Super Clogged Extruders

[brightorange 3]

hi again 

 

@SandervG thanks for forwarding and working towards more transparency regarding this issue.

 

they were kind enough to call it a "coin" but it felt like a medal

 

regarding solutions - If UM is smart enough to encourage community projects using the camera for diagnostics, I am optimistic such problems can be first candidates  ( from now on to be known as "spaghetti classifier" and "blob classifier")

 

thx for all responses , waiting for updates

 

 

 

[geert_2 557]

If you have 1 in 15 parts coming loose, then I would suggest you review your bonding method: there is a lot of room for improvement. Improving your bonding method until it is 99.99% safe will give you a lot more peace of mind; it is definitely worth the effort.

 

I have maybe 1 in 2000 parts coming loose, even though my models are not always easy to print, and even though I don't particularly care about cleaning the build plate.

 

[brightorange 3]

I believe you that you may have 1 in 1000 prints come loose , but consider this:

1. not all users are knowledgeable. beginners may send things that are not recommended to print. these prints may have much higher tendency to cause problems. 
2. when I have 1 in 15 coming loose is when working with CPE and nylon. 

say you are a school sending beginner-level projects, or a lab working with nylon/CPE/other, so you have 1 in every 50 prints come loose. let's say this severe clogging problem occurs only once out of 3 times the print comes loose. if you send 1 print every day on average, you will have a failure every 6 months on average. this is still not acceptable! now imagine you have a lab with 6 new shiny printers.

You will have , on average, this severe problem every single month! every month an argument with your dealer who'se fault it is, and maybe a charge of $200 or more. if your printer is out of warranty, you are looking at $500-600.

 

 

[kmanstudios 1,120]

2 hours ago, brightorange said:

I believe you that you may have 1 in 1000 prints come loose , but consider this:

1. not all users are knowledgeable. beginners may send things that are not recommended to print. these prints may have much higher tendency to cause problems. 
2. when I have 1 in 15 coming loose is when working with CPE and nylon. 

say you are a school sending beginner-level projects, or a lab working with nylon/CPE/other, so you have 1 in every 50 prints come loose. let's say this severe clogging problem occurs only once out of 3 times the print comes loose. if you send 1 print every day on average, you will have a failure every 6 months on average. this is still not acceptable! now imagine you have a lab with 6 new shiny printers.

You will have , on average, this severe problem every single month! every month an argument with your dealer who'se fault it is, and maybe a charge of $200 or more. if your printer is out of warranty, you are looking at $500-600.

 

 

This would be a false equivalency based on this logic:

 

1. Schools should have somewhat experienced or supervised people running printers.

2. You cannot take a professional level machine and dumb it down to the lowest common denominator and still have it be a professional machine.

3. Having been around school printings (3D and otherwise) there are a variety of printers available, and not all one type or brand. These things evolve as well as the capabilities of the department they are functioning within.

 

What you are asking is kinda like those knuckleheads that have too much money to burn, buy a Lamborghini and wreck it because they cannot handle the power. It ain't a frikkin' KIA. LOL

Edited March 7, 2018 by kmanstudios

[engr 2]

On 3/4/2018 at 8:28 AM, SandervG said:

I assume @brightorange, @wydmynd and @rjd are different users right?

I'm curious to know this too, haha!

[geert_2 557]

When we were back in school, looooong ago :), we were not allowed near critical lab equipment unless we passed a test on how to use it. So, anyone who was not prepared well for the lab session, could immediately return home, and could "come back in september". This tradition worked well.

 

For the lab technicians and supervisors, I would recommend: before using any new equipment for serious work, first play around with it until you know it down cold. In case of a 3D-printer: have fun printing keychains and all sorts of little toys until you have your bonding methods, temperatures, and all other critical material settings correct. This playtime pays off really well. If you don't spend time playing around first, you will spend 10x more time while wasting real models later on. This is sort of a law of nature. But of course no one believes this, so they neglect the playtime, and they spend 10x more later on... 

 

For getting the bonding-methods correct, a little testprint like this might help: if you can print this thing well, without any corners lifting, the majority of normal prints will also print well. And if this test fails really soon and produces spaghetti, you are likely to have bonding failures in other models too. Feel free to use this one, or to derive your own version of it. Be sure to stay with the printer while doing this test.

 

warptest8.stl

 

 

[brightorange 3]

@engr I am @wydmynd too. I have reason to do so. I am not rjd. no conspiracy here. 

[kmanstudios 1,120]

5 hours ago, brightorange said:

@engr I am @wydmynd too. I have reason to do so. I am not rjd. no conspiracy here. 

