Water Pressure valves ?? are they necessary

[larryccf]

FIrst my specs -
CPU I9-9900K
Mobo: MSI MEG Z390 Ace
Case - Fractal Design Refine R6
cooling system comprised of two radiators, a XSPC TX240 in the "roof" of the case (ultra thin 20mm) and an AlphaCool ST30 280mm in the front of the case, both exhausting out, both with fans in "pull" mode


I'm new to liquid cooling, and thought i'd read all the guides out there, just went thru the one on this forum, and don't see any mention of pressure valves

about a week ago i saw someone post his system, commenting that the pressure valve was a great aid in bleeding off air in his system. I checked and found a number of the suppliers offering them (EKWB, Alphacool & XSPC) but not a lot of explanation of when you'd need one or why

any help in this area would be greatly appreciated


2nd question - i'm an old fart (72) and up until recent was totally adverse to the idea of putting liquid near electronic circuits. What brought me over to LC was A) my 9900k, i could fry eggs on it if i wanted when OC'd and B) numerous web explanations that distilled water and glycerol were "nearly" non-conductive - they may not have used the word "nearly" but they made it clear they were not totally non-conductive. Does anyone ever spray their motherboard with polyurethane to help water proof the exposed circuits - i'm talking a couple of light coats after masking off the ram, PCIe slots etc? Only reason that occurred to me, early in my employment life, i worked as a mechanic at a VW dealership (1960s) and we used to spray the ignition wires with either varnish or acrylic to help insulate them, otherwise on a rainy day, a lot of them would not start, and we'd have angry customers calling in or having their cars towed in



tks in advance

 

[Zerileous]

I don't believe they are necessary. If you could place one at the highest point in your loop, a bleed valve would help you get trapped air out. Not sure if pressure valves and bleed valves are the same, just equalizing the pressure, or if the pressure valve has some sort of "pop off" pressure. I really don't know a lot about them, working on my first custom loop also.

Secondly and probably more importantly, make no mistake, coolant in contact with powered components will likely cause shorting and damage. There is no such thing as non-conductive coolant. Even the coolants sold as non-conductive will pick up ions from the metal in the loop and become conductive over time. The risk of leaks is mitigated by thorough leak testing of the loop with the PC components unplugged from power.

I've heard of using vasaline to protect PCB from condensation during rounds of extreme overclocking with sub-ambient (LN2, etc) cooling, but not of any long term waterproofing solutions.

 

[EarthDog]

Not necessary.

You can spray a conformal coating... but it isnt needed for ambient cooling.

 

[larryccf]

thanks for the replies guys - if no one else is using a pressure valve, i'm going to forget about it - already have too many fittings in my system design

i've worked with silicone (pure silicone) - we hired a retired Dow chemical lab chief as a consultant, and learned a lot about silicone - totally non-conductive, and will not float off in water, it prefers to cling to "aggregate", so it would serve nicely but problem with it is, it's what's called a low tension surfactant (not sure i'm recalling that correctly) but basically it will wick thru out the computer, which would be messy

i'm putting the radiators into a new case, the Fractal R6, so what i'll probably do, before i move the Mobo over, mask it off and spray it with a couple of coats of polyurethane. I know to test for leaks, but we can test till we're blue in the face, and still have an accident or an unexpected leak

 

[EarthDog]

It really isnt necessary to do this silicon spray....but...you can. I've never used it for water cooling.

 

[larryccf]

EarthDog - you're probably right but i seem to live with Murphy (of Murphy's law) sitting on my shoulder. I realize i can't protect or waterproof the PCIe slots or the ram slots, and it's probably a 50% shot that if a leak occurred, it would not hit either of those. But still, it's not much in the way of effort.

i'm not going to use silicone, either acrylic or polyurethane

btw, thanks for posting about conformal coatings - i googled it, didn't realize the electronics industry did employ coatings for various purposes

 

[Robert17]

Also, from your VW days, the wiring harnesses weren't sealed very well, if at all, so a jury-rigged coating at the connector points was a good idea back then. Since those days the technology has improved dramatically with newer insulation and shrinkable sleeves to make them moisture resistant.

You're correct, you can't Murphy-proof your rig, but if you take your time assembling it, pressure/leak test the connections outside the case before final assembly, you'll likely not have to concern yourself unless you drop your rig from height.

