Project: Omega [56k warning]

More of an update should be forthcoming in the near future. Started taking it apart to get a photo and discovered that the tubing coming into the 10W pump is pretty much flat. This suggests lower pressure in this pipe, and the tubing recovers if the 10W pump is turned off. Not worked out what this implies yet. Probably just that I don't want a long, curved, flexible intake tube going into the impeller of a ddc, but it also suggests that the pressure in parts of the loop is fairly high. This probably means it's a good time to put things in parallel.

Some research into centrifugal pumps suggests I'll be alright with two matched pumps adjacent to each other, so I think I'll try that before any extensive rebuilding. Need to get the psu out somehow (difficult) to change the pvc tape for a K200 pad. Have another radiator on the way for good measure too. Realized I don't have any pictures up of the assembled case yet, so should change that.
 
Last edited:
Computer spent last night folding. Air coming out of the psu is definitely cooler than it used to be, but warm enough that I don't think I want to disable the fan. Shall have to have a think on that one, it's possible I've mounted the block badly. It's not dead yet though, so it seems induced voltages in a water loop don't matter.
 
I'm going to play devils advocate here and say that unless you have the electrical knowledge to be able to predict what will happen when you mod something like a psu, e.g. whether some part of the case is going to become live with 240VAC or not, then its an accident waiting to happen. Sorry to sound harsh but from reading the above its sounds like allot of trial and error is going on which surely is unbelievably risky when dealing with potentially 240VAC flying about? Trial and error is generally not to be advised with electrics I would say.

I'm not an electical engineer either but from what I have gathered the power transistors in the psu will either be n-p-n or p-n-p. On one of these the aluminium heatsink will be at ground and therefore safe to connect other sinks to. The other however will be at the full voltage of the transistor. Sounds like you have the non ground ones in yours. If these are short circuited to ground as you seem to have done then what is happening to the power supply, in the short and long term? Unexpected current loops are being set up as a minimum.

The fact that you appear to be relying on pvc tape to protect you from mains voltage sends shivers down my spine tbh, what happens when it becomes nicked, or deteriorates or breaks down?

I may be wrong and you may know what you are doing but tbh it *sounds* like your obvious enthusiasm and inquisitiveness is leading you into unkown terrirory that could kill you. Just my gut reaction.
 
You're not the only one who's taken this approach, I thank you for your warning much as I thanked the others. I shall attempt to set your mind at rest.

I think you'll be pleased to hear that the pvc tape is gone, although I did check it was rated for mains AC and layered it very carefully. I've no idea what the odds of deterioration were though, so I'm glad it's gone. I'm now using this, a 50mm piece was very kindly sent to me by PhillyDee of these boards. So while I agree wholeheartedly that I'm approaching this in trial and improvement fashion which may be unwise, at least I'm using the right material now.

The base of the waterblock has an induced voltage in it equal in magnitude to the 240V ac running a couple of mm away from it. This is a very weak current source, but doubtless it is leading to current. This results in a charge on the water. The barbs are therefore live, and therefore wrapped in electrical tape. Fortunately the waterblock is acetal, which is an excellent insulator so avoids a direct path to ground through the psu casing. As far as my multimeter can tell tygon is an electrical insulator. Tygon themselves state a dielectric strength of 19.3 kV/mm, it's 1.5mm thick. Again induced voltages could be an issue, I'm relying on the charge in the water being sufficiently negligible that this isn't a problem. At each end of this tubing is a radiator which is convincingly screwed into the case. The casing the radiators are attached to is grounded (multimeter verified) via the psu casing.

Now, if things go wrong. Ignoring briefly dead hardware, which would be gutting. Worst case scenario puts the casing of the computer at mains ac. Briefly. Short circuit is established between mains and ground. PSU has a fuse, cable has a fuse, system has a circuit breaker attached. I've verified (accidentally) that connecting the heatsink direct to ground hurts and trips the circuit breaker, psu continues to work much as if nothing had happened. So in a sense the worst has happened (I've had three fairly nasty shocks off this by connecting heatsink to casing with my finger) and no particular harm occurred.

I have a fairly good idea what I'm doing, the only real surprise so far was that the heatsink was live in the first place. I hope I've been clear throughout that I strongly recommend no one else attempts this, as it is clearly foolish.
 
Last edited:
JonJ678 said:
I have a fairly good idea what I'm doing, the only real surprise so far was that the heatsink was live in the first place.
Click to expand...

That is a pretty substantial and significant "surpise" considering you say you think you know what you're doing. I would imagine it is a fairly fundamental concept for anyone who makes power transmformation euipment, i.e. sparkys.

That stuff you have sourced which says it has a 1000V breakdown voltage. Fair enough, looks like it would work in theory. However, a scenario for you:

I place a metal plate in front of you, attached to the plate is a wire going off to a plug in a wall marked 1000V. On the plate is a some of that stuff. I show you a piece of paper saying "the film shows no known deterioration over time". You have no multimeter to hand. I ask you to touch the film. Would you touch it?

Unless you're constantly going to be using a multimeter whenever you go near your pc this is effectively the choice you are making whenever you touch you pc.

Then again, personally I would go back a couple of stages and ask myself would I even bother to water cool the psu? Especially as you seem to still have a fan on it. The psu is designed to cope with the temperatures that it self generates, it will have a mtbf of umpteen thousands of hours. I'm really struggling to see what the point of it is, other than a possible aim of just to water cool everything in a pc just because it can be done.

I suppose you should be thankful that BS 7671 doesn't appear to apply to bodging your computers psu, yet :)
 
A belated response to the above. I'm not a sparky, more of an applied physicist really. I don't know if I'd touch said wall or not, curiousity might overcome reason or it might not. I wonder if the film degraded, or whether it was simply induction. I hope I haven't lost the material. The time dependent aspect didn't occur to me.

Not much going on with this, need to spend some quality time with a Bridgeport really. The water-cooled psu approach ended badly though, see this thread. I don't think I'll try it again. I'll post the pictures here as well.

2193tiw.jpg

21eyvc6.jpg

1ht8hz.jpg
 
You must log in or register to reply here.
Back
Top Bottom