When i started this project 8 months ago i did know it would escalate so much, but i have become obsessed with the idea of building the perfect water cooling block for a 50W LED.

But what means 'Perfect' in that case?

It has do be:

- modular

- easy to build

- cheap in material costs

- repairable if the LED dies

And it took me many prototypes to come anywhere close to that.

I started out with a very simple design, a 3D printed box with 3D printed fittings and a place where you could fix a LED in place with epoxy:

I was surprised that it actually worked, but the design had many flaws.

It wasn't repairable if the LED or one of the 3D printed fittings broke.

Also it took way to long to print for making many of them.

I would not recommend using this module version, but if you want to print it here are the files:

https://www.thingiverse.com/thing:4508165

The next prototype, i build tryed to cool the LED from immersive from both sides:

This looked very nice, but had some major problems.

Since epoxy was used repairing is not possible.

The Phosphor layer on the Cop LED dissolves in water over weeks.

But this was the first module where i integrated real fittings into the design.

Also it was leaky....so i spent some time trying to get my printed parts watertight with Plastidip and epoxy:

Sadly both are expensive and a pain in the ass to work with.

So i decided to step away from 3D printing and try something else.

My options where:

- Milling

- Casting

- Injectionmolding

Since i didnt had a Mill back then, and also no injectionmolder, i picked resin casting.


I started Making modules from epoxy resin, using silicon molds from 3D prints:

The resin made prototypes where promising, manufacturing speed was improved since i could just make more silicon molds to make 3 modules at once.

But the downsides where:

- Lots of epoxy vapors

- 12/24h cure time

- dependent on Mixing not temperature stable

- absolute not repairable

Nerveless i build about 20 resin  LED modules:

and few of them are still in action of them, the fail rate was way to high:

At this point i realized Resin casting wasn't the way to to, since i now had hard to recycle electric trash  in module from...and that's not good.


So i rethinked my design, i had to get it tight...but without epoxy.

My options where:

- Milling

- Injectionmolding


Since i dont have a shop and do everything in my living room Mill was not an option.

A DIY Injectionmolder on the other hand....i could place it on the balcony maybe.

So i did that spend way to much time on building my own Injection molding press:

https://hackaday.com/2021/01/01/diy-injection-molding-press/

Only to be then disappointed when the Modules still where not tight.

After that i was so frustrated that i made a lab out of my balcony so i had a place for the Mill.

And with the CNC Mill i finaly managed to make some usable Modules:

https://www.thingiverse.com/thing:4770617

The modules finaly worked.

But why should you watercool your modules anyway?

Well heres is a short presentation:


This project kinda keened me sane during the lockdown, and it made me try out 3 typed of manufacturing i had never tryed before: Resin Casting, Injection Molding and CNC milling.