[Geschlossen] MMU unreliable, Prusa engineer are amateurs and Josef an fraud!  

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

While I can understand your frustration - I think you are wrong with your measured diameters. 1.93 / 1.91 - this is for me identical and not the root cause of the problem. I don't think it has to do with the hotend diameter nor the lower diameter of the multiplexer part above.

Check this thread and then you might understand the issue:

https://shop.prusa3d.com/forum/original-prusa-i3-mk2-multi-material-f26/monoplexer-and-multiplexer-redesigns-no-metal-part-t12801.html

Beside of that your language is rude. Please behave normal and don't get emotional.

John

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

I think he might mean the heatbreak inner diameter? If this is true it would explain the blobbing below the lower entry of the PTFE which many affected users see (including me). Its a very small ridge that very slowly grows until it blocks. Could be the guy packing up the kit just grabbed heatbreaks from the wrong box.
His observation (if accurate) would at least be consistent with all my issues, and would explain why Seth's printed multiplexer improves the situation (blob rubs off the inner diameter, so after a few prints it slides in smoothly). Also it means it would be easy to fix.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

While I can understand your frustration - I think you are wrong with your measured diameters. 1.93 / 1.91 - this is for me identical and not the root cause of the problem. I don't think it has to do with the hotend diameter nor the lower diameter of the multiplexer part above.

Check this thread and then you might understand the issue:

https://shop.prusa3d.com/forum/original-prusa-i3-mk2-multi-material-f26/monoplexer-and-multiplexer-redesigns-no-metal-part-t12801.html

Beside of that your language is rude. Please behave normal and don't get emotional.

John

Hello John,

I'm not wrong with my measurements if you measure the same thing. And these diameters causes problems (with my prints) that results in filament that would be stuck in the Steel Tubes. I had to disassemble the Festo fitting from the Y-splitter and take out the filament that was stuck. Then I heat the Steel Tube with my lighter otherwise I couldn't get the filament out of the Steel Tube. This happened to many times in my case!

On my case the heatbreak diameter makes a huge difference in reliability. Changing this to Ø1.9 mm gives me less problems. But care to explain the logic of making the Steel Tubes inside Ø smaller then the heatbreak inside Ø? Let my try to explain this:

So your filament goes threw the Steel Tube with a Ø1.93 hole, then it goes to the heatbreak that has a Ø2.1 hole, where it goes into a nozzle with a inside Ø1.9/2.0 hole (the diameter is different with each nozzle).

I actually have a nozzle in my possession with a Ø2.03mm hole. Let's say I'm Bad Luck Brian here because I just bought a new nozzle with a Ø2.03mm hole, when the filament is switched (pulled out of the nozzle), it goes into the Steel Tube that has a Ø1.93 hole. Could you explain how this would fit? I'm pretty sure you will end up with this https://imgur.com/gallery/Na8pn

A smaller heatbreak inside diameter could certainly make a difference here. So Bad Luck Brian didn't get a heatbreak with a Ø2.10mm hole, this is now Ø1.90mm. When you take out the filament the tip inside the nozzle is still hot, when it has to go threw the heatbreak forcing the hot PLA or ABS to change the diameter. Odds are that the filament now doesn't get stuck in the Steel Tube, but you are still asking for problems with a Steel Tube that has these diameters.

I'm actually an engineer, and I say that the engineers at Prusa made a rookie mistake. They didn't take into account that you have manufacturing tolerance, if they would have offered me a job as an Prusa Engineer then this would never happened! So when I say they are fucking amateurs, they are FUCKING amateurs!!!!!!!!!!!!!

Thanks for sharing that Topic, but I already seen and read it. I also 3D printed the part with Modified ABS and it makes a huge difference in reliability. It isn't perfect (yet) and from time to time it goes wrong, but way less then before. This is a really big gain!

