Basic watercooling questions

[BrainEater]

Woop !

I did a test of my watercooling setup, d5 vario from a reservoir of my own design to dual socket 3647 cpu blocks.
Seems to work quite well, however I have some questions.
---

So a G1/4 watercooling fitting uses the o-ring to provide seal. The threads only provide 'compression' for the O-rings.Is there any reason NOT to use something like a medium strength threadlocker to keep things from unscrewing ?

--
Radiators :
I have to assume if you are using a radiator horizontally , barbs 'down' is bad.

I'm sure I'll think up a couple more dumb questions ,

TYIA all the same !

 

[Zarathustra[H]]

I've never heard of g1/4 threads "loosening". Probably wouldn't hurt anything to use some thread tape or something, but it is unlikely to help much either.

You'd just be wasting your time and energy

 

[cyberguyz]

The o-ring alone should be fine. I would use some thread tape myself being careful to keep it away from the silicone o-ring and its seats. Threadlocker well, I guess it depends on where the fitting is sitting. If the joint is where there will be lots of vibration like a rad with fans running or a particularly unbalanced pump I might consider a small drop of blue locktite. Vibration and heat changes can loosten tight-ish joints.

 

[Zarathustra[H]]

The o-ring alone should be fine. I would use some thread tape myself being careful to keep it away from the silicone o-ring and its seats. Threadlocker well, I guess it depends on where the fitting is sitting. If the joint is where there will be lots of vibration like a rad with fans running or a particularly unbalanced pump I might consider a small drop of blue locktite. Vibration and heat changes can loosten tight-ish joints.

It's possible. Hasn't happened in any of my builds, but possible I guess.

I'd be worried about getting them too stuck though, and making them difficult to remove down the line.

When I recently serviced my loop I I couldn't remove one of the fittings with my bare hands. I tried using a large screwdriver in the slot in the barb, but that just broke the barb. Eventually I got it loose with a large plumbers wrench I had handy, but I then had to replace the fitting. Luckily I had spares.

I didn't use any kind of thread lock on my fittings.

 

[cyberguyz]

Same in my case. I am looking back at a time when there were no G1/4 threads, o-rings and compression fittings. Back then I used silicone sealer for the threads. The wc builds I am doing now are the first custom ones i'm doing in about 8 years (though I'm loaded for bear with parts of every type)
Locktite comes in different grades. The blue one is considered the mildest and you only use a very small drop of it. It is meant to be easily re moveable.

 

[owcraftsman]

I would recommend agaist thread lock of any kind including tape paste or drops and be more concerned about how the chemicals of those products would interact with your loop components, biocide or whatever you are planning to use and the water clarity long term. I haven't used anything like that for 15 years of watercooling and never had a leak.

 

[Zarathustra[H]]

I haven't used anything like that for 15 years of watercooling and never had a leak.

I haven't been water-cooling as long as you, but I too have never used anything like that, and the only leak I've ever had came from defective XSPC swivel bends, not from any G1/4 threads.

 

[VanGoghComplex]

No threadlock and no tape. The fittings are designed to seal on the o-ring, and if your installed fitting is undergoing stresses sufficient to unscrew it, you're doing something else very very wrong. As for tape, BSPP is not a tapered pipe thread. Tape will not help it to seal, and potentially will just put little shreddy-threads of teflon in your loop.

You can mount your barbs in whatever direction you want. If your tubing fits over your barbs loosely enough that they're at risk of falling off from gravity alone, you're doing something else very very wrong.

There aren't any dumb questions. Keep 'em coming! =)

 

[Zarathustra[H]]

There aren't any dumb questions. Keep 'em coming! =)

Agreed. This is a judgment free zone, and a lot of this stuff is difficult to wrap your head around until you have parts in hand and then all of a sudden it makes a lot of sense.

Building your first loop can be a bit overwhelming, but once you e done it the concept is fairly straight forward.

So, I'm not a Jaystwocents fan, and this video below is a little old, but it is still the most comprehensive summary of what it takes to put together a custom loop, and nothing much has changed since he made this. Every first timer should study it.



