Can your UM2 printer achieve 10mm3/s ??? Test it here...

[abstract 3]

abstract: would be interesting to see a print just changing the flow to 96% with the temp at 210C to better compare the result to your earlier test with the red filament.

What does your filament look like from the section that the feeder was driving at 4 and 5 mm³/s? Are the tooth marks clean and spaced the same as the section fed at lower rates?

 

OK, here is a print at 95% flow at 210 degrees.

I used a fresh bit of filament which had not been retracted.

My UM2 simply cannot feed fast enough to get past 3mm. First I see thin extrusion (does this mean blockage has started further up ???) Then I get feeder kicks, and then major under extrusion.

So, to be clear - I do see under extrusion starting sometimes BEFORE the feeder kicks.

Where to now ??

 

Mark.

 

[cyan 0]

Is this latest print with the teflon coupler in the condition in your photos from the head disassembly or did you clean up the deformed lip?

 

[abstract 3]

Is this latest print with the teflon coupler in the condition in your photos from the head disassembly or did you clean up the deformed lip?

 

Well, i didn't really see any way to clean up the deformed lip. I put a drill through the teflon piece, but the 'rut' in the teflon is a temporary *widening* of the shaft - so it could be a 'catch' point for unevenness in the filament - where the filament might have been carried along in the tube, with the kink riding along the inside of the bowden tube - then it gets to the rut in the teflon and it pops into the rut and stays there.

So this print is after I reassembled the head, with some very minor clean up of the teflon.

I think there is more going on.

I would really like to replace the teflon. No answer as yet from the 2 emails sent to Marrit at UM requesting a replacement (under warranty or chargeable)

Do I have the worst results of anyone on this thread ? Anyone failing earlier than 3mm cubed per sec ?

Mark.

 

[woofy 12]

Do I have the worst results of anyone on this thread ? Anyone failing earlier than 3mm cubed per sec ?

Mark.

 

A lot of us have been there. In my case Sander/Marriot sent me replacement teflon and nozzle parts which completely fixed the issue, as you can see After.

 

[gr5 2,167]

Anyone failing earlier than 3mm cubed per sec ?

 

The test is meant to be done at 230C so it's hard to say but clearly you have a bad printer. How far do you get at 230C?

When you had the white teflon isolator out, did you try pushing some PLA filament through it? I found that to be very enlightening. I'm particularly curious if you pushed through that filament with the big chewed up bump on it.

I think something is wrong with your feeder that it makes those chewed up bumps. Remind me again your tension? I have mine set to the minimum although around March 1 2014 they started shipping UM2s with a weaker spring and the tension set at the half way point.

I have done a print which averaged 8 retractions on every point of the filament. It was the Eiffel tower. In other words for every point on the filament it went back and forth through the feeder 8 times. On average -- therefore some pieces more than 8.

I don't get those chewed up bumps. I have used ultimaker red with no trouble (along with many other PLA filaments with no trouble).

Note that I always print now with the filament on the floor to avoid that sharp angle of the PLA going into the feeder.

Also I think a test at say 80% flow would be more interesting than 95% flow. But I still think something is wrong.

Also I wish you took the nozzle out and burned it clean while you had it all apart.

 

[cyan 0]

A lot of us have been there. In my case Sander/Marriot sent me replacement teflon and nozzle parts which completely fixed the issue, as you can see After.

 

Did you ever try isolating which part fixed the issue, or if they both contributed to it, and how the working replacements differed from the problematic ones? That would be really valuable information to have.

 

[Nicolinux 288]

With my first UM2 I had similar problems. But sadly it didn't help even with a new teflon piece. I think I was the only one where it did not help.

I'd do the folowing now:

 

1. Request a new teflon piece (and maybe a nozzle) - contact Sander directly.
   
2. Print Robert's feeder in PLA and replace the stock feeder. https://www.youmagine.com/designs/alternative-um2-feeder-version-two
   
3. Disasemble the head again, burn the hell out of that nozzle, let it cool and then put it into acetone for a few hours.
   
4. Print Robert's low friction spool holder (https://www.youmagine.com/designs/low-friction-um2-spoolholder) or put the filament spool on the floor or on a lazy susan.
   

 

By the way, speaking of clogs and chewed filament. I had a very long failed print (25h , failed after 21h) - http://umforum.ultimaker.com/index.php?/topic/5595-resume-long-print/.