I think he might mean the heatbreak inner diameter? If this is true it would explain the blobbing below the lower entry of the PTFE which many affected users see (including me). Its a very small ridge that very slowly grows until it blocks. Could be the guy packing up the kit just grabbed heatbreaks from the wrong box.
His observation (if accurate) would at least be consistent with all my issues, and would explain why Seth's printed multiplexer improves the situation (blob rubs off the inner diameter, so after a few prints it slides in smoothly). Also it means it would be easy to fix.

I got your PM but I couldn't answer because I'm a rookie here and they don't let me send PM's to other users.
But yes, I'm talking about the inner diameter of the heatbreak. I see now that I used the wrong name for a part, sorry. I will change my text so there is no misunderstanding in my explanation.

But I don't understand your explanation on the problems you have.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Thank you. What I mention is basically what you explained in your second post. The steel tubes won't budge, i.e. +/-0.1 mm makes a huge difference when you retract often. To clarify, I think this might be a combination of two problems:
The fully printed multiplexer will rub off a bit of material every time you pull filament through, so it would improve the resistance in the multiplexer itself.
The second problem is blockage below the PTFE inside the hotend, i.e. the heatbreak outlet has a wider diameter than the PTFE ID. See for example the post from today, the guy made a foto. I have made a replacement PTFE piece and of course also tried both parts I originally got - the results were absolutely identical (failure by blockage on exactly the same layer). I don't recall the ID of the PTFE but it is smaller than 2.0 mm (1.85?). If some unlucky person has a good fan and a 2.1mm bore I can see the filament might already be quite hard on PTFE entry and this would add to overall feed path resistance.
The solution would be what you implicitly suggest: send out 1.9mm bore heatbreaks with proper tolerances.
Actually, I just assumed this is what they meant here: https://www.prusaprinters.org/update-multi-material-upgrade-2/ ("custom heatbreak")

I tried systematically to change the print parameters to circle in the problem and found there is absolutely nothing systematic. Only since I have the modified MUX installed things started to make sense again and I could empirically improve my prints. A logical explanation would be a summation of two independent problems as I described. The real question remains why are we bothered with this 😡

Edit/afterthought: ...and how many end users can be expected to have a set of bore gauges around and the means to mill a heatbreak?
I suggest toning down your post title a bit. I think your results are very valuable and interesting, as it is now you couldn't really blame an admin if they delete it.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Maybe I should make a quick post to clear this up...

The heat break has 3 holes drilled; one from the bottom at 2mm and two from the top at 2.1 and 4.1mm

This is all made clear in the blockage troubleshooting thread and is by design. It is necessary for smooth(er) operation of the MMU.

In addition, the ID of the PTFE tube is (from memory) 1.85mm and the steel cooling tubed (again from memory (1.9mm)

Again, this is by design and is required for the MMU to operate.

There is a reason for all this. The cooling tubes can be replaced by PTFE tubing, but some types of filament may require more cooling movements.

The PTFE tubing needs to be the correct length, with both ends properly formed and inserted into the heat sink in the correct manner (you need to loosen the heat break as described in the thread mention previously). The PTFE tubing may require changing on a regular basis (frequency will depend on use and filament type) as the diameter will increase in use and the tip formation will become less effective over time.

I hope this clears a few queries up; it does not however, excuse the lack of help from PR on these forums.

Peter

If anyone requires further clarification, please PM me; I don't visit here very often but I do receive PM notifications.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

@bo.b Your explanations make no sense - plain wrong. Simple as that. Your language and attitude is rude and I'm wondering where you got your education from.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

@bo.b

This is not the way how to react on a forum. I do really understand your frustration. This forum is for helping each other, and yes I am also not happy with my MMU (see my reaction on this forum), but that does not mean I can spit out accusations and behave with disrespect.

That said:

Yesterday I got a sensor problem. The cable of the PINDA was broke. Thanks to this forum I could do the diagnostic and repaired it. Thanks forum!!

My MK2s is not one year in operation and this should not happen, but it happens. Just maintenance I told myself, as my car and all other stuff I have sometimes need (faster then I wanted 😀 ).

