Ultimaker 2 Cooling Hack for ABS (and for PLA fine details at end of print)

[Wisar 86]

When I first got my Ultimaker I struggled to print ABS.  This is a little ironic as I had printed nothing but ABS on my previous printer, a Wanhao Duplicator 4S, as I struggled to print PLA on it!  

I have since been able to get past my issues of build plate adhesion and warping and print ABS quite successfully for many things.  I now have an enclosure for the printer and have decided that Klapton tape is my printing surface of choice.  I still struggle with fine details and in particular when printing at high resolution (<100 microns).

Speaking of fine detail at high resolution...I have similar issues with PLA but only when printing the final layers.

This first picture shows two tanks with the top printed in PLA and the bottom in ABS.  They are 1:100 scale models and are about 2.5 inches long so there is quite a bit of detail there.  It is not a completely fair comparison, however, since the PLA print was done with a 250 micron nozzle and the ABS print with a 400 micron nozzle.  The PLA print is 60 micron resolution and the ABS print is 80 or 100.  

Ok, not a great apples to apples test but what is clear on both prints is the problem that I am trying to solve.  On the lower, ABS, print you can see the poor quality on the small dots on the engine compartment, on the periscopes forward of the turret hole, and on the top of the fuel tank.  The lines are not as sharp either (corners you would expect to be better on the PLA due to the nozzle).

On the top print, the PLA one, the issue is the top of the fuel tank and the periscopes at the bow of the vehicle.  In both cases they are somewhat smudged.

My solution to the PLA problem has been to either print multiple parts or to print a "sacrificial" part or two that accomplish the same thing.  I have already slowed down the print using the S3D feature that allows one to adjust speed for layers below some number of seconds.  I use 15 seconds as the threshold and 20% as the factor.

I have not had a good solution for ABS but I know from my Wanhoa Duplicator 4S that cooling is the key.  On that printer I had installed a highly directed ducted fan and it made a huge difference.  So enter the below hack:

This is not the first of these experiments that I have tried in the cooling space.  My first attempt, which had limited success, was a piece of plastic that clipped onto the print head assembly and focused the cooling flow onto the area below the nozzle.  I talk about it elsewhere but it was not the cat's meow.

Below is a test print done in PLA that shows the same g-code printed on a printer with standard cooling and the other with my directed fan.  It is obvious which is which.  

My first test with ABS shows a significant improvement as well but as I do not have an apples to apples comparison that will have to wait (yes, I know, under extrusion).

I would be interested in knowing what people think of this idea.  I would also like to hear any feedback into whether my printer is going to explode from running with the Olson block exposed!  I am doing a two hour long print in PLA as we speak and will then try an even longer one later.

[gr5 2,071]

Olsson block exposed is fine. The cover is for safety to reduce the chance of burns but I touch my nozzle all the time with bare fingers. You just have to be quick. Also you want to keep wind off the nozzle if you can but you have taken care of that mostly.

You obviously want to create higher quality prints. The trick is slow and cold. Cold because at 240C PLA pours out like honey and at 190C it comes out like toothpaste and stays where you put it and the nozzle doesn't leak so much.

I'm not sure what "cold" is for ABS. Certainly 250C is too hot if you want extra good quality. Probably 230C? maybe?

Slow because at 190C it requires more force so you have to slow down. Also going over 20mm/sec (the JERK parameter) reduces quality because you don't want your feeder speeding up and slowing down all the time.

So just for fun, for PLA, try 200C and if .1mm layer height or thinner try 30mm/sec. If .2mm go no faster than 20mm/sec at 200C.

The problem with your ABS prints I'm pretty sure is that the nozzle is too hot. One can get just as good quality with ABS if you experiment enough.

The main things to play with here are speed and temp. Then you can also mess with fan speeds. For PLA you want as much fan as possible. For ABS you will get hugely weak parts if you have too much fan. You want the minimum fan where overhangs are still okay. Even better use a cross flow fan and enclose the printer (google it - foehnstrum has a great one of these).

