There is no going back, true. I am also dreading the lack of Loctite and safety nuts. 
What do you mean? Didn’t you get any with the kit? 
1 Like
I effing called it: https://blog.prusa3d.com/development-diary-update-on-mmu3-for-the-mk4_85043/
The problem lies with the little strings and the detector not noticing the filament has been unloaded. I have been chatting about this in two threads on the bug report forums and as soon as they announced it would take longer I was sure it was this problem though they never got back to any of the posts in the forum. Hahahaha… 
At least I am going to get a voucher:
As a gesture of appreciation for your continued patience, we’ll send you an email with a $50 voucher next week. This applies to all MMU3 for the MK4 orders.
3 Likes
In the long list of things I don’t understand: why for heaven’s sake is grid infill the standard for input shaping? The rattling… Aaaargh. Change the standard infill for all profiles? Nono, can’t do that! No changing the standard profiles.
1 Like
Isn’t it a cheat for speedy prints by only printing alternative layers in alternative directions?
Somewhere I’ve guessed there are 50,000 Mk$'s in the wild give or take, so here’s my one in 50,000 story.
Tested the filament sensor yesterday, and it detected run out no problems at all, it just couldn’t work out that I’d removed the filament to enable me to load it.
Went through all the trouble shooting stuff, documented it all, then waited more than an hour for the chat line to free itself up and pasted the story complete with glossy 8 x10 photographs.
Over the next FIVE HOURS with customer service bloke in tow, I tested everything, pulled stuff apart, factory reset the firmware and loaded a virgin version of the latest and greatest which fixed the problem but introduced a fake fault with the x axis, so we dropped down a version by .0.0 and pulled stuff off again and put it back together again, which proved that it was all built correctly and magically the filament detector started working again, as it had done in every test prior, and the x axis wasn’t a problem any more.
So what we know now, despite all the evidence to the contrary, is that we never had a problem, but when we upgrade to the next firmware version, it’s going to render the machine unusable! 
Even though there are tens of thousands of machines using the next version!!
If someone could enlighten me as to how one machine can be allergic to one version of firmware - it’s like having a kid with a peanut allergy.
Delighted with the customer service, sad that I didn’t get to spend that time on the LR3.
Magic. Like, really. Why didn’t my screen work? They said in a blog post that most screens they got back were working fine, which I just can’t believe, because people were sent the other version and it worked without problems.
If there weren’t thousands working perfectly I’d be more skeptical-
It is doubtful (But surely possible) that it’s a firmware issue. All you did is shuffle an intermittent problem around until something that wasn’t working started working.
This seems more evidence for something intermittent being perturbed and all that really happened is you provided enough entropy to change the state of whatever is having trouble from bad back to good.
Impressive both that both their service agent and you had the patience and endurance to put that much effort in.
How interesting that it has no trouble detecting filament run out but struggles to detect filament unloading. I can see where that could be firmware, or some weirdness with the way the sensor works in the deliberate unload process.
Out of curiosity, what kind of fake fault?
1 Like
They had a blog post about it in connection to the MMU. Tiny strands of PLA block the sensor.
All credit to Prusa’s staff training and selection I say. It got to a point where we both had had enough, but in my case I would be left with an unusable machine so no point in giving up, and in his case he would be left with an unhappy customer or worse, a replacement machine on order I guess. He was very methodical, and didn’t take me over much previously trod ground.
That may have been helped partly because I did document the issue and what I’d done very clearly in the first place, and sent photos of the build - which he described as “nice”. I wonder what they see in that regard, when the assembly manual in describing how to get the frame level (it’s not hard) says don’t worry if you’ve got it within TWO mm!!
I am inclined to think that there’s some mechanical tolerance in my machine, I think the part that carries the magnet and ball in the sensor, that they are trying to cope with somehow in firmware. If the problem returns or if I could be bothered I might try making that part either as a super fine print or machined in aluminium, but this would be out of curiosity only as I am reluctant to make any mods (see numbers of other machines that work perfectly).
Ahh - the X axis sets up correctly in the setup - but fails the check immediately prior to commencing a print, resulting in a “crash detected” error and the thing just keeps flogging itself against the Z rod. It’s very much like watching a toddler have a tantrum - the banging of the head against the floor is funny the first couple of times but then it keeps going and it’s quite disturbing!
I am completely bemused at how this sort of error just effects my machine. My build tolerances there are zero, all I can think is that somehow the firmware has offset the x axis somehow. I didn’t think to try a different print file - I was using one of the standard pre-sliced ones that came with the machine and it’s possible that … well anything’s possible when you have a gremlin like this!
I’ve had that happen as well, but only twice with X but several times with Y. Hasn’t happened for quite a while though. Love your analogy. 
1 Like
Is that with the latest firmware version? How did you convince it that it was wrong?
It was with the 4.x and 5.0 Alpha. I actually never changed anything. I did not really pay attention whether it went away with the new screen or later. It happened when I moved the printbed by hand before/after a print, like it lost count to where it was. Sorry, this is pretty incoherent rambling, but it’s midnight and I have to sleep…
2 Likes
Six weeks in, and although I’ve been away for three of them, the Mk$ is copping a flogging and after it’s little hiccup at the start, is working flawlessly.
Zero failures, zero imperfect first layers. Astonishing.
1 Like
Still no MMU. 
Just buy one of those cheap Chinese ones!
(I suspect that there’s a bit of propping up of a certain company there - even for a Chinese product it’s too cheap… ) 
I’m jeaalous! Please let us know how it works! I’m tempted… but I don’t have the space. (atm…)
Will take a week to arrive. But I will surely keep you posted!
2 Likes
Jo P will be in Sydney next week, and since Google maps have told me that’s only 250 hours walk away, I’m off hopefully to see an MMU in action on their stand at the industry exhibition.
Would have been nice to have while printing that portrait instead of being on hand for 500 colour changes!! For all my disparaging remarks about colour changes and printing useless stuff - I am buying a big bottle of sauce and getting ready to eat my words!
On the other hand, I’m only about $7000 AUD away from buying an XL!
6 Likes