Edit:
Only thing I disagree with him on is the use of distilled water and PTnuke. Distilled water is great, but PTNuke has copper ions in it and that can accelerated corrosion if you have nickel coated blocks. Also adding ions to the distilled water makes it conductive, which can make any leak worse. Similarly I wouldn't use a silver kill-coil as it can also contribute to galvanic corrosion.

If you want to DIY the coolant, the best choice per Watercools rep on these forums Watercool-Jakob is plain distilled water with a glycol based anti-corrosive which also serves as a bio/fungicide.

Personally I like EK's Cryofuel concentrates. Just mix them with distilled water for a dyed product with an optional mix of anti-corrosives, biocides and surfactants that works very well and has perfectly protected my system for years . They are pretty cheap on Amazon too.

 

[VanGoghComplex]

Agreed. This is a judgment free zone, and a lot of this stuff is difficult to wrap your head around until you have parts in hand and then all of a sudden it makes a lot of sense.

Building your first loop can be a bit overwhelming, but once you e done it the concept is fairly straight forward.

So, I'm not a Jaystwocents fan, and this video below is a little old, but it is still the most comprehensive summary of what it takes to put together a custom loop, and nothing much has changed since he made this. Every first timer should study it.

Agreed on Jay. I used to be a fan of his, but he's been a bit of a turd lately. His older watercooling videos do a great job of covering the basics tho. Every now and then he drops a new vid that covers WC stuff, but lately his content is not very practical. =P

 

[Arcygenical]

I've been using "barbs down" fittings for around 5 years for a 3.120mm radiator. No leaking issues. I tighten to hand tight, then give another 1/4 turn using a wrench. They never, ever, move.

 

[VanGoghComplex]

I've been using "barbs down" fittings for around 5 years for a 3.120mm radiator. No leaking issues. I tighten to hand tight, then give another 1/4 turn using a wrench. They never, ever, move.

If it worked for you it worked, but I'd advise against ever using a wrench on watercooling fittings unless you're out of options to loosen a stubborn one. They seal on the o-ring - hand tight is enough.

 

[cyberguyz]

If it worked for you it worked, but I'd advise against ever using a wrench on watercooling fittings unless you're out of options to loosen a stubborn one. They seal on the o-ring - hand tight is enough.

In my own case I have arthritis in my hands so my grip & dexterity is not as strong as it once was. A little mechanical helper is mandatory in my case (used with extreme cawshun of course!!).

 

[VanGoghComplex]

In my own case I have arthritis in my hands so my grip & dexterity is not as strong as it once was. A little mechanical helper is mandatory in my case (used with extreme cawshun of course!!).

Sure thing! If you need it you need it. For me, finger tight is already shoulder-to-port, and the idea of turning those thick-pitched threads another 1/4-turn fills my ears with the sound of cracking plexi!

 

[Zarathustra[H]]

Sure thing! If you need it you need it. For me, finger tight is already shoulder-to-port, and the idea of turning those thick-pitched threads another 1/4-turn fills my ears with the sound of cracking plexi!

This is why I hate plexi tops.

I wish all blocks had their G1/4 threads in metal. My EK Supremacy EVO all copper edition does, and all my radiators do (obviosuly) but my pump and my GPU fullcover block are both plexi, and I hate that.

 

[VanGoghComplex]

This is why I hate plexi tops.

I wish all blocks had their G1/4 threads in metal. My EK Supremacy EVO all copper edition does, and all my radiators do (obviosuly) but my pump and my GPU fullcover block are both plexi, and I hate that.

I'm torn on plexi. On the one hand, an easy peek into the channels of my GPU block gives me peace of mind, but yeah... So easy to crack. x_x

 

[Zarathustra[H]]

I'm torn on plexi. On the one hand, an easy peek into the channels of my GPU block gives me peace of mind, but yeah... So easy to crack. x_x

Actually, I take my comment back. I have no plexi in my loop. I have plenty of black plastic (Acetal?) though. Either way, I'd prefer if it were all solid metal.

