Mains AC, water and unforseen consequences

[JonJ678]

A while back I decided to watercool my power supply. The forums told me I was nuts, see here. I cheerfully ignored some sound advice and went ahead anyway. Here I present the results.

In short, steel compression fittings, near a piece of copper, with mains AC (with very limited current available), will attack each other.

A considerable amount of the thread is gone, leaving behind what looks a lot like rust. The coloured stuff on the piece of copper isn't algae, I had to chisel it off with a steel ruler. Eat your heart out feser 1, nothing clogs up waterblocks like rust.

Naturally the iron has gone somewhere. A significant amount has gone into blocking up the pipes going to and from the waterblock. These tubes no longer allow water through, and go quite a long way towards explaining why the flow rate through my loop was so crap. At the time I believed my pump had died. I can't imagine it enjoyed the overall experience, but it still seems to be working well. Fair play to laing.

The photo didn't come out great but I think it gets the point across. In the background you can see the scraped clean base soaking in ketchup. I don't know if the copper is gone yet, vague recollections of chemistry suggest the copper should be fine as the iron was plated onto it, rather than the other way around.

I am somewhat discouraged from repeating this experiment.

 

[dan1903]

WTF?
really struggling to get my head around this...
why would you put a load of water right where you are getting all your electricity from

surely you saw this coming?

 

[JonJ678]

Chemistry wasn't really my strong point. I reasoned that 240V but hardly any current, plus copper and iron being similarly reactive, would mean no corrosion problems.

The plan was to draw heat out of the psu using the loop, thus leading to a quieter psu. I didn't think the heatsink would be live, but by the time I discovered this I'd cut the hole in the psu case and felt committed to the idea.

 

[Darkpilgrim]

Watercooling a PSU you must be freaking mad lol

 

[wonko]

Heh, I remember that thread - "mental, but brilliant" summed it up nicely

Now it's broken you can stop wearing all that electrically insulating rubber

 

[PhillyDee]

As said before, still good going!

I am amazed, it must be the induction currents which has essentially corroded the fittings. Is it just the threads that are corroded?

 

[Marvin]

WTF?
really struggling to get my head around this...
why would you put a load of water right where you are getting all your electricity from

surely you saw this coming?

It's not that stupid an idea, Koolance have had water-cooled PSUs for some time now. Not cheap mind.

Kudos for trying it - that's one seriously blocked, well, block...

 

[Stryda09]

I still don't get how it was conducting electricity unless you had it attached to exposed wires. It looks like you have had some serious electrolysis going on though.

 

[Squakingcow]

IIRC from reading about someone else who had water-cooled a PSU there can be heatsinks/metal plates in the PSU that are live, which could be the cause of the problem here.

 

[JonJ678]

I am amazed, it must be the induction currents which has essentially corroded the fittings. Is it just the threads that are corroded?

I'm fairly surprised too. It looks like it's the steel physically closest to the copper which took the hit. One of them is badly corroded, the other might be usable after a clean. The copper had loads deposited by one of the fittings, and was nearly clean on the other side.

It was a piece of copper at mains AC, then a layer of insulating material, then the copper base of the waterblock. The insulation was good, in that I poked it and didn't get shocked, and a multimeter failed to pick up a voltage through it. When the other piece of copper was in place a voltage was induced across the gap. Further down the tubes the next fitting was also live, but didn't have enough current available to notice when I touched it. I earthed the fittings against the case, checked there was no voltage downstream and thought nothing more of it.

I suppose it's a learning curve.

 

[cameraheaven]

It's called electrolytic action! or electrolysis....it will happen when even a minuscule current is present between dissimilar metals that are present and in contact with each other,you can't stop it entirely but you can control it. It can be controlled rather simply with the use of an anode,that is a piece of zinc (half ounce slug would do it) placed in the reservoir , you can also get a product from any good shipping chandler, looks like bright yellow tooth-paste and is a barium-chromate based paste and inhibits galvanic and electrolytic action between dissimilar metal parts even threads...........Tomato friggen sauce!!!..........Your a chemical engineer are'nt you. Yes it does contain acid, are you trying to de-solve your copper fins or just the rust???....for rust use oxalic acid, will only attack the oxide (rust) a few of the better paving cleaners will have this chemical present or even a rust converter from Halfords....Tomato sauce............Gem!

Aussie Allan in the UK

 

[JonJ678]

That's wildly inapplicable as well as patronisingly written Allan. It's evident that you looked at the pictures, but couldn't find the motivation to read the text.

1/ The metals were not in contact with each other

2/ It wasn't passive electrolysis, there was a (considerable) voltage applied

3/ A piece of zinc in the reservoir wouldn't have been an anode, as the reservoir was at ground

4/ Oxalic would have been better, I had none to hand. Ketchup isn't going to eat copper.

Now, I agree that electrolysis will occur in the presence of a current. It'll be proportional to the current. As the voltage switched polarity at 50Hz, and as the current was believed to be negligible, it is a reasonable belief that the steel wouldn't dissolve. It turns out the alternating current electrolysis does occur, I imagine because the system was asymmetrical. Do you know anything about this, or would you rather recite gcse chemistry at me?

 

[Gundog48]

Never do this! Many of you will remember the incident of the person who applied a copper water block directly onto the capacitors where the heat sink went which sent mains electricity all through his presumably conductive coolant basically wiring the whole system to the mains...

EDIT: Apologies, seems that person was you! Nothing like the 'why not' approach of modding that you have, but a cooler and more silent PSU is not something I would like to die for!

 

[ColonelCage]

This is just great, didn't care, gave it a shot, the learn from your mistakes approach. Love it sincerely!

 

[cameraheaven]

Johnj678....You have to understand Aussie humour....."Giday you old ugly Bastered" would be a term reserved for my very dearest and closest friends !!! and would be excepted that way! And I assure you I read every bit of your post. My intention is to learn what I can and if I can , help a few people out on the way, Tomata sauce tickled my humour because tamoto sauce is for Pies and to an Australian what tea is to a P-P-P Englishman!
I'll try and keep my answers a little more simplistic....Whats GCSE anyway?...never covered that text book at uni of NSW....Replies to your replies follow.

1/ Unless your using plastic bolts or air screws the blocks where in fact in metallic contact, regardless of this, dissimilar metals in a fluid environment will act like cathode and anodes and produce a current...and acid in the presence of CO2
2/ If there's an external power source this will just speed up the process your experiencing.
3/ By bolting on a piece of zinc on to the side of a boat or outboard motor just gives the "Action" we'll call it, something much more appetising/attractive to attack and devour instead of of your fittings , make no mistake this is a consumable (zinc) and would need to be replaced periodicity.
4/ Oxalic is hidden in a lot of off the shelf products, you just gotta goes a huntin!...one of the reasons tomato sauce works is it infact contains Oxalic acid which is perfect for desolving rust but leaving copper alone, ...unfortunatly it also contains Citric acid,Malic and Glutamic acid as well...all will attack copper.

So to get away from the bloody chemistry which I in fact hate ...If you have steel and copper in a given circuit with a fluid ........you have an anode, you have cathode, and water.........you have a battery and electrolysis !!! ... Oh Yeh! ..and Acid.