Still happy with my PRUSA!!
Frans

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Holy cow, where to even start. This thread blew up!

First, off, I need to get a couple of things out of the way that seem important to clarify now.

1. I'm going to ignore all the MUX stuff for now, because I've discussed it TO DEATH on my thread that has already been linked here.

2. This is the pilot release of the MMU, it hasn't even been in production a full year. Someone mentioned this on another thread, and I think it holds true, we are the pioneers of this technology. Prusa are the ones doing a lot of ground work, but ultimately, users like the ones in this thread are the ones who will fine tune the technology over time. Hell, 3D printing has a relatively short history still, and this kind of set up is still cutting edge. Even with a team of engineers, you can't account for everything. I'm sure that they had multiple prototypes that worked great before sending it to production, but as bob put it, there are manufacturing tolerances that need to be accounted for as well (while I might have put it a little more tactfully, the point is still valid).

3. Continuing my thought about manufacturing tolerances, this might not always be the fault of PRUSA, (i.e. they probably buy the steel tubes in bulk from a different manufacturer). All we can do is try to help make it a better product (no, it is not our job, but I like to think I've contributed to this awesome community by adding my ideas into it). As the first end users, we have taken a risk in buying the mk 1 version of the MMU, and that risk shouldn't be brushed off as all Prusa's fault.

4. No reason to be heated here, nobody on this forum is from Prusa, we're all end users, so lets just remember that when we are talking, we're not addressing Prusa, we're addressing each other, who have all had our own frustrations to be on this forum in the first place.

Now for the meat and potatoes:

Someone posted a picture on my thread on page two that puts a bunch of this into perspective, but for some reason, it won't upload for me at the moment, so I'll try to post it later.

To kind of paraphrase what has already been said, the heat-break steel tubing is supposedly a larger diameter (2.1 mm as measured by bob) than the steel tubing in the MUX (1.93 as measured by bob).

In order to get from the heat break to the steel tubing in the MUX, there is a piece of PTFE tubing, that is 1.9 mm. From what I understand, this is supposed to sort of help shape the hot/malleable filament on it's way from the heat break to the steel tubes in the MUX. If we follow the path, it goes from 2.1mm in the heat break, to 1.9 mm in the PTFE tubes, to 1.93 mm in the steel tubes (ignoring the plastic tolerances in the MUX). It seems like people have been having problems with the filament being too bulbous to even make it through this PTFE tube though, and it gets stuck at the bottom of the PTFE tubing.

You'll notice I haven't mentioned the nozzle diameter. If I'm correct, the 2.03 mm input into the nozzle is smaller than the heatbreak, and therefore shouldn't be causing any of the blockage problems. The wide point here is the heat break, which has been measured as 2.1 mm.

Now, I'm going to attempt a solution that may fit everyone's needs.

I'm going to assume that you have replaced the MUX at this point, since it has it's own set of extensive issues, so keep reading with the implication that the MUX has already been replaced with the redesigned version that I have posted. I would also like to mention that in order to get the redesigned MUX to work properly, I already had to increase the diameters of the feed paths, and the filament paths need to be broken in.

With all that in mind, the problem discussed in this thread seems to be with the PTFE tube directly above the heat break and directly below the MUX, or with the heatbreak diameter itself being 2.1 mm, not allowing for smaller cooled filament diameters. Since replacing the heatbreak with a smaller diameter seems unlikely without a machine shop, I'm going to focus on the PTFE tube.

If you were to drill out the center of the PTFE tube to say 2.0-2.05 mm (keeping the shape of the bevel in the bottom and top ends), it would give you a whole extra .1 - .15mm of tolerance to help prevent jams here. The only thing this might affect down the road is retracting the filament all the way into the long tubes since they should still have a 1.9mm diameter (I say this without measuring, I don't know the actual ID of the upper tubes, they could very well be larger, I'm just assuming it's the same as the lower end). Worse case scenario, it jams during filament unloading and you have to remove some of the top tubes to clear it. The end goal here will probably be making the feed paths taper from larger to smaller going from the extruder to the heat break.

