Athlon 245 Under DICE And How Not To Make A Pot (56k Warning)

[apt403]

So we (DrakeZ and I) managed to ghetto up a Cu pot for some benching... It wasn't as successful as we had hoped, but the preliminary results were decent enough. After we get the pot patched up, we're going for another run, only with neoprene insulation on the board and mounting plate instead of paper towels, and with the thing properly brazed together.

The main components of our ghetto pot:



Test assembly:



DrakeZ cutting out the mounting plate:



Almost there:



Clamping together the top half:



Epoxy to get everything to stick:



Interior shot:



Testing the pot for leaking and to see whether or not the epoxy would just shatter completely:



Air benching with the Mugen:



Insulated and mounted:



All DICE'd up and ready to play:



DrakeZ on DICE pouring duty:



Lowest the temp sensor went, then the isopropyl froze in the tube and the temps skyrocketed.



Here's what happened after the epoxy got eaten by paint thinner (A 2in section of copper really doesn't have a lot of load capacity...):



Max speed I got to before all the insulation started melting the VRMs and everything died:



And the two of us afterwards, I'm on the right:



Highest I could get validated:

http://valid.canardpc.com/show_oc.php?id=639460

We learned 4 things from this:

1.) If you try to use epoxy for a DICE pot, the epoxy cracks and everything falls apart.

2.) You should really braze copper together...

3.) When your VRMs get covered in insulation, they get really hot... And then Windows crashes.

4.) 70% isopropyl alcohol really doesn't work. Paint thinner is better, but then your epoxy starts to dissolve.

So tomorrow if I've got time between class, we're going to braze together the pot, then at some point get more DICE. This chip has at least another few hundred Mhz in it, just waiting to be eeked out.

Oh yeah, and here's how you REALLY don't make a pot: