Discussion on mixing metals, other than "don't do it"

Subterranean · 11 Feb 2019 at 16:42

Hi all

I'm most of the way through assembling, cleaning and testing my first custom water cooling loop. Everything's looking good so far and before anyone asks, it is all name brand copper/brass with mayhems fluid to go in it.

At the moment I'm using a Corsair H100i AIO for my CPU that I've had for nearly five years and never had any problem with. I was surprised to find out that the vast majority (all?) of the major brand AIO cpu coolers like mine (other than the much more expensive ones from specific watercooling brands) use an aluminium radiator with a copper cpu plate. Obviously every newb guide on watercooling screams "don't mix metals" with the occasional picture of the inside of a heavily corroded copper cpu plate, but no real context for the picture. For all we know it could have been run on saltwater.

For anyone that doesn't know galvanic corrosion, think of it like the two ends of a battery. If there is nothing conducting between them, the battery will stay charged. If you put something conductive between them, like a wire or normal water it will conduct and the battery will run out. This electric current is what causes the corrosion. In a water cooling system, the copper is one end and the aluminium is the other.

I have a physics and engineering background, and have seen how small a volume of sacrificial anode is needed to keep an entire ship from falling to pieces. I am also a caver/climber and have seen how much and how quickly corrosion happens to a high purity stainless steel bolt in rock when constantly splashed with seawater and due to run off from vegetation above.

This means the lifespan of all these AIO controllers is controlled by what sort of fluid is put into the system. Clearly big names like Corsair have RnD departments that can ensure a correct mixture is chosen that will outlast it's warranty and ideally the full 6 years it is expected to perform for (in the UK at least).

Has anyone ever taken apart an old 5-6 year old AIO and actually seen any problems inside from the alu/copper mix? I realise their pumps are weaker and aluminium rads are not as good as copper/brass, so there is a good reason to upgrade to a D5 pump with copper rads, but if I was to take mine apart, does anyone really think it would be that bad inside, remembering its 5 years old?

I have changed cpu since buying it so I cannot give any results on its continued performance to be sure that it has not got worse over the years, but my current CPU is a 9700k bought in December which I can run at 5.1GHz with 24hr prime95 stability etc etc etc. The fans do get a touch loud at those levels though.

Glanza · 11 Feb 2019 at 16:50

MikeTimbers · 11 Feb 2019 at 16:58

Ethylene glycol so no galvanic corrosion.

Subterranean · 11 Feb 2019 at 17:10

Glanza said:

I did see that video. My problem is he said he has had that cooler for four years, but didn't mention whether it had been used for four years. Maybe other videos of his show it in use if anyone wants to correct me. He has a lot of parts that sit around on shelves in the background for a long time but warming and cycling the fluids will be different.

He found no problems from mixing the metals other than the liquid was "greenish" (i'm surprised he didn't pour it out into a glass container to look for actual contaminants) and that it smelled bad, both of which could have been completely normal from when it was filled.

Subterranean · 11 Feb 2019 at 17:21

MikeTimbers said:
Ethylene glycol so no galvanic corrosion.

So effectively anti-freeze/refrigerant. Is it much more than expensive than name brand fluid? Is it more viscous and hard to move, or else what are the disadvantages?

MikeTimbers · 11 Feb 2019 at 18:03

There are quite a number of videos on YouTube of people draining the fluid out of AIO and it generally looks like water - i.e. not viscous. A 120mm AIO only holds around 100ml of liquid.

The point is that they are using liquids designed to prevent corrosion and to inhibit any possible bio presence. if you are using a premix that includes inhibitor and biocide you have nothing to worry about (although premixes bring other problems like particles gunking up blocks and pumps, YMMV)

Holst1981 · 11 Feb 2019 at 18:05

Subterranean said:
So effectively anti-freeze/refrigerant. Is it much more than expensive than name brand fluid? Is it more viscous and hard to move, or else what are the disadvantages?

Its specific heat capacity is less than water.

Its poisonous.

If you do get a leak it's messy.