I think it would be a bad idea to go 2.1 mm or more ID for the PTFE tube above the heat break since that would only make it easier to have more bulbous filament during retraction, but in the end, that may be the direction it needs to go.

So to summarize, try drilling the PTFE tube to 2.0 first, replace the MUX with the redesign, and try it from there.

Try all of this at your own risk. I'm just throwing out an idea that may or may not work. Keep in mind I do NOT have this problem, so I can't test it. I'm just hoping that it will help some of the people having jams in this portion of the printing process.

Sorry for the lengthy post, just trying to get a lot of info out there all at once.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Seth I agree with most of what you say, but not point 2.
Users having problems now probably ordered their units in July 2017 or later. By then the upgrade had been available for half a year. What I did back then: Watched all the youtube vids, read the articles and reviews, checked the forums. There was absolutely nothing negative, everybody seemed happy. I usually assume that youtube reviewers get hand-polished, massively tested units. Also I assume 50-75% of the marketing statements are outright lies that none of the developers would agree with. Even then I was left with the impression that if there were issues I could work them out in 1-2 weeks with the tools I have available and that there would be some helpful support by PR guiding me in the right direction. But then again, many customers would just watch the reviews, then order and feel cheated.

I'd very much like to be a space pioneer, but definitely not a pioneer in fixing a commercial product. If they had called it "pilot release" on their website, I would have stayed away. At this stage I think PR would be well advised to come out and tell the community what they have been doing to adress these issues, with a big apology. Not because there are problems, but because they are handling it terribly. Most affected users have voiced no unreasonable demands. I totally agree there is no reason to insult anyone, but posts like Bobs will happen again if they keep damaging their reputation. I cannot blame him at all if this is his only printer and/or he had to save up the money instead of just buying cheap chinese. There have been times when I had to calm myself down and not throw the printer through the window. Many 3D printing enthusiasts have been burned by companies going bust, bad quality products etc., making them at least a bit careful and sceptical. If that kind of customer spends 5 minutes browsing this forum they'd stick away. Everybody wants to print, not fix their printers.

Don't get me wrong, I'm happy for anyone who got a working unit. We have absolutely no idea how many users are affected. It could be 5% (great) or 90%. But then just check out the number of quiet readers e.g. on the blockage thread...I don't believe all of these have resolved their problems. A relatively cheap approach would be an "official" thread describing a systematic algorithmic procedure to set up and test, ending in -> "it works" or -> "contact support to get spare part X". PJR has tried this, but without support from PR it is pointless. And it assumes manufacturing tolerances are the reason for problems.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Hi all.

I can only encourage all of you/us to continue the communication and to share observations and recommend solutions. To some extent I agree to all positions stated - even the harsh ones. I am not an expert engineer like some of you and I rather want to be happy about successful prints instead of successful trouble shooting sessions :-)-

Anyway - on the point of my printer's readiness. I got the MK2S early 2017 (after running a Mendel90 for almost 3 years and feeling like a pioneer for many times) and upgraded to MMU X-mas. The MK2S worked perfectly, though I did not really like the design of the XY axis in the beginning.
Now the MMU upgrade is creating a challenge - absolutely. When doing the assembly, I ran into exactly the same problems as all other users (you can see this in the online manual at these steps, where the comments increased suddenly. In retro I would say that the too long PTFE tube of the hotend and the recommended shortening and necessary reaming were the main failure (and disaster from a quality control point of view). Somehow I found a spare PTFE tube in my kit and that had the correct length from the start so I did not have to ream the upper entry. And with this one, feeding problems are gone (for the time being).
I did buy the MMU kit as I wanted to work with soluble filament for structures. The PRIMASELECT filament I got has not worked out until now. After 3 or 4 prints (which I always stopped after the first layers), I decided to put this on hold and continue with "normal" filament until I new that the mechanical structure of the printer is not at the root cause of the PRIMASELECT problem.
Anyhow - 2 weeks later I am 2 weeks wiser as well. Frustration has not yet won - I would say that I am still breaking in the MMU. Single color prints are working now reliably - on 3 out of 4 extruders. I did print Seth's new splitter but have not yet exchanged the original one.
The first dual color print is running now (4 little Marvins are printing as we speak with filament from PRUSA (orange) and Filamentworld (snowwhite). The first 10 filament changes have taken place now and the Marvins start to become visible. It'll take another 2-4 hours I suppose before the print will be finished buy I'll give the update as soon as I am done. So far it is looking ok - I cannot yet complain.