 

[EarthDog]

EarthDog - you're probably right but i seem to live with Murphy (of Murphy's law) sitting on my shoulder. I realize i can't protect or waterproof the PCIe slots or the ram slots, and it's probably a 50% shot that if a leak occurred, it would not hit either of those. But still, it's not much in the way of effort.

i'm not going to use silicone, either acrylic or polyurethane

btw, thanks for posting about conformal coatings - i googled it, didn't realize the electronics industry did employ coatings for various purposes

Yes... we typically use a silicon based conformal coating on boards when using sub ambient cooling methods.

I mean, go for it, but this is being a bit paranoid about it all... see Robert's post... we're here to help as well.

 

[larryccf]

if you guys are that serious that it isn't necessary i forego it then

i am building it in the new case as components come in - and when finished will be testing it in the case it's going to reside it, then bring in the motherboard and other components.

 

[maxfly]

ive been wcing for right at 20yrs now and the only two leaks ive had didnt get anything wet inside the case because of a... you guessed it, a thorough leak test. my desktop hutch still has damage on one leg from the 1st leak (2001 or 2) but nothing inside my case got it! that was from a cheesy plastic koolance pump/res barb that was broken during shipping. i found it after an overnight leak test without any gear in the case (silicone caulked the little sob and ran it that way for years). the second leak was from bad xspc 45/90 degree fittings in my last build (2yrs ago) but they were found during the leak test as well. as long as you get your leak test done right AND use high quality gear youll be golden! btw if your using 45/90 degree rotary fittings, if they turn hard as hell by hand they wont leak. if they turn easily by hand throw them away/return them. theres a really good chance they will leak at the rotary joint.

 

[larryccf]

thanks and nope, not using any swiveling 45/90 degree fittins. As i stated earlier, i mechanic'd for a number of years, then chased my hobby for a living, and ended up in mfgring. Just retired - i like to eliminate any fail points possible and those looked like a good item to avoid. In fact, even the screw in 90 degree fittings i stayed away from as i don't like the reduced flows their tight radius will cause. I went with soft tubing

I do have one other question - when testing, like i plan to in the new case w/o mobo or anything electrical, do you heat the radiator(s) or the cpu waterblock to elevate coolant temps to what they'll be in use? It only occurred to me, as hotter water is thinner water plus there has to be some pressure from the expanded water ie water volume when heated

 

[EarthDog]

Temp differences are around 10-20C in properly radded loops. It's not going to make a difference.


Have you read our guides? They may help fill in some blanks.

 

[maxfly]

as long as your fittings are tightened down properly, temp fluctuations shouldnt change anything. its never a bad idea to do your leak test as close to real world as possible but it isnt necessary. if your using barbs and worm drive clamps you should never have to worry about fitting leaks of any kind. just dbl check your clamps are tight prior to filling hehe yeah ive done that...

haha ED beat me to it!

 

[Robert17]

I've got a couple of these if you'd like to give 'em a try.........or you could find them at some parts stores. Good for taking apart a loop for cleaning.

 

[larryccf]

EarthDog - actually i went thru them first looking for info on the pressure valve - don't recall seeing any mention of it or the other two items i asked about. But i've read so many guides, by my calculations, one more guide and i should at least qualify for a B.S. in custom loop design lol

- - - Auto-Merged Double Post - - -

actually i'm going with compression fittings, but since we're on the subject, don't know if you guys are aware of mcmaster.com. They are to manufacturers what sears & roebuck used to be to the commercial market - i don't care what you're looking for, they'll have it or find it for you. Prices are retail, but service and availability are what they offer. I've ordered from them on Monday (their nearest warehouse is NJ) and had it here on Tuesday or Wednesday the latest UPS Ground. Metric or inch standard, any type of clamp, 3000F ceramic caulk you name it

when i got my hands on their catalog, it was 3500 pages, small print - i spent till after midnight trying to digest all they cover. Now they're on the web, it's a whole lot easier to find stuff

 

[JLK03F150]

Those thumb-turn clamps look nice Robert.

I use a combination of the auto store worm drive clamps along with Swiftech's clamps. The only leak I ever had (since ~2006) was a barb I failed to fully tighten in the radiator of my FX 8320 system. No water got onto electronics though, just dripped into the bottom of the case. whew!

Oh, and to the forums Larry!