 

[BrainEater]

Thanks folks , I appreciate the input a lot !!
-------

I have an odd bias when it comes to watercooling.
I run a half million BTU glycol loop for my brewery, but I have never done a watercooled PC.
The two are not the same in any way really, hence some questions.....

----

I'm going to use the fittings as-is....no threadlocker.I was kinda just curious about this.
----
Arcygenical , the 'barb's down' thing is in relation to air gaps in the rad.......Fermenters hafta run input on the bottom, or we get air locks.

-------

I normally run 50% propylene glycol/water at the brewery, the loop normally runs at -4 C.Biocide is not normally an issue for me.
I am looking at running mebbie a 10-15% propylene glycol/water mix....I don't want my loop running below dewpoint ever, but I want freezeproofing.......a quick -5c blast of air into my pelts could drop the temp fast.

My next question , biocide..... Silver kill coil would be my choice.....Any opinions ?

------

Last question for now
Galvanic corrosion related effects.

In this loop I have a copper rad, my custom resevoir/chiller is aluminum , I have brass,stainless,and nickel plated brass.
To me , everything is expendable except the 2 cpu blocks.
Galvanic corrosion requires electrical conduction, I can electrically isolate my aluminum reservoir no problem, but I'm not sure about the copper rad.
Should I worry ?

Cheers everyone , thanks again.
It's my sincere hope to wrap this up in the next two weeks.
It is, after all, ski season.

 

[VanGoghComplex]

Just to be clear, are you planning on chilling the water? As in sub-ambient cooling? Because if you are, you get into a whole new ballpark of crazy: mitigating condensation on/near your powered motherboard.

Also, running such a diverse mix of metals is begging for trouble. I'm not a chemist but I've gathered that the aluminum and copper in your loop will attack each other, and nickel plated things will also war with silver, so forget the "kill coil." I've never heard of electrically isolating things to prevent galvanic corrosion - again, not a chemist - but how can you electrically isolate your copper rad from your aluminum res if they're both in contact with the same volume of water?

You can probably mitigate the corrosion risk for some time by using a strong anticorrosion additive, but that's kindof a band-aid. If you're interested in getting any long-term service out of the thing, I'd nix that aluminum reservoir.

 

[BrainEater]

Yes.
I am chilling the coolant. Custom peltier reservoir.

But just a bit......as I said, never below dew point.
In reality , I just want 20 C all the time.I don't think I have enough power, but we'll see.

edit :

I tried for 8 months to find copper /a big enough and b/ a welder to make my reservoir from pure copper.
The cost to that exceeds the already insane cost of the 20k rig.

 

[Tsumi]

Any particular reason why you want to keep it at 20 C? If you're doing it just because it's a nice number to be at, I would can that idea right away. Just throw more radiators and more flow at your system, it'll be more cost effective.

 

[VanGoghComplex]

Why not use one of the non-metallic, inexpensive, commercially available reservoirs built specifically for watercooling?

I know it's not a big deal with something like the chiller in your brewery, I'm betting the waterways in that equipment are all large enough that a little corrosion won't hurt performance much... but you do not want to mix metals like that in a PC loop. You said the CPU water blocks are the parts you care about? Well, those blocks are the parts that corrosion will hurt the worst. The fine machining in the microfins can get wrecked by a minimal amount of corrosion, permanently hampering the performance of the block.

I might be worrying about nothing, and you seem to have a fair amount of technical acumen. The chiller and phase change stuff is well beyond my experience level, too. I think I'll bow out of this one, as long as you promise to provide pics of this crazy contraption you're building. =D

 

[cyberguyz]

propylene glycol as an additive is fine to resist galvanic corrosion. Aluminum has a high anodic index of .75-.95 volts. Copper is much lower index of .35v with silver and copper/nickel alloys sitting at .15v. Brass is sitting around the .4v - .35v range. The ideal system be made of those metals that have the lowest anodic indexes and stay away from the unplated aluminum parts as much as you can.

Here's a useful link to a pdf with a blurb that relates to galvanic corrosion to use when shopping about for parts.

