[Gooserider]

bigben2k: Since you'll be making your own NB and GPU block, the actual flow resistance isn't known, but in what's commonly available, it isn't hard to see that even a DD Z-chip block would be less restrictive than a lot of waterblocks. If you design your NB and GPU blocks specifically to be restrictive, then that'll change everything.

Well, I'm designing my CPU blocks to be non-restrictive actually. While I'm not specifically trying to make the NB and hard drive blocks restrictive, I don't see how they couldn't be to a fairly large degree - I'm going to be making them with 1/4" ID tube, (CPU's get 1/2") with fairly long passages that will have as many bends as I can figure out how to fit on the drive plate (I'm going for simplicity - bend tubing into a maze, solder the maze onto a 1/8" x hard drive sized plate. No milling, no seams, should be no leaks, efficient enough to cool the drive.) I am guessing each plate will have between 1 and 2 feet of tubing, and there will be at least 3 plates, and possibly an NB block (which would be relatively restrictive internally, but still use 1/4" plumbing)

sevisehda: Failure recovery. If your flow slows to lets say half more than likely your temps would rise X amount and level off at some point. If your pump were to fail then the time would be dependent on the thermal mass of the WB and the water inside it. So a small cooler would die faster a large one would give you more time. Without doing any math I'd guess you'd have about 10-20 seconds.

Thanks for the input on failure recovery, it is somewhat reassuring. Big is relative, but I think I'll be on the heavy side of that set of numbers. I might do a limited simulation test once I get things built to get more definite information.

A minor planning concern I just had - I haven't dealt with one of the modern ATX type power supplies from user space yet. Do you know how it behaves when the AC input line is interrupted and then restored? Obviously it goes off when the power does, but when the power comes back does it turn back on and try to bring the PC back up?

I'm planning on implementing a 'last chance kill' relay that would interrupt power to the PC and pump if it triggered. If the supply stays off, then I only need a momentary relay. IF it trys to come back up, I would need a latching relay that would keep the power disconnected until I reset it.

jaydee116: Couldn't you just make a bypass valve around the rad to keep flow going. Kinda what aspirit did with his rack mount deal. Kill two birds with one stone. More flow rate through the blocks without giving up serial routing. Wonder if the loss of cooling would defeat the purpose.

Depends on what your purpose is If it's to get lots of flow, it would probably work. If it's to cool the system, it seems rather pointless to bypass the cooling loop. I haven't seen Aspirit's setup, but if its a rackmount (presumably w/ multiple processing boards) I'd suspect the function of his bypass valve is to isolate a part so that it can be serviced without having to shut down the entire rack. At least if *I* were doing something like that (and I used to work on high reliability rackmount setups for telcos) it is the kind of thing I'd do.

Gooserider