For your info - in parallel I am in touch with one of the PRUSA support engineers and the forth and back is working (it usually takes 2-3 days between mails).

So - this is just for your info, as I have mixed feelings about the product readiness of the MMU myself. 329 USD is quite a bit of money and upgrading your printer from an excellent workhorse to a non-working joke is nothing I want to experience. So, some of the feedback has to be taken seriously by readers and hopefully PRUSA departments that are listening to this.
Crowd competence and crowd troubleshooting is efficient and can help - but they need to listen.

BR, H

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

I agree with what you say Nils.

The main take away point is that these forums don't reach PR employees about 99% of the time.

I totally understand that frustration, hell, I even started my own Prusa hate thread when I first started diagnosing my problem. But did any PR representative ever get back to me? No. Lol. But you know who did? All of the wonderful people here.

Just so you guys know, I was able to get ahold of Jakub also. We will see what happens but I really wouldn't hold my breath.

BTW, just to clarify my point a bit, I'd strongly prefer a product ready and reliable out of the box. I was just trying to point out that there is always another side to it, and I was more trying to throw a little devil's advocate into the mix. $300 for a failed upgrade is the worst, no doubt.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

It would be fine if these threads never reached the employees at Prusa *IF* the link to this forum wasn't hanging off their 'Support' link on their home page. This doesn't *feel* like a Prusa support forum,. it feels like a 3rd party forum dedicated to Prusa printers.

Not a place to have support questions answered by Prusa. It's 100% community driven support - which, again, is fine, except for it being linked to from their e-shop.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Well I spent some time browsing the MK3 forum tonight, and my remaining hopes for the MMU dropped into a bottomless pit.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

I'm back with some good news guys.
I did some more research and reading because of a few comments here. I was wrong on a few things I mentioned before but I got it figured out now and I can also show it.

So I reprinted the Y-splitter (original). I used support and I did this with a 0,25mm nozzle (0.15mm layer height). I used the filament, TitanX from FormFutura with the Atlas Support as support. It didn't go well and I had to stay with the print for the first 4mm to correct any issues while changing the filament. After that there was no support so the print went well.
The reason for the small nozzle is because of the small details in the Y-splitter. In my opinion you need the Y-splitter as smooth you can print with high accuracy. I think that PLA would be a wrong material choice and I would not recommend it, the TitanX is a modified ABS where I managed to get a really smooth surface (almost Aceton finish smooth) where the material is also very strong and with a high softening temperature.

So I made this part in Solidworks and this are the dimensions I used. If you are going to replicate this I have to warn you for problems that will occur and I will show you engineering wise how foolish they are at prusa.
This part is a replacement for the Steel Tube, I made it out of brass because it's easy to machine and it has a better heat transfer then SS (stainless steel). Also I can drill the hole really smooth because of the material characteristics.
With milling there are a few problems, the hole of 1.90mm is not easy to drill and you have to take the time to prevent breaking. I suggest drilling the hole first on 1 side and use a center drill beforehand. But with brass this is relatively easy and an experienced miller shouldn't find this difficult.

