Hey! This thing is still a Work in Progress. Files, instructions, and other stuff might change!

Water-cooled hot end

by adrian-bowyer Jul 15, 2012
Download All Files

Please Login to Comment

Have you experienced an increase in maximum print speed?

In my Ultimaker the Hotend is the bottleneck that limits maximum print speed and the maximum flowrate of the hotend is in turn limited by internal friction between hotend and plastic. This modification should not only be capable of increasing the reliability of your printer but also its maximum speed.

I haven't tried running things faster, but the water-cooled hot-end should reduce the friction between the filament and the nozzle where the filament enters the melt zone, so I would expect that your prediction is correct.

Have you considdered making a tiny peristaltic pump for that silicone tube? I have a minuscule one cooling the brush motor on one of my iRobot Roomba floor vac robots. It is an enclosed system with water jacket on the motor, and radiator in the brush airflow

Yes - that would work well, and it's quite easy to print a peristaltic pump. But you can get greater flow with a gear pump, and it's also easy to print that - especially as it doesn't have to be watertight (you just run it under water in the sump with the motor on a long shaft above).

Have you considered making a tiny peristaltic pump for the silicone tube? I made a minuscule one to watercolour the brush motor on my Roomb

Have you considered making a tiny peristaltic pump for the silicone tube? I made a minuscule one to watercolour the brush motor on my Roomb

Have you considered making a tiny peristaltic pump for the silicone tube? I made a minuscule one to watercolour the brush motor on my Roomb

Have you thought to replace the pump with a small peristaltic pump? The silicone tube lends itself to this pump type and I have successfully made an enclosed system to watercolour the brush motor on one of my iRobot Roomba floor cleaning robots

I was wondering when ideas from computer overclocking would trickle into 3D printing!

This is something I would definitely go for once seeing convincing evidence that it results in significant improvements in either print quality or speed.

Are "before
amp; after" data or photos available?

Where do you have the fan's air flow directed, and how well focused (concentrated) on that location?

Not at the moment - but I can say it works at least as well as the fan cooled version (which works very well).

i have build a test setup and now i have to find the sweet spot fore the Teflon to end near the hot end.

and see wear the cooler sweet spot need to com.

great id.

what is the diverseness in temp hot in cold out of the brass ?? it looks like you mount the brass to the pla of abs frame.

and Wat is the length of the brass tube you use.

I haven't neasured the inflow and outflow temperatures with a thermometer, only with my finger :-) I'll try and be a bit more scientific and post the results here asap. The brass has a PTFE pad between it and the PLA frame. But it (the brass) is only around 35 C in operation, so that's probably not needed (it's a good safety feature if the water fails, though).

I don't understand why we extract heat from a device that we are deliberately heating.

Along the same line, it's my understanding that the benefit of a cooling fan is to solidify the extruded plastic quickly to avoid sagging over spans. It seems a bumber of people think differently.

Can you help me understand these issues? (I speak fluent Engineering-ese)

The idea is to keep the melt zone as short as possible near the tip. This has the advantage of minimising ooze, and minimising friction (both coulomb and viscous) as the filament is driven into the nozzle. In particular, you don't have to retract the filament very far to stop the flow altogether.

To achieve this you want a high-powered but compact heater at the sharp end, with a good PID controller to keep it at temperature no matter what. Then you want as quick as possible a drop in temperature as you move back up the filament being fed in. Jean-Marc's design achives this by having a sh
ort stainless-steel tube between the nozzle and a cooled block. Stainless has a surprisingly low coefficient of thermal conductivity for a metal. We normally cool the block with a fan. This does it with water.

For details of the construction, see: http://reprap.org/wiki/RepRapPro_Huxley_hot_end_http://reprap.org/wiki/RepRapP...

The point here is that in order for an extruder to run smoothly, the heat needs to be concentrated in a smallish zone, close to the nozzle. If the plastic softens too high in the hot end, pressure from the extruder causes it to expand, and jams the hot end. This mechanism cools the upper half of the hot end - not the print.

Nice! What's on the other end? Tank, radiator, drain?

A 1 litre water jug...

I'm considering a fan and heatsink with a copper coil on the back to make a closed system, though. I think I'll take a look at what all these maniacs who cool their CPUs with water do, and what kit is available to support them.

I tried this with a CPU WC kit, i found that it cooled it just to much. ill have to have another go at this and try to hack the rad/pump and gear the motor down (assuming it has one)

hey Adrian! iv been having a little play with it, i have a H100 (http://www.ebuyer.com/277523-corsair-hydro-series-h100-high-performance-liquid-cpu-cooler-cwch100?utm_source=googlehttp://www.ebuyer.com/277523-c...

the pump was just being a pain in the but for me and it wouldn't go slow enough
lt;' iv added a few step downs going into and out of the block, its helped a little. but i think ill need to take the motor out and replace it or gear it better, its just to great of a flow (or just to good at its job) if found its cooling my hot end down to much for extrusion atm. ill get there. keep trying :)

Maybe add more power to the heater by lowering the resistance?

Could you just drive it slower? If the CPU pump is a simple DC motor, maybe just drop the voltage, or drive it via PWM. I found a flow of 5 ml/s was just about perfect.

PS - I suspect that a printed peristaltic pump would work pretty well. And one needs the silicone tube anyway.

There are external router speed controllers that can control the speed of any 110v AC router motor regardless of the motors original speed (even 1 speed ALL on types).

This external box is capable of varying the power and therefor speed to any standard 110v ac powered item you plug into it.

The pumps speed could easily be controlled with this without any pump modification at all.

I use one to control the flow speed of the pump on my cnc mill currently.


I have no doubt this would function perfectly in this application.