 

[Zarathustra[H]]

propylene glycol as an additive is fine to resist galvanic corrosion. Aluminum has a high anodic index of .75-.95 volts. Copper is much lower index of .35v with silver and copper/nickel alloys sitting at .15v. Brass is sitting around the .4v - .35v range. The ideal system be made of those metals that have the lowest anodic indexes and stay away from the unplated aluminum parts as much as you can.

Here's a useful link to a pdf with a blurb that relates to galvanic corrosion to use when shopping about for parts.

I've always used this table on Wikipedia.

Whenever there is a difference between the anodic index of metals the more anodic of the two will corrode (and the more cathodic of the two will collect oxides and gunk from the corrosion). The larger the difference, the faster the corrosion. For a very small difference in anodic index, the corrosion is slow enough to be negligible.

Design guidelines are typically to keep normal dry indoor differences to 0.25v or under. In wet or harsh environments, that goes down to 0.15V. This is the figure I use for water cooling loops.

Most loops use copper, so Nickel, Titanium and Brass are all safe.

The difference between copper and aluminum is 0.55 to 0.60V, so this is clearly very bad.

This next one is controversial. Many people use a silver kill coil to keep bacterial growth under control, but silver and copper have a difference of 0.20v, which is above our 0.15v target, so IMHO, silver is best avoided.

Adding corrosion inhibitors to the coolant helps reduce/slow these corrosive effects, but I still wouldn't really stray beyond the 0.15v difference.

 

[Zarathustra[H]]

My next question , biocide..... Silver kill coil would be my choice.....Any opinions ?

As mentioned above, you have to be careful with this. PC Water cooling blocks have microchannels and jet plates not present in larger loops (like those in cars) so even very small levels of corrosion can be a problem. I would just avoid adding a silver coil.

Dow's pages on propylene glycol (as well as WATERCOOL-Jakob , Waterkool's rep on these forums) have suggested that the propylene glycol in and of itself can have a anti-microbial effect.

Dow's product pages suggest it is best to have 20+% propylene glycol for this effect.

Apparently in sub 1% concentrations, bacteria actually feed on propylene glycol.

In the 1%-20% range, some bacteria still survive.

Above 20% it appears to be an effective growth inhibitor. 25% or above is ideal.

Honestly, I'd normally recommend just going with a pre-formulated fluid, like EK's Cryofuel, as they are both relatively cheap and effective, already being ideally formulated with the best balance of anti-corrosives, surfactants (to help get air bubbles out of the loop) and biocides to maximize loop performance. That way you don't have to worry about corrosion. They come in many colors, here is the clear concentrate to make 1L of coolant on Amazon for $14.50 with Prime. I don't know if they have enough propylene glycol or other anti-freeze like components in them to make them resistant to freezing though, so they may not be ideal for your corner case.

For your application, I'd just go with 25% propylene glycol (or higher if you need to protect against freezing) and 75% distilled water. The Propylene glycol will be a corrosion inhibitor, growth inhibitor and a surfactant all in one, so it looks like the ideal silver bullet for you.


Edit:

Interesting sidenote: It is actually the surfactant properties of propylene glycol that make it a growth inhibitor. At 25+% concentration by volume it has so little surface tension that it breaks down the bacterial cell walls and kills them.

 

[Tsumi]

9 years of using a silver killcoil without any observed problems. Mine sits at the bottom of my reservoir.

 

[VanGoghComplex]

9 years of using a silver killcoil without any observed problems. Mine sits at the bottom of my reservoir.

Got nickel plated blocks?

 

[Zarathustra[H]]

9 years of using a silver killcoil without any observed problems. Mine sits at the bottom of my reservoir.

I agree that the risk of a silver kill coil is small, especially if you use a corrosion inhibitor in your fluid.

A small risk is not no risk though. There will be a very slight slow corrosion, but depending on your blocks it may not have a noticible impact.

There are a wide variety of radiators, blocks, and fittings out there. Chances are that like in your case a silver coil won't cause any issues, but personally I'd rather be safe than sorry just in case I happen to have randomly bought parts that are more sensitive.

Of particular concern with silver may be any brass fittings or adapters that are in the loop. These are rare, as most seem nickel plated these days, but some people use them, and then you have a potential difference of 0.30v.