As you can see I drilled the hole (i managed to get the hole 1.89mm on all the four parts). I also machined it down to 3.00mm on the outside on a length of 23.50mm. But because you have a 0.55mm wall thickness the tip of the tube will bend while machining, so my tip was 3.34mm and I corrected this with a small angle (around 5 degrees) to 3.00mm
I assume that this is a bit hard to understand, but let me explain this. You wanna avoid a 0.55mm wall thinkness with machining because it's very difficult and also expensive to produce in big numbers. If you would have a wall thickness of 1.55, you get a outside diameter of Ø5.00mm (I would like to see Ø6.00mm) then you have something that is easy to machine and produced in big numbers.
I also think there is a big gain in a thicker wall thickness when you 3D print. There is more material, so when switching filament you cool down the tip more and the brass will not heat up as much. And theoretically speaking you can speed up the switching filament process and gain time in your printing process.

Because I used a Ø10mm brass I machined it down to Ø7.00mm and I drilled a M5 screw thread inside for the Festo fitting. Again a big gain here, you have more material to distribute heat and there is no play on the brass tubes. Where with the Steel Tubes fitted I had a huge gap and you could hear it rattle inside the Y-splitter. In my opinion a prone to problems when printing.

I also made a small change to the heat-breaker. The inside diameter was first 1.90mm, this is changed to 2.00mm (i got this 2.01mm). The idea is as following: When the filament goes out of the nozzle the heat-break will force the filament to 2.00mm, when it goes into the brass tube it again be forced to change to 1.90mm diameter where it will cool down and shrink. Now the filament will always fit in the heat-breaker because the tip is always smaller! This will result in a reliable printing machine.

So far I did 2 test prints today. I printed a magnet holder where I use 2 different coloured PLA and PVA for support, I printed this before so I know the problems that I had before. But this print went without any problems, like zero problems! And I'm not exaggerate when I say that this never happend before.
The second print is still printing as we speak, but at 4mm layer height still no problems. Here comes the best part, I used again the TitanX with PVA and it works without any problems. I'm printing the TitanX with 258 degrees and the PVA with 210 degrees, a huge difference! This is outstanding considering that I read nobody managed to got this working on a MM prusa. Ofcourse I'm going to make a picture tomorrow of the result as proof.
I have to do more test, but at this point I'm more then satisfied with the results.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Glad to see you're having some success. Maybe this is a case of what works in the prototype lab isn't scaling well in manufacturing.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Hi Bob,

impressive - you are lightyears ahead of me when looking into the problem and doing the additional engineering. Let us know how the second print is coming out - will be interesting to see.
In the meantime I continued to do my printing. What shall I say - I printed 4 Marvins with dual colors last night and this morning - and the result is visible on the pictures.
What do you think? I am very positively surprised, as it shows that the printer is able to produce a certain result - reliably. Both prints took about 3h5min.

.
The result is consistent (including the small faults). The feeding of the filament was no problem.
I will now add the 3rd extruder - I am still working with the original splitter.
BR

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

I'm back with some good news guys.
I did some more research and reading because of a few comments here. I was wrong on a few things I mentioned before but I got it figured out now and I can also show it.

So I reprinted the Y-splitter (original). I used support and I did this with a 0,25mm nozzle (0.15mm layer height). I used the filament, TitanX from FormFutura with the Atlas Support as support. It didn't go well and I had to stay with the print for the first 4mm to correct any issues while changing the filament. After that there was no support so the print went well.
The reason for the small nozzle is because of the small details in the Y-splitter. In my opinion you need the Y-splitter as smooth you can print with high accuracy. I think that PLA would be a wrong material choice and I would not recommend it, the TitanX is a modified ABS where I managed to get a really smooth surface (almost Aceton finish smooth) where the material is also very strong and with a high softening temperature.

So I made this part in Solidworks and this are the dimensions I used. If you are going to replicate this I have to warn you for problems that will occur and I will show you engineering wise how foolish they are at prusa.
This part is a replacement for the Steel Tube, I made it out of brass because it's easy to machine and it has a better heat transfer then SS (stainless steel). Also I can drill the hole really smooth because of the material characteristics.
With milling there are a few problems, the hole of 1.90mm is not easy to drill and you have to take the time to prevent breaking. I suggest drilling the hole first on 1 side and use a center drill beforehand. But with brass this is relatively easy and an experienced miller shouldn't find this difficult.

