needed
6) rinse and repeat.

So during the time between 1 and 2 the tube between the multi material switcher and the extruder would become empty.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

I don't think it's so easy. If you cut filament and continue to print it's end will come past primary gears and stop extruding at the bottom, leaving some of the filament still inside top part of extruder(cooler), causing the clog. I think that extracting back is the only way to go.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

I'm trying to figure out why everyone seems to think retraction is needed.

Wouldn't it just make more sense to cut the filament when the appropriate length between the cutter and what's left to be extruded is reached in the gcode.

1) Cut filament
2) continue printing until the sensor (near the extruder says it's empty)
3) select new filament and feed that until that sensor reads it has filament
4) turn the revolver to the slip all position so that the feed is only happening from the extruder bondtech
5) print purge block or infill as needed
6) rinse and repeat.

So during the time between 1 and 2 the tube between the multi material switcher and the extruder would become empty.

Cause they said so. 😛

The thing is, unless the multimaterial unit is doing 100% of the "measuring" (aka: MMU for mk2) the MMUv2, will not know where to cut the filament precisely.
Your concept would have the need of a MUCH larger purge block... That takes time... *AND* filament.

1. To do the "cut filament", how do you propose it is measured? What if your bowden tube is 5mm too long? 5mm too short? 25mm too long! Your system would have to be re-calibrated by an "Engineer" minded person. Sadly, not everyone is born an "Engineer". Your system would then have to be "smart" to rule out the "engineer" needed to calibrate it. And as I mentioned above. "smart" is *BAD*.

2. And then the PETG/Clear/Green filament throws a false alarm, becuase the "SMART" optical sensor, screws up.

3. Now your MMUv2 attemps to ram 200mm of PLA down the bowden tube with only 50mm of free space, causing the gears to grind a hole in the side of the filament about 50mm from the end. Your extruder gear for that filament likely needs to be taken out and cleaned before it will work properly again.

4. This will work.

5. There is a high likelyhood of a jam happening once the "ground" bit of filament gets to the bondtech, as it may not be able to feed it. At this point everything goes to hell, and you pray the printer throws an alarm and turns itself off.

6. Assuming the printer don't turn itself off, you get to spend the next 1-2 hours, cleaning gears, and replacing the nozzle (ever see what happens when you leave the hot end on for 6+ hours, without extruding? Just replace the nozzle, quality ones are cheap. (about 10$)

Your system also requires each "strand" be minimum ~225mm in length, which is rather... long.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

Aside from multiple mechanical challenges and limitations of cutting ahead of the material change, I am thinking of software limitations. I think it would be ideal for the new mmu to be operable with the slicer of your choice, so ideally a tool change script and a purge block combined should allow that if you do all filament changing activities at the time of the filament change instead of looking ahead to cut at the right spot. But I don't know that prusa has this viewpoint since they are able to provide a slicer and post processor that can handle whichever way they want to do it. I would just like to see us move away from the need for using just one toolchain.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

These are all good points I hadn't thought of. I am really curious to see this in action.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

The Palette+ filament splicer is sort of like this, in that it attempts to cut the filament and splice in exactly the right length segment.

I think the retract and reload approach is likely to be more reliable. From my experience, even with some fancy tricks and careful measurement of the filament path, the Palette sometimes gets out of sync with where the printer is in the print. The advantage of retract and reload is that the printer will know exactly when it needs to change filament. It can get the new filament loaded at the precise moment when it's needed and not have to guess at the amount left in the feed path.

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

My understanding operation.

Filament #1 is loaded. Call for Filament 3 to be loaded.
Bondtech on printer AND MMU both "retract" on filament #1, once past optical sensor on X carriage, speed up retraction, and retract until MMU filament sensor no longer detects filament.

But the filament sensor near the hotend is a bit limited at the moment right? So unless Prusa Research has a better filament sensor up their sleeve, the MMU will be unable to handle quite a few materials from what I understand in this thread and what I've heard about Prusas filament sensor in reviews (I don't have my mk3 yet).