Copper blocks with very fine small machined microchannels could also be more sensitive over time. The potential difference between silver and copper may only be 0.20v but even very slow corrosion over extended periods of time can damage very small copper fins at a microscopic level.

Corrosion inhibitors are a good idea, but at the same time it is not a good idea to be overly reliant on them, because they do break down over time and become less effective.

This is one of two reasons why replacing fluid regularly is a good idea.

1.) Corrosion inhibitors break down

2.) Previously non-conductive fluids become conductive over time as they pick up ions from the contact metals making any potential leak riskier.

Typically they seem to recommend replacing fluid annually. I just did mine after two years and three months and it doesn't appear any worse for the wear, but your mileage may vary.

For reference my loop has exposed copper (radiators, CPU block) nickel (GPU block coating, and fittings) and no silver coil.

That's what I know of. Who knows what kind of solder they used in the radiators, but Alphacool have been doing this a long time, so I'm pretty sure they know what solders to use in a radiator.

 

[Formula.350]

I can't really offer much here, except that if you want a good real world example of why bare aluminum is bad in a water cooled application, then automotive and Apple's original Power Mac towers that were water cooled. When vehicles started using Aluminum cylinder heads, or even just aluminum thermostat housings, lots of problems started happening. Mind you, this is on average-Joe vehicles, not performance ones with people who constantly change fluids, etc. But yea, the galvanic corrosion mentioned is crazy and the aluminum just turns to a white powder.

In Apple's situation, I believe it even resulted in holes in their pipes, but most of the leaks were just in the rubber-hose over aluminum barbs that would corrode, expand and start to weep. A system I was given suffered this and the motherboard had a couple drips on it and caused further issues when it interacted with the copper, corroding traces completely. It was nuts.

But yea, with General Motors, they came out with DexCool coolant that was specifically designed for systems with aluminum, going as far as making the coolant orange to differentiate it. The corrosion would clog up radiators in vehicles, so it can become a problem there, too. As Zar said though, the microchannels in the waterblocks are super tiny compared to radiators, but it goes to show that it'll cause problems regardless.



On that note, BrainEater, why not get a copper radiator that has copper tanks and affix your TECs to that instead? (vs the custom copper res)

 

[Tsumi]

Got nickel plated blocks?

No nickel plated blocks, but do have nickel plated fittings and quick-disconnects. No bare brass though.

I agree that the risk of a silver kill coil is small, especially if you use a corrosion inhibitor in your fluid.

A small risk is not no risk though. There will be a very slight slow corrosion, but depending on your blocks it may not have a noticible impact.

There are a wide variety of radiators, blocks, and fittings out there. Chances are that like in your case a silver coil won't cause any issues, but personally I'd rather be safe than sorry just in case I happen to have randomly bought parts that are more sensitive.

Of particular concern with silver may be any brass fittings or adapters that are in the loop. These are rare, as most seem nickel plated these days, but some people use them, and then you have a potential difference of 0.30v.

Copper blocks with very fine small machines microchannels could also be more sensitive over time. The potential difference between silver and copper may only be 0.20v but even very slow corrosion over extended periods of time can damage very small copper fins at a microscopic level.

Corrosion inhibitors are a good idea, but at the same time it is not a good idea to be overly reliant on them, because they do break down over time and become less effective.

This is one of two reasons why replacing fluid regularly is a good idea.

1.) Corrosion inhibitors break down

2.) Previously non-conductive fluids become conductive over time as they pick up ions from the contact metals making any potential leak riskier.

Typically they seem to recommend replacing fluid annually. I just did mine after two years and three months and it doesn't appear any worse for the wear, but your mileage may vary.

For reference my loop has exposed copper (radiators, CPU block) nickel (GPU block coating, and fittings) and no silver coil.

That's what I know of. Who knows what kind of solder they used in the radiators, but Alphacool have been doing this a long time, so I'm pretty sure they know what solders to use in a radiator.