[Wisar 86]

Thanks GR5. I am printing my PLA at 195-200 and in the range of 20-30mm/sec depending on the quality that I am looking for. With PLA, and a 250 micron nozzle, the quality and detail is amazing. Except for the top layers if there is some small detail sticking up and I am not printing multiple parts (or have a sacrifice out there). I want the directed fan to help there and it seems to do so. I will keep the speed versus layer height guide in mind but I had sort of stumbled onto that anyway, going faster with the thinner layers.

With ABS I am printing at what seems like crazy low temperatures of 215 (starting at 245 and winding down after the first layer). The idea behind the directed fan is that I can cool the extrusion area with the lowest possible fan setting and not have that same fan on the rest of the print...helping to prevent the print popping off. The same kind of setup worked brilliantly on my Wanhoa Duplicator 4S and it seems to be working pretty well here too.

I think you are confirming the direction that I have been taking so thanks. I am particularly glad that I don't have to enclose the rest of the Olson block as that was going to be tricky!

What surprises me is that no one else seems to have been done this?

Thanks again,

Will

[gr5 2,071]

ABS has an issue that PLA doesn't because the glass temperature is so high - adhesion.

I suspect if you are printing at 215 and have even the tiniest 10% of fan you will get bad layer adhesion which means your parts won't be very strong and can break easily along layer boundaries.  For parts that don't need to be strong this is absolutely fine.  But if you want something as strong as PLA you need good bonding.  Now I could be wrong - especially if you print slow and thin layers the nozzle might be in contact long enough to melt the layer below but I suspect your parts are weaker than you think.

Try printing a 1cm cuboid that is 1cm by 1cm and maybe 5cm tall.  Or much taller.  Print it with fan off and nozzle at 240C and then print it with fan on 100% (maybe the second half?).  Then try to snap it in half like a pencil and see how strong the zero fan portion is compared to the part that had 100% fan cooling.  That will give you a good feel for layer adhesion issues. Turn off infill for this test. It will save you lots of time and help make the part a bit weaker for a more valid test. An ABS part with good layer adhesion will be very difficult to break even hollow with two .4mm shell passes (shell around .8mm or .75mm for .25 nozzle).

Enclosing your printer helps quite a bit as now the fan is blowing 50C air instead of 20C air.  Glass temp of ABS is around 100C.

Edited March 2, 2016 by Guest

[Wisar 86]

Thanks gr5,

I will give it a shot. In the meantime I am also trying a couple more materials that might also get the same test. I have a reel of PET that I am liking and have some Colorfabb co-polyester on the way as well. Chasing the promised land of a material with the advantages of both PLA and ABS but without the disadvantages of PLA or ABS!

Will

[izzy 33]

Hi Wisar, I've done a design for ducted nozzles you may want to have a look at, they can swing out of the way to give you full axis to the nozzle and you can adjust the height from the workpiece,and doesn't need any screws. if you do try it let me know what you think

https://www.youmagine.com/designs/um2-fan-duct-v2

[Wisar 86]

izzy:

I like your design, good idea on the swing and being able to adjust it.   Instead of a swing out I went with a slide out...though I have not put in a retainer yet.  

The goal of my cooling design is a little different to yours though.   I actually have two different fan outlets though I only show the ABS one above.  The other one is more akin to what is on the UM now in that it has a pretty wide aperture.

My design goal with the one pictured above is to direct the air flow as much as possible to just the extrusion area.   I got this idea from an add-on that I had installed on a Wanhao Duplicator 4s that allowed it to print ABS with amazing precision:

Uber Cool Active Cooling for Makerbot Clones

Here is a sample of the quality that I was able to print with the above cooling solution on my old printer.   I want my much more expensive UM2 to be able to do the same quality (in ABS)!

Cheers,

Will

Edited March 3, 2016 by Guest

[eldrick 64]

Best UM2 fan shroud design of the half-dozen that I've tried.