There is a big difference between having a sensor working for some materials in regards to printing where you can still print with the sensor off, versus having a sensor working for some materials for auto-loading from the MMU. For printing it's a nice to have a sensor that saves material, cleaning and possibly the print if there is a failure. For auto-loading it's a "no go" with quite a few materials that you can't use a workaround for.
Some translucent filaments focus more light on the sensor and this can result in weird behavior. Certain colors like ivory, matte white or lime green
(From: https://help.prusa3d.com/l/en/article/UaW5ZNPYR8-filament-sensor )

Re: Prusia i3 Mk3 MuliMaterial @ MRRF 2018

My understanding operation.

Filament #1 is loaded. Call for Filament 3 to be loaded.
Bondtech on printer AND MMU both "retract" on filament #1, once past optical sensor on X carriage, speed up retraction, and retract until MMU filament sensor no longer detects filament.

But the filament sensor near the hotend is a bit limited at the moment right? So unless Prusa Research has a better filament sensor up their sleeve, the MMU will be unable to handle quite a few materials from what I understand in this thread and what I've heard about Prusas filament sensor in reviews (I don't have my mk3 yet).

There is a big difference between having a sensor working for some materials in regards to printing where you can still print with the sensor off, versus having a sensor working for some materials for auto-loading from the MMU. For printing it's a nice to have a sensor that saves material, cleaning and possibly the print if there is a failure. For auto-loading it's a "no go" with quite a few materials that you can't use a workaround for.
Some translucent filaments focus more light on the sensor and this can result in weird behavior. Certain colors like ivory, matte white or lime green
(From: https://help.prusa3d.com/l/en/article/UaW5ZNPYR8-filament-sensor )

I think I am planning on getting this addition but I can confirm that the filament sensor is a real hit and miss on a lot of filaments. I use a lot of black PETG from a few vendors and so far none have been able to reliably use the sensor. I remember reading a post that there are some vendors that have more of a matte finish that work but who wants to give up a reliably proven filament that has been used for a long time to make this work. I do hope over time that the sensor issue will be figured out to better support all filaments especially if the MMU upgrade requires it to work reliably.

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

Technically it can be done without the sensor... the sensor is convenient but if they assume filament isn't slipping and it's actually loaded it will work without one...

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

I have not read the whole thread yet, but in the future, maybe there should be a small purge platform which is driven so that its next layer is always kept on plane with the active print layer. That way extraneous material would not need to be printed just to keep the purge plane level with active print plane.

I just thought of this on the spur of the moment so maybe it is half baked.

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

I have not read the whole thread yet, but in the future, maybe there should be a small purge platform which is driven so that its next layer is always kept on plane with the active print layer. That way extraneous material would not need to be printed just to keep the purge plane level with active print plane.

I just thought of this on the spur of the moment so maybe it is half baked.

That sounds like a $50 way to save $0.05 of plastic, to be honest.

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

The thought was in response to some reviews where I have seen purge towers that greatly outsize the small part (or even large part) that they were co-created with. Some materials are expensive and a large purge tower might not be financially wise. Also, a purge tower takes time to create, no? I dont know the delta in time to create a model versus a model with purge tower, but I assume it is significant enough to make the extra $50 worth it in the end maybe?

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

When you unload a filament, there is almost 50mm^3 melted filament left behind in the hot end which needs to be purged out.

Coupled with the fact that you are feeding a 1.75mm diameter filament into a "pot" of melted filament that is 2mm diameter, the actual purge needs to be significantly greater than 50mm^3 - in some cases as much as 250mm^3 (or more) for each filament change.

Therefore the purge tower must be quite large. And when you are looking at variable (or low) layer heights, the purge tower can be huge.

The purge tower must also be fully supported - even when there are no filament changes in the lower layers. This is achieved with a sparse fill tower.

Actually, given that the user knows about the waste, the waste volume is down to the user (model design, size and number of models on the bed).

Peter

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

The thought was in response to some reviews where I have seen purge towers that greatly outsize the small part (or even large part) that they were co-created with. Some materials are expensive and a large purge tower might not be financially wise. Also, a purge tower takes time to create, no? I dont know the delta in time to create a model versus a model with purge tower, but I assume it is significant enough to make the extra $50 worth it in the end maybe?