Just checked my reservoir. I have an Alphacool Repack dual D5 bay reservoir, which I believe has brass inlet filters (flow distributors?) on the inside. The killcoil has some tarnishing, and the filters show some signs of corrosion. This is on a system where the water hasn't been changed in almost 3 years and running straight distilled water.

 

[Zarathustra[H]]

No nickel plated blocks, but do have nickel plated fittings and quick-disconnects. No bare brass though.

Just checked my reservoir. I have an Alphacool Repack dual D5 bay reservoir, which I believe has brass inlet filters (flow distributors?) on the inside. The killcoil has some tarnishing, and the filters show some signs of corrosion. This is on a system where the water hasn't been changed in almost 3 years and running straight distilled water.

Interesting. I'm curious, when you say "some corrosion" what does it actually look like?

I'm basing all my comments on engineering textbook type materials, but I have actually not seen it in person.

 

[Tsumi]

Interesting. I'm curious, when you say "some corrosion" what does it actually look like?

I'm basing all my comments on engineering textbook type materials, but I have actually not seen it in person.

I suppose discoloration would be the more accurate term. The killcoil looks like it has some yellowing on it with some tiny black spots near the edges. Brass appears to be greening with some slight pitting.

 

[Zarathustra[H]]

I suppose discoloration would be the more accurate term. The killcoil looks like it has some yellowing on it with some tiny black spots near the edges. Brass appears to be greening with some slight pitting.

Ahh. That's pretty much what I expected.

It all depends on how long you want these parts to last (and how rigorously you want to clean everything after)

I decided to go the abundance of caution route with mine as I want to reuse as many parts as possiböe in future builds.

 

[Formula.350]

I suppose discoloration would be the more accurate term. The killcoil looks like it has some yellowing on it with some tiny black spots near the edges. Brass appears to be greening with some slight pitting.

The killcoil is either pure silver or is a well-plated piece of metal yea?
If so then that looks like tarnishing (oxidizing) to me, which I don't think is anything to be too concerned over.

Take a look at the first few rows of image results to see what I mean. Silver tarnishes with a color of anything from a yellow hue at the start, to a very dark black/brown over time.

Beyond that, I'd speculate the color depends on the purity in addition to the gases its exposed to (aside from 'normal air', so like smog). Considering it's always submerged underwater in this case, the oxidation wouldn't be as severe.

 

[Tsumi]

Considering the relatively minute amount of damage that has happened to the brass fittings (the ones most likely to be damaged) sitting right next to the killcoil after two years, I think I'll keep using a killcoil and distilled water. Also, the brass will act like a sacrificial anode? cathode? so my blocks should be fine.

 

[BrainEater]

Any particular reason why you want to keep it at 20 C? If you're doing it just because it's a nice number to be at, I would can that idea right away. Just throw more radiators and more flow at your system, it'll be more cost effective.

No reason at all.
It's an active cooler so I picked a number to shoot for.
I don't think it has enough power/surface area to get there anyways.
------

Why not use one of the non-metallic, inexpensive, commercially available reservoirs built specifically for watercooling?

It's a custom reservoir, designed to be a peltier chiller.I made it from scratch,to use up a bunch of my crap from my workshop.
Here's a pic ;



----
I had intended to fabricate the reservoir from quarter inch copper , but I could not find a 400 amp TIG machine.I ended up having to use the next best heat conductor , aluminum.
There is nothing cheap about it, just the parts in that picture are worth about $ 1200 US , not including my time.
The PSU shown is a meanwell RSP-1000-27 .It's controllable from about 13volts to 30.5 volts (the pelts are 31 volts, 17 amps max, each).

------

Zarathustra[H] ,

Good info on Propylene glycol !!! I did not know about the antimicrobial properties of it....very cool.
Thanks !!

No wonder I never see crap in my coolant at the brewery.I've got a 55 Gal drum of clear 100% propylene glycol on the way already for the brewery, so that's done.
-------

As far as the galvanic corrosion goes ;
Galvanic corrosion requires a current loop.....The aluminum reservoir part is already electrically isolated, so it will not be an issue.
It's trivial for me to electrically isolate my copper rad too, I'll do that and be done with it.
-------

I've just taken the whole reservoir assembly apart for final lapping and insulation, I'm putting it all back together this week.