The filament was worn down so bad that I had to dismantle the feeder to get/cut it out - fun fun :/

 

On the second picture, see that strand that's about to fall off? Thats ABS that will be transported through the bowden tube and into the nozzle where it will definitely cause trouble.

I couldn't push any filament through the nozzle at first - even at 260°. Then I remembered the hypodermic needles I had bought for this reason and poked inside the nozzle with one. The filament that came out features some dark spots - I guess that's the ABS stuff.

 

[takei-naodar 0]

Abstract how much space is between the metal coupler and the teflon piece on your machine?

 

[abstract 3]

The test is meant to be done at 230C so it's hard to say but clearly you have a bad printer. How far do you get at 230C?

When you had the white teflon isolator out, did you try pushing some PLA filament through it? I found that to be very enlightening. I'm particularly curious if you pushed through that filament with the big chewed up bump on it.

I think something is wrong with your feeder that it makes those chewed up bumps. Remind me again your tension? I have mine set to the minimum although around March 1 2014 they started shipping UM2s with a weaker spring and the tension set at the half way point.

I have done a print which averaged 8 retractions on every point of the filament. It was the Eiffel tower. In other words for every point on the filament it went back and forth through the feeder 8 times. On average -- therefore some pieces more than 8.

I don't get those chewed up bumps. I have used ultimaker red with no trouble (along with many other PLA filaments with no trouble).

Note that I always print now with the filament on the floor to avoid that sharp angle of the PLA going into the feeder.

Also I think a test at say 80% flow would be more interesting than 95% flow. But I still think something is wrong.

Also I wish you took the nozzle out and burned it clean while you had it all apart.

 

OK, so when I had the head open I did fire a gas torch down the head for a while. Not quite exhaustive, but I wasn't ready to start disconnecting the temperature probe and heater.

I did try pushing filament through the teflon. For the most part it works fine, and even with quite curved filament it goes through fairly smoothly - HOWEVER, there most definitely is a tendancy for snags (like the start of the filament) to hook into the rut in the far end of the teflon - so if there are any dings in the filament I would guess they will snag inside the teflon (but we dont know how hot it is down there so maybe its different during printing.

The feeder tension has not been adjusted since new, the white marker is at the TOP - I guess this is minimum - its an early UM2 - spring seems *very* strong. Maybe I need the weaker spring ?

Now to some more encouraging information:

The next thing I started to look at is the feeder. I took it apart and found tonnes of plastic dust:

Anyway - I cleaned all this out, and I also looked at the filament path going into the feeder. It looked like the plastic cutouts where the filament enters at the bottom were messy, so I filed them out

Then, I put it back together and decided to try something different.

Basically I really babysat the filament entering the bottom of the feeder. I held the filament so it had a clear path into the feeder, without pulling against the curve. I didnt really apply much pressure, so mostly I was easing the filament in, saving the feeder from the extra problem of the curve-pressure.

Then I started a print - 240 degrees at 95% flow, with me babysitting.

The result ! Amazing improvement:

All the way up to 9 with only mild incidents. Interesting - this suggests that fundamentally the material CAN flow through my hot end (even if it might snag on dings in the filament).

next I wanted to see how it worked at 240 / 95% if I didnt babysit the feed-in :

So, pretty good still, barfed at 8mm. This is 240degrees, 95% flow, no babysitting (just making sure the filament was not stuck somewhere on the back- just very loosely spooled hungg on the holder, wide loops of faberdashery non-wound.

Next, to answer your question, I tried the same thing at the default temperature for the cylinder, which appears to be 220 degrees. This is 220 degrees, 95% flow, no babysitting of feed in:

barfed at 6mm.

So. Let me summarise what I think we have learned.

1) The Teflon part gets distorted over time and can potentially cause dings in the filament to snag and create an extra load on the feeder

2) The nozzle can be blocked, and we suspect it may not give 0.4mm beads, so some alteration in the volume calculations is required.

3) The feeder may be contaminated with old debris.

4) Printing at 240 degrees definitely helped - perhaps its generally reducing the pressure, or maybe its moving the heat up the hot end and making the teflon hotter and more tolerant to dings in the filament.

5) Feed in to the feeder seems to be pretty important on my machine, and the filament inlet seemed to be a bit sharp. All those people saying put the filament on the floor are probably right - the pain of getting off the spool and into the feeder seems significant. The end of the roll gets more curved and this hurts the feed-in to the feeder as well as creating more pressure all the way down the bowden tube and in the teflon part.