The purge tower isn't necessarily solid. On layers where no purge is needed, the purge tower might be as little as 5% infill. Since it's square, those parts can print fairly fast.

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

When you unload a filament, there is almost 50mm^3 melted filament left behind in the hot end which needs to be purged out.

Coupled with the fact that you are feeding a 1.75mm diameter filament into a "pot" of melted filament that is 2mm diameter, the actual purge needs to be significantly greater than 50mm^3 - in some cases as much as 250mm^3 (or more) for each filament change.

OK, quick calculation. I bought the MMU specifically for soluble support. So let's say I print a 150 mm figurine with 0.1 mm layer height. This makes 1500 Layers. If I have support material up to the top (not totally unrealistic for an organic sculpt), this means 1500 filament changes. At 50 mm^3 per layer change this makes 75000 mm^3 or 75 cm^3 of material. Five times that if you purge 250 mm^3. BVOH has a density of 1.14 g/cm^3 so 75 cm^3 gives 85.5 g. At Fr. 122/500g (price at 3djake) this makes 20.74 Fr. of wasted BVOH material for a single print. Assuming 250 mmm^3 purge we get 103 Fr, alsmost a full spool of super expensive BVOH wasted. At 10 mm^3/s extrusion flow the 75000 mm^3 give 2 hours just for extruding the BVOH waste, another two for the build material. Five times that for 250 mm^3 purge.

Any error in my calculation?
This is ridiculous! I hope PR sees this also and comes up with a dual hotend setup (like UM3, toolchanger, whatever).

- Martin

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

When you unload a filament, there is almost 50mm^3 melted filament left behind in the hot end which needs to be purged out.

Coupled with the fact that you are feeding a 1.75mm diameter filament into a "pot" of melted filament that is 2mm diameter, the actual purge needs to be significantly greater than 50mm^3 - in some cases as much as 250mm^3 (or more) for each filament change.

OK, quick calculation. I bought the MMU specifically for soluble support. So let's say I print a 150 mm figurine with 0.1 mm layer height. This makes 1500 Layers. If I have support material up to the top (not totally unrealistic for an organic sculpt), this means 1500 filament changes. At 50 mm^3 per layer change this makes 75000 mm^3 or 75 cm^3 of material. Five times that if you purge 250 mm^3. BVOH has a density of 1.14 g/cm^3 so 75 cm^3 gives 85.5 g. At Fr. 122/500g (price at 3djake) this makes 20.74 Fr. of wasted BVOH material for a single print. Assuming 250 mmm^3 purge we get 103 Fr, alsmost a full spool of super expensive BVOH wasted. At 10 mm^3/s extrusion flow the 75000 mm^3 give 2 hours just for extruding the BVOH waste, another two for the build material. Five times that for 250 mm^3 purge.

Any error in my calculation?
This is ridiculous! I hope PR sees this also and comes up with a dual hotend setup (like UM3, toolchanger, whatever).

- Martin

Yeah, but you are forgetting the 10 linear mm (24 mm^3) pressure cool extrusion before unloading... And that you really need a largish purge after PVA, otherwise the PLA will be a little porous after dissolving the PVA.

I am currently doing some tests on this; will be posting result in a day or two.

Peter

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

Let's hope PR listens to you. I don't realistically see me buying a UM3 or UM5S - just for fun.
On the other hand, I don't see the MMU1/2 concept work for PVA/BVOH with a realistically sized print.
From what I've seen now, the MMU seems more a "proof of concept" for multi material print than being of any realistic use, particularly with expensive soluble support material.

- Martin

Re: Prusia i3 Mk3 Multi Material @ MRRF 2018

Martin

It really does depend on the user as to whether the MMU (and filament wastage)is worthwhile.

I have 2 Mk2/S + MMU and 1 Mk3, waiting MMU2. I have been using MMU for almost 17 months.

I have had many highs and just as many lows with the setup, but when it works (which is currently almost all the time), it really does produce excellent prints which cannot be generated in any other way - at least not any way approaching this price range.

Peter