Thanks again everyone !!
Cheers.


I'll probably have another question or two as I do final assembly

 

[BrainEater]

Alright , alright , alright.

getting to important questions :'Fillport/drain' tee's
and
Waterblock airlocks.
-------

It is my intention to put inlet/outlet tee fittings on either side of the reservoir.....any opinions ?
----
Airlocks ;
So with most cpu's the chip is actually held in with a lockdown clip. The cpu retention mechanism is independent of the HSF assembly.
My socket P Xeons are not like that....it's the HSF/waterblock that holds the chip on.
Ultimately, this means I cant (easily)fill the loop and 'move-it-around' until the air bubbles are gone.....the waterblocks are part of the mobo, and I dun want to shake that around.
Worse yet, my cpu blocks are the highest point in the loop, air will collect there.

Advice on air elimination ?

 

[VanGoghComplex]

1.) For fill and drain ports, the fill port is only necessary if you do not have easy access to a top port on your reservoir. The drain port should be as close to the bottom of the entire loop as you can get it. If the outlet of your res is the lowest point, then a T fitting and ball valve located there makes perfect sense.

2.) If you put your D5 Vario on the highest speed, you will not need to worry about air getting locked in any blocks. Those little pumps move more than enough fluid to push out any by sheer volume of liquid flow. I never do the whole "case gymnastics" thing, and my loop always bleeds itself naturally over the course of a couple dozen hours.

 

[Zarathustra[H]]

No reason at all.
It's an active cooler so I picked a number to shoot for.
I don't think it has enough power/surface area to get there anyways.
------



It's a custom reservoir, designed to be a peltier chiller.I made it from scratch,to use up a bunch of my crap from my workshop.
Here's a pic ;


View attachment 128413
----
I had intended to fabricate the reservoir from quarter inch copper , but I could not find a 400 amp TIG machine.I ended up having to use the next best heat conductor , aluminum.
There is nothing cheap about it, just the parts in that picture are worth about $ 1200 US , not including my time.
The PSU shown is a meanwell RSP-1000-27 .It's controllable from about 13volts to 30.5 volts (the pelts are 31 volts, 17 amps max, each).

------

Zarathustra[H] ,

Good info on Propylene glycol !!! I did not know about the antimicrobial properties of it....very cool.
Thanks !!

No wonder I never see crap in my coolant at the brewery.I've got a 55 Gal drum of clear 100% propylene glycol on the way already for the brewery, so that's done.
-------

As far as the galvanic corrosion goes ;
Galvanic corrosion requires a current loop.....The aluminum reservoir part is already electrically isolated, so it will not be an issue.
It's trivial for me to electrically isolate my copper rad too, I'll do that and be done with it.
-------

I've just taken the whole reservoir assembly apart for final lapping and insulation, I'm putting it all back together this week.

Thanks again everyone !!
Cheers.


I'll probably have another question or two as I do final assembly

Honestly, I think this build is a very bad idea. Mixing aluminum and copper in the same loop, even with the best anticorrosive in the world is going to be a problem. I'm not sure how you are insulating them electrically, but the current loop occurs via the fluid. It will pick up ions over time from the metal contact surfaces in the loop and become conductive, and with metals as dissimilar as aluminum and copper this is a very very bad idea.

I'd argue that no matter what mitigation steps you take, there is absolutely no way to prevent having a very bad result when putting copper and aluminum in the same loop. This should quite simply never be done under any circumstances.

 

[cyberguyz]

Mixing copper and aluminum in the same loop can and will have a galvanic effect where free electrons of the aluminum migrate to deposit on the copper via the coolant that acts as an electrolyte.

Ethelene glycol (good, highly toxic), propylene glycol (best, non-toxic) are additives that provide lower electrolytic factors as well as surface tension reduction and biocide properties. Both chemicals can be found in most, if not all automotive and industrial hvac coolants. These chemical can lose their effectiveness when contacting oxygen over long periods of time. That last means you need to change your coolant occasionally if you any air at all in your loop (everybody does).