I think the overall message is that this filament path is a symbiotic system - all the pieces are playing a part and failure can be a combination of cause and effect across multiple parts. This may be why some people who replaced a single part did not solve the problem.

For me - I am going to design a guide to do what I was doing with my hand in the photo above - to keep the filament straight into the feeder, probably around a large radius.

I want to replace the teflon and nozzle. Maybe I will try one of the alternative feeder designs.

I think I am approaching the end of this investigation. I will test again with the guide etc, and if I get new parts.

Thanks to everyone for your comments and suggestions. I hope the information I have added to this thread will help those that follow.

Cheers

Mark

 

[abstract 3]

Abstract how much space is between the metal coupler and the teflon piece on your machine?

 

1.2mm

 

[takei-naodar 0]

1.2mm

 

Increase that to 2-2.1mm by turning the metal coupler clockwise..... you can use the holes in it to do so.

 

[woofy 12]

Did you ever try isolating which part fixed the issue, or if they both contributed to it, and how the working replacements differed from the problematic ones? That would be really valuable information to have.

 

I still have the old parts, but now my machine is working so well I don't want to mess with it.

 

[abstract 3]

Increase that to 2-2.1mm by turning the metal coupler clockwise..... you can use the holes in it to do so.

 

How do you think this will help ? I guess the only effect is to move the teflon away from the heat. However its creating a larger void into which stuff might snag ?

Doesn't turning the coupler change the space between the coupler and the heating block ? The teflon seems to just sit inside the coupler, pushed in with the spring.

Clarification ?

 

[cor3ys 28]

Increase that to 2-2.1mm by turning the metal coupler clockwise..... you can use the holes in it to do so.

 

I believe their should only be a 1mm gap 1.2mm sound ok but i would not increase this to 2.5-3mm

 

[skint 61]

@ Abstract. Excellent breakdown and feedback. That`s a very messy feeder! lol

I started printing again last night, had a few issues. After clearing the bunged nozzle and changing material, noticed a white strand of material inside my feeder! Now considering I have NEVER printed with white, I thought this was strange lol.

It was a circular strand of material that had worn and fell off the feeder wheel. :shock: Hoping I can start printing again later without trouble and hope this feeder wheel isn't an issue.

Why do we love these printers so much when all they do is treat us so badly?

 

[SandervG 1,521]

Why do we love these printers so much when all they do is treat us so badly?

 

You got it all wrong.. they keep things .. interesting!

OnTopic, Note to self: Never leave this thread alone for a few days. Took me 5 pages to get back up to date!

 

[gr5 2,167]

@abstract - I'm glad you had some limited success. I'm still a little frustrated that you still haven't printed the test cylinder at 230C. Everyone else printed at 230C - that way you are comparing your printer to others. However you did produce some very interesting data.

Interesting theory about the heat reaching the upper head but mostly I think the issue is that the difference between PLA at 200C and at 240C is like the difference between toothpaste and honey.

The black feeder is made of ABS which melts at a different temp and if left hot (above 220C) for hours turns into a magic kind of cement that causes lots of clogging-like issues. I think 90% of all your problems were that feeder dust getting into your nozzle and that your best fix will be to replace it with one printed from PLA and also to (after at least another 5 meters of PLA printed) take that nozzle off and burn it out good (don't melt it though) and soak it in acetone for at least an hour if not overnight.

 

[abstract 3]

@abstract - I'm glad you had some limited success. I'm still a little frustrated that you still haven't printed the test cylinder at 230C. Everyone else printed at 230C - that way you are comparing your printer to others. However you did produce some very interesting data.

Interesting theory about the heat reaching the upper head but mostly I think the issue is that the difference between PLA at 200C and at 240C is like the difference between toothpaste and honey.

The black feeder is made of ABS which melts at a different temp and if left hot (above 220C) for hours turns into a magic kind of cement that causes lots of clogging-like issues. I think 90% of all your problems were that feeder dust getting into your nozzle and that your best fix will be to replace it with one printed from PLA and also to (after at least another 5 meters of PLA printed) take that nozzle off and burn it out good (don't melt it though) and soak it in acetone for at least an hour if not overnight.

 

I did print at 230, with 95% flow:

For my tests, temperature was just one variable. Up to 240, it adds a percentage improvement, down to 210 its worse.

The biggest improvement comes from manually feeding the filament into the feeder, and praying there is nothing going to get stuck in the rut in my Teflon.