Probably cheaper than branded fluid.

Subterranean · 11 Feb 2019 at 19:08

MikeTimbers said:
There are quite a number of videos on YouTube of people draining the fluid out of AIO and it generally looks like water - i.e. not viscous. A 120mm AIO only holds around 100ml of liquid.

The point is that they are using liquids designed to prevent corrosion and to inhibit any possible bio presence. if you are using a premix that includes inhibitor and biocide you have nothing to worry about (although premixes bring other problems like particles gunking up blocks and pumps, YMMV)

Thanks for your reply, that's what I was looking for. I guess that's why they all use black opaque tubing, since the liquid looks like ****.

Would everyone who has used liquid cooling for lets say the past 6 years agree that the right liquid additive more or less takes away any worry of internal corrosion, while obviously newb guides say don't mix metals, just in case someone pours tap water into their loop?

I have bought some XT-1 NUKE concentrate from OCUK to use because it was cheap and looked very specific application oriented rather than others that go for looks but need more regular maintenance.

MeddlE · 11 Feb 2019 at 21:41

Having recently gone from a CLC AIO (Enermax LiqTech II 120s) that had stopped working to AlphaCool loop with copper rad I decided to take the CLC apart to see what makes it tick and see if I could get it working.

The block on the CLC was completely gunked up with very small black particles.
I drained the CLC and got about 30ml out, it took about 5 times that amount to refill it. What came out looked like very dark urine with lots of particles in.
The alleged 30mm radiator was actually 15mm thick with 30mm fake end plates.

I flushed it, stripped the pump and cleaned it, put it back together and refilled it and it works.

While researching the replacement I quickly came across the "don't mix metals" suggestion so narrowed my choices down to full copper rads only.
What I also found was the information that your pump will not be made of copper. So you will always be running mixed metals.

What I ended up with is a dual bay res that takes the D5 vario, copper CPU block, 60mm thick 120mm copper rad, clear flexi tubing and clear premixed coolant.

Putting it together was a fairly logical process that was also fun. Who doesn't like playing with new gadgets?

I believe the idea is that you replace the coolant once a year and give the system a flush at the same time.

Kei · 12 Feb 2019 at 11:41

AIO's are far more likely to use propylene glycol as it is non toxic. Neither propylene glycol or ethylene glycol by themselves are a corrosion inhibitor. There are additives that serve that purpose.

Subterranean said:
Thanks for your reply, that's what I was looking for. I guess that's why they all use black opaque tubing, since the liquid looks like ****.

Would everyone who has used liquid cooling for lets say the past 6 years agree that the right liquid additive more or less takes away any worry of internal corrosion, while obviously newb guides say don't mix metals, just in case someone pours tap water into their loop?

I have bought some XT-1 NUKE concentrate from OCUK to use because it was cheap and looked very specific application oriented rather than others that go for looks but need more regular maintenance.

I've run diluted traditional car antifreeze concentrate in my system now for 6 years without any gunking up or corrosion. It's designed to prevent corrosion between, aluminium, brass, copper, steel and iron which are the common components in a car cooling system. They come in a variety of colours and in many cases, they are UV reactive too. I simply use whatever I have on hand that goes in the car as it saves me money buying loads of different coolants.

GM Dexcool, virtually clear when diluted. Fluoresces lime green under blue LEDs and UV. No issues with corrosion or gunk over 4 years use. Does discolour tubing over time, it eventually goes a yellowy pink.

BMW nitrite free. Dark teal in colour, no sign of fluorescence under blue LED's, not tested with UV. No signs of corrosion or gunking up in 2 years use. Discolouration of the tubing is not obvious so far. Not compatible with dexcool, mixing causes silicate displacement and will cause white powdery gunk to form.

MikeTimbers · 13 Feb 2019 at 10:26

You can do all you can to avoid mixed metals and still have problems:

This was a "fully anodised" block - I forget the manufacturer.

This was the infamous Swiftech Apogee GTX - again bad anodising. I had a bad run of luck with blocks in the mid noughties.