 

Hi Wisar, I've done a design for ducted nozzles you may want to have a look at, they can swing out of the way to give you full axis to the nozzle and you can adjust the height from the workpiece,and doesn't need any screws.  if you do try it let me know what you think

https://www.youmagine.com/designs/um2-fan-duct-v2

 

Edited March 3, 2016 by Guest

[Wisar 86]

Does PLA really have twice the layer bonding strength as ABS or have I screwed something up?

3D Print Filament - Strength Tests

PLA w/ 100% Cooling ==> 14.3

ABS w/ No Fan (215c) ==> 7.1

ABS w/ No Fan (245c) ==> 6.6

ABS w/ 40% Fan (245c) - My Mod ==> 4.8

ABS w/ 60% Fan (215) - My Mod ==> 4.6

ABS w/ 60% Fan (245) - UM Standard ==> 2.6

My test was pretty simple. I printed a little test piece as shown below (two of each). Little hole at each end to secure it to a leg of my desk and a baggage scale. Pull and record the point where it breaks. Recorded results on iPhone so as not to miss anything!

The results were interesting. PLA was twice as strong as the strongest ABS which had been printed at 215c with no fan. Strangely ABS printed at a higher temperature was a little weaker? Using a fan did indeed weaken the ABS print but my mod caused less weakening than did the standard UM cooling solution.

I may do a little more testing but I like what I see from the results relative to my directed fan causing less weakening than the standard UM fans ... and from my other tests I know that it improves print quality when it comes to fine details.

[gr5 2,071]

Wow!  Very nice experiment.  Hmm.  Try 255C no fan and try enclosing the printer?  Was your bed at about 100C during this time?  I find I can bet the air temp up to about 40C by enclosing the front and putting a huge box over the top that includes the bowden in the back.  I'm curious if this helps.  Although I suspect you won't get any better bonding than you already got.

ABS tends to be more shatter resistant if you use it say as a hammer because it is just a tiny tiny bit more flexible and that's enough to often make a huge difference.  For that reason alone people say ABS is "stronger".  In my own personal tests testing the other way (tests that didn't care about layer adhesion - e.g. printing your parts flat) I found ABS and PLA gave way at the same tension.

[gr5 2,071]

Hey where do I get one of these tekbox things? When I googled it I found only electronics stuff.

[Wisar 86]

Scales r Here!

Bed was at 110 and I have a full enclosure from the German guy (3DPrinterStore24) that makes them.  Chamber temperature was 61c.  I think I am good to go though as the models that I want to print don't need much strength...details are more important.

Printer Cover is Here

Edited March 3, 2016 by Guest

[Wisar 86]

3D Print Filament - Strength Tests

PLA w/ 100% Cooling ==>  14.3

ABS w/ No Fan (215c) ==>  7.1

ABS w/ No Fan (245c) ==>  6.6

ABS w/ 40% Fan (245c) - My Mod ==>  4.8

ABS w/ 60% Fan (215) - My Mod ==>  4.6

ABS w/ 60% Fan (245) - UM Standard ==>  2.6

 

Another result to add to the above. Just got a reel of ColorFabb Ngen. Expensive. Nice.

Ngen w/ 100% Cooling ==> 14!

[Labern 774]

nGen Also prints really nice with no fan or fan on low. So would be good to do a strength test with lower fan %

[Wisar 86]

nGen Also prints really nice with no fan or fan on low. So would be good to do a strength test with lower fan %

 

I was pulling the leg off my desk as it was! How much stronger do you want it to be? :-)

I designed that part to be reasonably fragile but was surprised by how tough the PLA was...and the test that I did with the Ngen was actually a part that was under-extruded so I was not expecting much and was planning on a re-do. I may still do another test with and without the fan.

[izzy 33]

Nice, I like the tensile test, I take it your using Kg Force? But what is your Cross Sectional area, and are you testing in both a flat and vertical build orientations?

[Wisar 86]

Nice, I like the tensile test, I take it your using Kg Force? But what is your Cross Sectional area, and are you testing in both a flat and vertical build orientations?

 

Test Print with Dimensions

Printed with two perimeter shells.  Expectation was that break would occur on single column!  Only did test of layer shearing as that was  the aspect of cooling versus not cooling that I wanted to explore.