Speaking for myself, and, I for one am unanimous in this, That is weird...being two different accounts with the same person is just odd.

[gr5 2,167]

23 hours ago, brightorange said:

I believe you that you may have 1 in 1000 prints come loose , but consider this:

1. not all users are knowledgeable. beginners may send things that are not recommended to print. these prints may have much higher tendency to cause problems. 
2. when I have 1 in 15 coming loose is when working with CPE and nylon. 

say you are a school sending beginner-level projects, or a lab working with nylon/CPE/other, so you have 1 in every 50 prints come loose. let's say this severe clogging problem occurs only once out of 3 times the print comes loose. if you send 1 print every day on average, you will have a failure every 6 months on average. this is still not acceptable! now imagine you have a lab with 6 new shiny printers.

You will have , on average, this severe problem every single month! every month an argument with your dealer who'se fault it is, and maybe a charge of $200 or more. if your printer is out of warranty, you are looking at $500-600.

 

 

You make good points.  We are all guys responding on this thread which means you probably will get "solutions" and not sympathy.  Sorry.

 

One solution is to switch to Ultimaker 2s.  

 

Another is to not let students print anything other than PLA. 

 

Waiting for UM to come out with a video algorithm to detect blobs is not a solution.  I suspect that will take more than 2 years.  The spaghetti solution is easier but less useful to you.  I don't think that will come out within a year either.  I've worked for many companies and you get this "move forward" attitude.  This is the kind of thing that gets fixed in "the next product".  Then *maybe* they'll have time to port the solution back into the older product.  Um is not full of computer vision experts.  Even though openCV should be able to do this it's tough because filaments come in many colors.  But it *is* doable and hopefully they will hire some summer intern to do this.  So you'll need to come up with something else.

 

Another solution is to disable autoleveling on all UM3 printers, have an expert do a manual level once every 50 or so prints.  Enforce a rule that cura should make the bottom layer 0.1mm (this really helps a lot - it's another solution to help get that first layer to stick extremely well).  Make sure the person who does manual leveling does NOT use the calibration card to get the filaments to stick real well.  I'm not sure this will work with Nylon.

 

Nylon is a tough one - I've had trouble with Nylon sticking before (years ago) - I don't print nylon much.  I've printed about 10 nylon prints in the last 2 months but they were all small and had lots of brim and the bed was very hot I think (I'd have to check my notes).  Maybe if they are going to print nylon they need a dedicated printer for that - one with a different build surface?  I'm not a Nylon expert so I'm not sure what build surfaces might work well with Nylon.  I've only tried blue tape and glass.  I prefer glass.

 

In my town we have a local middle school - children aged 11-14.  They tried putting a printer in a teacher's room and having a teacher in charge.  That didn't work so well.  Instead they put the printers in the IT office of the school (it's a large school with about 1000 students).  I trained the IT people and they do a print service for the teachers in the school.  All over the school (e.g. history students can print things like medieval artifacts such as war machines).  Maybe this model/concept can be used where students and teachers don't touch the machines.  I used to think students should "learn" to use the printers but have since changed my mind and I think the students should be learning CAD and design.  Not how to use a printer that will be obsolete in a few years.  The other advantage of having these printers scattered around the IT department is that people are constantly walking by the printers and if a head flood were to occur they would notice within 10 minutes.

[SandervG 1,521]

Hi @brightorange, I replied to your DM. It would be helpful when you have time, if you would be willing to reply and supply us with some more information. Thank you! 

[SandervG 1,521]

22 hours ago, kmanstudios said:

1. Schools should have somewhat experienced or supervised people running printers.

 

I think you would be surprised how often this is not the case actually. Sometimes someone just decides a school needs a 3D printer.

[kmanstudios 1,120]

Just now, SandervG said:

I think you would be surprised how often this is not the case actually. Sometimes someone just decides a school needs a 3D printer.

I was basing this assumption off my limited experience. Oddly, it did come about as you said, but not available to students until they worked out the bugs and had a semblance of a system in place to operate. Never was great printing due to the huge volumes, but it was something they did set up while I was teaching at a university.

 

How others operate, I am at a loss and kinda baffled that a place of learning would just expect high end tech to be a simple thing with so many parts to make even the most simple output.

[SandervG 1,521]

Hi @rjd and @brightorange, I just verified internally and you should indeed be able to count on Ultimaker in this case (well, I like to think you can always count on us :))

 

What does it mean?

Your reseller or our partner in your area should replace your print cores or service the print head, depending on the intensity of the clog.