Switching from 100% to 95% flow feels like maybe it helped, but really I think the problem is NOT about over-pushing beyond what the nozzle can deliver and its more about achieving a clear, unobstructed path for the filament from the reel to the nozzle. I do also feel like my feeder is overly aggressive when it kicks, and I hope to put in the newer (weaker) spring. I also want to make it a gentler path along the tube and hot end somehow.

So, my plans include:

1) Replace Teflon

2) WD40 in Bowden

3) Weaker Spring in Feeder (to reduce kick-damage and later problem with nicks)

4) Burn out the nozzle

5) Finish my design for a 'feeder - feeder'

"Feeder-Feeder version 1.0" - needs further work: Not 100% certain it will help, but feels like it might.

Meanwhile, I can print slowly and mostly it works OK. I am having problems loading filament and I am pretty sure the end is getting snagged in the Teflon. I plan to start sharpening the filament when loading.

Cheers

Mark

 

[skint 61]

I sometimes cut my filament with pliers, always gives a nice sharp point for loading

 

[aaronalai 2]

Yikes, the activity in the forum seems to be increasing every day. This certainly is an interesting thread :mrgreen:.

Anyway, @abstract thanks for the exploded view of the printer head! For some additional perspective I got little bits of the motor extruder case ground down and could see little flecks in the bowden tube, this was very minor and did not affect the extrusion rate. But, once I noticed the bits in the bowden tube I did what you did and bent the filament more normal to the bottom of the motor extruder case. I have not yet seen any bits of black plastic accumulate since.

On another note, your filament looks very similar to mine when the nozzle is blocked. The filament tends to grind down, it will skip backwards but only in extreme cases like if the bed is too close to the nozzle. Do you see black flecks in the bowden tube? Also, has your filament guide proved useful?

 

[aaronalai 2]

@abstract I saw in your post in the ABS printing thread. I had a couple of very persistent clogged nozzles after printing with ABS plastic. To mostly fix this I used the move material option in the menu options and moved the PLA material through the print head at the ABS recommended temperature, I was extruding plastic the whole time once the head got to 200C. After I was convinced the head was clear I continued to extrude material until the nozzle got back down to 200C and spun the dial in the other direction to retract it. I say mostly fixed because I would continue to get little bits of ABS ejecting from the print nozzle, but nothing that would clog the nozzle.

 

[wallan 0]

I also had some problem and the thing that did the most difference for me was to build a low friction holder that I placed on the floor below and behind the printer so that the filament goes straight up in the feeder and in a direction that makes the curve of the filament on the holder match the curve on the bowden. That way I have minimal friction. I think that fixed most of the print speed issues on my part. I also spray a little WD40 on the filament roll. That way the bowden gets lubricated as well. Before doing this I could not print at 50mm/sec 0.1 mm layer without feeder skipping but now I have no issues at all at that speed. I can still not reach the 100 cubic mm per second unless increasing temperature to 240deg but that's a none issue as I never print that fast anyway.

 

[illuminarti 18]

The gap between the coupler and the Teflon piece should be around 1mm. The spring keeps the Teflon piece presses down against the top of the nozzle. So increasing the gap between the coupler and the Teflon piece doesn't affect distance from nozzle to Teflon (they're always touching). However, it does compress the spring more, increasing the force on the Teflon, and making it more likely to deform. It also makes it more likely that filament will get up under the fan hood, since it reduces the clearance between nozzle tip and fan shroud.

[Dim3nsioneer 557]

First, sorry about hijacking this thread a bit. I remember having read somewhere on the forum that the hotend heating power of the UM2 is higher than that of the UM1. Is this correct?

With this in mind, I wonder what volume flow could be achieved with an UM1 (0.4mm nozzle). I see that for an UM2 8mm^3/s seems to be an acceptable value. I remember illuminarti's UM1 flow tests with a 0.65mm nozzle.

Or asked in a different way: Would you expect the UM2 cylinder test to be completed on an UM1?

 

[illuminarti 18]

No, the heater on a UM2 is lower powered than on a UM1 - and has a much smaller melt zone, so in general it's going to have a harder time getting heat into the filament as it moves through the hot end, compared to the UM1.

But the limiting factor to extrusion on both machines isn't the hot end, so much as the ability of the extruder motor arrangement to exert force on the filament without simply tearing up the filament.

The throughput limit on a standard UM1 is about the same as most people see on a UM2 - around 8-10mm³/s. A well-tuned UM1 should be able to complete the standard test, but probably not go a lot higher, whereas a UM2 on a good day can get up to about 15mm³/s.