TPU95A with 0.25 printcore

[AurelienC 0]

I need to print TPU95 with small details, on my UM3

According to this article : 

https://support.ultimaker.com/hc/en-us/articles/360011940440

 

UM2 is compatible with TPU95A and 0.25 nozzle, but not UM3, which is 0.4 or 0.8 only

 

Is there a technical reason about it ? How can you loose a spec upgrading the machine ?

Actually Cura don't want to slice in this configuration,  like it is totally forbidden.

 

Is there a way to damage the printcore ou the printer ? Anyone already tried ?

 

Thanks

 

[SandervG 1,521]

Hi @AurelienC, thank you for your message. The Ultimaker 2 printhead is entirely different than the Ultimaker 3. The Ultimaker 3 uses print cores, which has different heating elements and hot / cool zones etc compared to the Ultimaker 2. Printing such small details on an Ultimaker 3 requires a lot of retractions and detailed extrusion in order not to over/under extrude. The flexibility of TPU makes that very difficult. That is why the profile does not support printing with a 0.25mm print core. The results will be unreliable. It won't really damage your machine, but the risk of a clog is also significant due to the TPU staying in the print core for a relatively long time. (This should be able to cleaned with our cleaning filament, but it would interfere with your print). 

 

If you are really determined to print with TPU and the .25 print core, you would need to slice with a different material and tweak the settings for TPU, and remove the NFC from TPU so your Ultimaker does not read it as TPU. Obviously I don't know what you wish to print at what urgency, but generally speaking, I would not recommend this route. 

[AurelienC 0]

Thanks for this answer.

I can ask an access to UM2 machine, probably better to try it in this way, isn't it ?

Thanks

[SandervG 1,521]

Hi, I think printing TPU95A with a 0.25mm nozzle will always be on the 'experimental' side given the flexible nature of the filament and the amount of control necessary with the nozzle but you should get better results. Perhaps if you can share your model, or an img of it, we would be able to tell you with more confidence how difficult or easy to print it would be. 

[AurelienC 0]

Hi

My goal is to print a 1mm timing belt, type FHT ( http://shop.sdp-si.com/catalog/product/?id=A_6B18M190030 )

As SPI/SI is the only one supplier that I found, and they are stopping the production of the length I need,  I would like to try printing my own timing belt with the length I need.

 

You can download my stl file here : https://cloud.univ-grenoble-alpes.fr/index.php/s/RyXxcdw4HziJXsW

 

This doesn"t retractation, but teeth are really smal and definitively need a 0.25mm noozle

[geert_2 557]

All timing belts I have seen are composites of cord made from hemp or flax (NL: hennep or vlas), or from metal wire, embedded in different sorts of rubber. Much like car and bike tires. The cord is absolutely necessary for strenght, durability and dimensional stability.

 

If the belt is just a drive belt, like in a battle robot, it does not really matter if it fails (apart from that you lose the game). But if it is a valve-timing belt in an internal combustion engine, failure would completely destroy the engine.

 

If you would try to 3D-print that, you are going to have a hard time getting the cord in there. I think you would be better off printing a mould. And then wire the cord in it, in a spiral enless loop. And impregnate that cord with some flexible and very strong composite or PU or liquid rubber? Use plenty of mould releasing agent, and a mould in multiple pieces, or you would never get the thing out of it.

 

If I were in your place, I would try to find a real one, thus search more. In the Netherlands, companies like INDI have a huge amount of mechanical components for machines, including lots of timing belts. If you couldn't find double-sided teethed belts, what about single sided belts, and glueing two together on their backside? There does exist rubber glue for belts and O-rings and seals.

 

Or, depending on the equipment, modify the equipment so it could work with a standard belt of different length? For example by adding or changing a couple of tension wheels or springs?

 

[AurelienC 0]

Hi everyone

 

As the timing belt I need is the smallest timing belt ever ( 1mm pitch ), it doesn't have and doesn't need a cord inside. I already cutted one and they are 100% polyurethane. Also, the mechanical force are limited, as load and speed are really low.

 

I succeded in printing one ! This is a first try but I am much more confident to print a usable one. 

 

What do you think about it ? How can I improve the rack shape ?

 

Thanks

 

 

[geert_2 557]

Yes, with that tiny size you won't get far with a cord, indeed. I did not realise it was that small, I thought it was a regular size belt.

 

Probably print even slower? And adjust temperature down, maybe to well below recommended minimums, so the filament does not burn inside the nozzle due to sitting there for too long? And of course the thinnest layer height.

 

For printing accurate PET models, I have gone down to 10mm/s speed, 0.06mm layers, and 10°C lower than minimum temp range for that material. But I don't know if that would work for TPU too?

 

The idea behing printing slow is that there is very little pressure build-up in the nozzle, thus the flow accurately matches the model, and the filament has plenty of time to melt. But at the normal temperature, it might degrade, decompose, or burn. At least, this happens with PET and PLA, I don't know for TPU. So that is why you also need to lower temp, to below its decomposing temp. Thinner layers give a lot more accuracy when going around corners than thicker layers, and they squeeze the new melt better into the previous layer, so the filling is better.

 

See this test with transparent PET. Each block is 10mm x 10mm x 20mm. The text is a hollow watermark halfway in the blocks. Top row = 50mm/s, bottom row = 10mmm/s. Layer height from left to right: 0.4mm, 0.3mm, 0.2mm, 0.1mm, and 0.06mm. All with the standard 0.4mm nozzle of an UM2. Brown discoloration is degradation due to sitting in the nozzle for a prolonged time at low speeds and layer-heights. But the clarity of the 0.06mm layers at 10mm/s shows that this gives a much better filling, with far less entrapped air. Also, corners are way sharper and cleaner. Bottom photo shows this clearest block as printed, and after cleaning-up and polishing, so the internals become more visible. Text is 3.5mm caps height, text legs are 0.5mm wide.

 

 

 

What I sometimes do with a new material, is remove the bowden tube. Manually heat up the nozzle, and insert a piece of the new filament from above. Just like when you would do a cold pull / atomic pull. And then manually adjust temp from low to normal, and manually push some filament through. Then gradually reduce temp again, and feel when the flow stops. Do this a few times. This gives some feeling of when the material starts to flow well.