When your print got loose and you have a flooded print head it is important though that you contact your reseller before you attempt to clean it yourself. This is to determine the best course of action to a fix, like to have it serviced by an Ultimaker partner or you clean it yourself and receive necessary parts which could not be cleaned. This needs to be mutually agreed to with your reseller. When you try to clean it yourself and damage parts without having consulted your reseller, you may not get parts replaced. 

 

So, I don't know any details or context of your previous experience BrightOrange, but this is the service you should be getting from Ultimaker. 

 

Hope this helps!

 

 

[SandervG 1,521]

2 hours ago, gr5 said:

Another solution is to disable autoleveling on all UM3 printers, have an expert do a manual level once every 50 or so prints.  Enforce a rule that cura should make the bottom layer 0.1mm (this really helps a lot - it's another solution to help get that first layer to stick extremely well).  Make sure the person who does manual leveling does NOT use the calibration card to get the filaments to stick real well.  I'm not sure this will work with Nylon.

 

In general I think gr5 has the best tips but I would advice against this I believe active leveling calibrates your bed to 0.1mm distance between your nozzle and glass plate, and Cura's first layer is 0.27 which should ensure a good squish on the build plate. For Gr5 the above may work, but in my opinion 'an expert' who does 'a manual level' without a calibration card leaves too many insecurities/room for error.   

[brightorange 3]

hi @SandervG

 

I notified the lab I am working in, that we should be in touch with our dealer to coordinate regarding 4 printcores that were discarded because of these issues. This is already $400 that this thread has saved them..

 

To summarize this long thread I kinda hijacked, I think transparency regarding these kind of problems is very important. You admitted that your dealers were getting special updates regarding this issue and how to deal with it. The issue can cause a relatively minor thing like print detachment to full-blown disaster (over $400 in damage) 

 

Why is this so critical to some customers? because with open source  3D printers you can play with a lot of settings, sometimes the dealer can decide some issue is your fault because you did something wrong, and you are helpless. If an issue is known to happen more rarely or frequently, this can be a factor in deciding if you deserve service or not.

 

I hope this forum post can be enlightening for other users and save them undue costs and unnecessary arguments with their dealer.

 

 

 

 

[Digibike 107]

Hi,
I think you could have saved the BB Core.
Here a sprayer and hot soapy water would have been helpful.
Have seen the thread too late, because I can only bad English and
mostly follow the German thread.
I have had the problem too and have a) the extrusion back a bit
taken and b) significantly lower than in manual leveling determined.
Since then no more problems, no matter how long the prints last.
It is important that the first layers sit properly. But I also print with an oozschield around the objects (prints mostly ABS, because proper adhesion is even more important ...).
The core I have roughly cleaned with a Dremel and plastic milling attachment and then heated with hot air dryer and repeatedly removed (different fine slotted screwdriver). Also helpful is an essay that you put on top, which contains a fan and the cooling fins and just below blows. Then you can also heat up from inside (via the core) and warm up from the outside with the hair dryer and remove, without the problem can arise that the heat of the core moves upwards and in the cold zone even more "mischief" caused by constipation ...

Greetings, Digibike

[brightorange 3]

I see more people suggesting manually/mechanically offsetting the automatic leveling to make the first layer more squished to the build plate.

is there a way to add a "correction value" to the automatic leveling value? offset each value by X mm ? Or if I lower the first layer height it will give the same result? 

[Digibike 107]

Hi,

I do not know if an automatic correction so purposeful is ... Here also the thermal expansion plays a role.
The hotter the bed and the nozzle, the tighter the gap will be.
So if I print nylon or ABS, I have other distances
as with Ninjaflex or PLA, since I already clearly at the pressure plate
other temperatures drive, not to mention the nozzle, the
but has much less mass.
As can fast 0.1 to 0.16 mm (I have in ABS at
set me as a correction in the slicer).
And the temperatures vary so synonymous with the same material ever
according to speed and nozzle size because the amount of material
which heats per second and must pass through the nozzle.
I can currently imagine no automatic correction.
With my RF1000 yes, since it permanently via measuring cells the
Counter-pressure detected and thus conclusions about the pressure
can lead to the nozzle and thus can counteract. This works
but only the first layer and is our unofficial FW Hack (Nibbels
did a great job there!). But this is not in the Ulitmaker
available but only an inductive measurement on the Heizbettlage
(automatic Bedleveling). But that does not work with one anymore
DDP, as I have you on it - that's why I complete automatic bedleveling
disabled.

There remains, I think, only clean, the difference between Sliceinstellung and real distance of the individual temperature combinations to determine and korriegieren in the slicer.
Then the device should also go through cleanly for hours ...

Greetings, Digibike

 

[9thoracle 3]

The hotend on the UM3 is poorly designed. Period. 