As you can see I drilled the hole (i managed to get the hole 1.89mm on all the four parts). I also machined it down to 3.00mm on the outside on a length of 23.50mm. But because you have a 0.55mm wall thickness the tip of the tube will bend while machining, so my tip was 3.34mm and I corrected this with a small angle (around 5 degrees) to 3.00mm
I assume that this is a bit hard to understand, but let me explain this. You wanna avoid a 0.55mm wall thinkness with machining because it's very difficult and also expensive to produce in big numbers. If you would have a wall thickness of 1.55, you get a outside diameter of Ø5.00mm (I would like to see Ø6.00mm) then you have something that is easy to machine and produced in big numbers.
I also think there is a big gain in a thicker wall thickness when you 3D print. There is more material, so when switching filament you cool down the tip more and the brass will not heat up as much. And theoretically speaking you can speed up the switching filament process and gain time in your printing process.

Because I used a Ø10mm brass I machined it down to Ø7.00mm and I drilled a M5 screw thread inside for the Festo fitting. Again a big gain here, you have more material to distribute heat and there is no play on the brass tubes. Where with the Steel Tubes fitted I had a huge gap and you could hear it rattle inside the Y-splitter. In my opinion a prone to problems when printing.

I also made a small change to the heat-breaker. The inside diameter was first 1.90mm, this is changed to 2.00mm (i got this 2.01mm). The idea is as following: When the filament goes out of the nozzle the heat-break will force the filament to 2.00mm, when it goes into the brass tube it again be forced to change to 1.90mm diameter where it will cool down and shrink. Now the filament will always fit in the heat-breaker because the tip is always smaller! This will result in a reliable printing machine.

So far I did 2 test prints today. I printed a magnet holder where I use 2 different coloured PLA and PVA for support, I printed this before so I know the problems that I had before. But this print went without any problems, like zero problems! And I'm not exaggerate when I say that this never happend before.
The second print is still printing as we speak, but at 4mm layer height still no problems. Here comes the best part, I used again the TitanX with PVA and it works without any problems. I'm printing the TitanX with 258 degrees and the PVA with 210 degrees, a huge difference! This is outstanding considering that I read nobody managed to got this working on a MM prusa. Ofcourse I'm going to make a picture tomorrow of the result as proof.
I have to do more test, but at this point I'm more then satisfied with the results.

Sooooo.... how much for the brass tubes??? LOL

Would it be of more help if the PTFE tube was swapped for metal, and perhaps even something with a narrowing inside diameter to further help resizing the filament?

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Narrowing inside diameter is a nice idea, but how do you want to machine it...

If the problem is indeed wobbly or poorly centered metal tubes, maybe something like this:
https://www.landefeld.de/artikel/de/reduzierung-6mm-stecknippel-x-4mm-schlauch-iqs-standard/IQSG%2060H40

We use the IQS system at work, it is quite nice. 6mm OD is the smallest I could find in the catalogue, so probably 4mm ID. If it were 3mm ID the metal tubes could fit directly. However, the metal tubes are not the bit that is hard to get/make. One would have to redesign the MUX so that the IQS part can be press-fitted. Stick the metal tube inside the IQS part, push IQS part in MUX, done? At least this way there would be absolutely no movement in any direction and the metal tube could be pushed directly against the PTFE in a controlled way.

Re: MMU unreliable, Prusa engineer are amateurs and Josef an fraud!

Guys, Peter just approached me with a very interesting idea - using PTFE instead of metal tubes, since it could fit directly into the part I linked. Went rummaging and found something similar (see pic, the pushfit is too big and it's angled but from the same manufacturer so the ID should be identical). It is exactly 4mm on the inside. With this and a reamer tool, the tubes could be self-built, and I think the printed part can be redesigned to accomodate that assembly. I'll take it home and try to stick a 4mm PTFE tube in, see if it wobbles around.