[kmanstudios 1,120]

3 minutes ago, 9thoracle said:

The hotend on the UM3 is poorly designed. Period. 

Anything to back that up?

 

Been working fine for me since inception.

[9thoracle 3]

22 minutes ago, kmanstudios said:

Anything to back that up?

 

Been working fine for me since inception.

 

The fact that this thread exists.
The same thing has happened to me twice. 
Also, the fact that you could separate a super clogged heat block on the UM2 much easier/better. I should say it's designed poorly in comparison. 

Edited March 6, 2019 by 9thoracle

[kmanstudios 1,120]

34 minutes ago, 9thoracle said:

The fact that this thread exists.

Threads exist for a variety of reasons. By this reasoning, I am the greatest elephant hunter in NYC. Why? Because there are none around.

 

35 minutes ago, 9thoracle said:

I should say it's designed poorly in comparison.

Not really a rigorous  explanation.

 

38 minutes ago, 9thoracle said:

Also, the fact that you could separate a super clogged heat block on the UM2 much easier/better.

Sounds like a recurring issue for you regardless of machine or hotend.

 

 

Also, the 'super clogs' posted are not because of the hotend. It is because parts come loose, get dragged around and the material backs up into the mechanism. This is not a nozzle or hotend issue. This is 99% of the time, poor adhesion to the buildplate.

 

This does not mean that things will not happen, but for the mechanism to fail on its own is actually rare.

[Digibike 107]

I've noticed that when the core tears a component off the bed, blocking the nozzle, it pushes the filament out the side of the nozzle. As a result, he pours the core into filament and eventually it pushes the flap with the fan. This flap is only held by magnets. Why not just make 2 reed contacts that simply bridge the temperature sensor for the left and right core to ground when the door is open? This would prevent a lot, since, as soon as the flap is pressed, the firmware would shut off via implausible temperature values of the core. One should not print with an open flap, because the core can not be cooled and the flap can get stuck. So would be cheap and easy way defused the problem, since the flap in such a case very quickly pushed, I have observed in any case in both cases, is due to the excess filament. The pressure part is then anyway not to save, but the Core's already ...

Greetings, Digibike

[kmanstudios 1,120]

6 minutes ago, Digibike said:

I've noticed that when the core tears a component off the bed, blocking the nozzle, it pushes the filament out the side of the nozzle. As a result, he pours the core into filament and eventually it pushes the flap with the fan. This flap is only held by magnets. Why not just make 2 reed contacts that simply bridge the temperature sensor for the left and right core to ground when the door is open? This would prevent a lot, since, as soon as the flap is pressed, the firmware would shut off via implausible temperature values of the core. One should not print with an open flap, because the core can not be cooled and the flap can get stuck. So would be cheap and easy way defused the problem, since the flap in such a case very quickly pushed, I have observed in any case in both cases, is due to the excess filament. The pressure part is then anyway not to save, but the Core's already ...

Greetings, Digibike

Interesting idea. But to be honest, I have not had parts tear off the print. When I first started I did have issues with plate adhesion and some things came loose, dragging the part around and creating the super clog.

 

Once I solved my adhesion issues, the problem went away completely.

[9thoracle 3]

39 minutes ago, kmanstudios said:

Threads exist for a variety of reasons. By this reasoning, I am the greatest elephant hunter in NYC. Why? Because there are none around.

 

Not really a rigorous  explanation.

 

Sounds like a recurring issue for you regardless of machine or hotend.

 

 

Also, the 'super clogs' posted are not because of the hotend. It is because parts come loose, get dragged around and the material backs up into the mechanism. This is not a nozzle or hotend issue. This is 99% of the time, poor adhesion to the buildplate.

 

This does not mean that things will not happen, but for the mechanism to fail on its own is actually rare.

 

19 minutes ago, kmanstudios said:

Interesting idea. But to be honest, I have not had parts tear off the print. When I first started I did have issues with plate adhesion and some things came loose, dragging the part around and creating the super clog.

 

Once I solved my adhesion issues, the problem went away completely.

 

Just because the issue's root cause is bad adhesion, doesn't mean there aren't issues down the line. We're not talking about the obvious problem of bed adhesion and the obvious need to solve that. I am talking about the $375 UM3 head and how it can be brought to its knees with a clog; potentially ruining the head itself (see the picture attached).

I have never had a clog on the UM2.  And even if there were a clog like this that also occurred with the UM2, the assembly is much simpler and way cheaper to fix. This is where I derive my position that the UM3 head is poorly designed in this regard.

 

I sense, based on the type of responses you've given so far, that you've decided I am wrong and are willing to go to great lengths to prove that. Respectfully, I will not be spending any more time responding to you here.