Anybody have a liquid coolant leak in their PC and did it go well after regarding the cleanup?

[Forman]

Were you on Xtremesystems?

Oh yeah. 3dmark, superpi,, Winbond BH5 torture, DICE, LN2 ... Overclocking and diy extreme cooling is how I got into physics and electronics.

 

[Chromatischism]

Oh yeah. 3dmark, superpi,, Winbond BH5 torture, DICE, LN2 ... Overclocking and diy extreme cooling is how I got into physics and electronics.

My time there slowed down after about 2007ish, but it was a great forum.

 

[Doodski]

My Corsair water cooler pump died! Corsair replaced it with a better and newer model under warranty. The PC was working and after about 3 years of operation I decided that I better check the thermals on my PC. So I loaded the CPU down with software and checked the temps and sure enough 40C at idle and 100C+ when loaded. It must have been just under the thermal cutoff temp because the PC never auto shut down. I still wish I had bought a Noctua air cooler though. The Corsair service was good.

 

[Grumpish]

As far as damage goes, it is going to depend on what the coolant is, and how quickly you get to it. I have saved flooded underwater cameras by keeping them soaking in freshwater until I could get them somewhere they could be dried out properly, and I have saved more than a few laptops after they have been given a cup of coffee (or in one case a strawberry flavoured protein shake) to drink, by getting the motherboards out and into clean water as soon as possible. The main thing is to get any power source disconnected as soon as possible, and to get whatever has got onto it off before it can do any damage - you would be amazed at how fast something as apparently innocuous as sugar can eat away at a circuit board.

 

[Doodski]

As far as damage goes, it is going to depend on what the coolant is, and how quickly you get to it. I have saved flooded underwater cameras by keeping them soaking in freshwater until I could get them somewhere they could be dried out properly, and I have saved more than a few laptops after they have been given a cup of coffee (or in one case a strawberry flavoured protein shake) to drink, by getting the motherboards out and into clean water as soon as possible. The main thing is to get any power source disconnected as soon as possible, and to get whatever has got onto it off before it can do any damage - you would be amazed at how fast something as apparently innocuous as sugar can eat away at a circuit board.

I've serviced a bunch of audio gear with water, coca cola, orange juice and whatever else they spilled so I've seen the damage from them. It's impressive that you saved the underwater cameras and notebooks. From what I've read the glycol mixture is not conductive but I would not rely on that because like you say disconnect from power ASAP.

 

[Raindog123]

Immersion cooling, Baby!

 

[Grumpish]

I've serviced a bunch of audio gear with water, coca cola, orange juice and whatever else they spilled so I've seen the damage from them. It's impressive that you saved the underwater cameras and notebooks. From what I've read the glycol mixture is not conductive but I would not rely on that because like you say disconnect from power ASAP.

Coca Cola is lethal to anything electronic, is is so acidic that I don't like to think what is doing to my stomach. The laptops that I was able to rescue were the ones that were bought to me straight away.

 

[Doodski]

View attachment 163540


Immersion cooling, Baby!

Is that distilled, water mineral oil etc?

 

[Raindog123]

Is that distilled, water mineral oil etc?

It’s a “secret sauce” (but mineral oil would do too).

 

[Doodski]

Coca Cola is lethal to anything electronic, is is so acidic that I don't like to think what is doing to my stomach. The laptops that I was able to rescue were the ones that were bought to me straight away.

If memory serves me correct I've replaced a CPU or two due to coke spills. The stuff must be conductive. Nasty stuff for sure. My favorite excuse was the planter ring on top from where a fern(example) was placed on top of a receiver and then over watering the plant caused a spill. I've seen that a couple of times too. Myself I have my notebook on the far left side of a 6 foot desk on a stand and the desktop PC is unreachable from the front of the desk on a riser too and I never put my beverage glass on the left side. I would be so embarrassed and angry with myself if I damage my own gear with a spill. It looks pretty idiotic if a electronic tech spills crap in gear.

 

[Berwhale]

It’s a “secret sauce” (but mineral oil would do too).

20 years ago, when I was into overclocking and building custom watercooling rigs, I had a friend who dabbled in full immersion cooling. I can't remember which dialectric fluid he was using, but I remember it was expensive - over £50 for 500ml. Anyway, full immersion cooling is a lot more common in DCs now... https://www.engineeredfluids.com/

 

[nerdstrike]

Pedant alert!

Firstly, basically all consumer CPUs after the Pentium 4 era have integral thermal throttling. They slow down rather than kill themselves. It's fairly common to see laptop chips bouncing off their limiters at 105 degrees, because the cooling is not intended for sustained peak power. In a desktop setting you can and should do much better!

Secondly, the phosphoric acid in cola is less acidic than your typical stomach content. The mucous lining is intended to allow high acidity to sterilise and hydrolise food. Then your gall neutralises the acidity and pushes it alkaline as it transits to the intestines. As such I think the stomach can handle some cola, unlike your dentine!

Lastly, the conductivity of deionised water is very low, but it doesn't take much material at all to turn it conductive. As such it's a poor choice for immersion cooling.

Do carry on!

 

[bloodshoteyed]

thought of going liquid cooled on my main rig quite a few times over the years
but with my CPU (R9 3900x) running below 65C with all the cores/threads maxed out on synth benchmarks and just about 71C with the GPU going full (6800xt) - on a stock AMD cooler - i just don't see the viability of it
mind that it's running 24/7 for both work and hobbies i don't OC it and introducing liquids is still a risk factor i never could lay aside

 

[Berwhale]

I've never had a leak in 20 years, that includes running custom rigs with coolers on the CPU, GPU and Northbridge and various AIOs. I think there's very little risk as long as you buy decent quality components and route pipes and cables sensibly and securely.

 

[jae]

Coincidentally, this happened to me personally very recently for the first time in nearly 20 years of water cooling due to laziness/tiredness. And twice in the same day. The proper procedure is something like this:

1. Immediately turn off power supply switch and unplug it.
2. Unplug motherboard cables (these can often carry residual power to the motherboard, especially with cables that have in-line caps for ripple control)
3. Remove CMOS battery
4. Soak up any residual puddles of liquid with paper towel
5. pour high concentration isopropyl alcohol on affected areas with board at an angle to displace the water. If the board was extensively covered, I would bathe it completely after removing the CPU/memory from the socket.
6. bake it in the oven at low temperature (around 40-50C) to evaporate, no need to wait days to see if your hardware is actually dead

Obligatory pic, I could have lost dual 3090s and an 8700k, not to mention 2x 2TB samsung nvme. The system was powered but luckily the fault protection kicked in when there was a short on on a non-critical area of the USB controller and after the cleanup, everything was in working order. I did get a bit of a numb tingly shock from the USB ports from the capacitors in the PSU cables, even after the PSU was off. When reassembling everything after a full cleanup (and total dismantling, drying, and cleanup of the GPU water blocks and boards), the EXACT same leak occurred and I had to redo everything all over again. Luckily it was powered off that time and I was running the pump solo. Fortunately, everything was also on a test bench so I was able to respond quickly.

By the way, the cause of the leak was a overly twisted/stretched hose causing a fitting to unscrew itself. Of course, my fault for not leak testing but I never leak test and this has never happened to me, I was just a bit impatient and this was only a temporary setup so I was not scrutinizing things well enough. Can't wait to go back to barbs+clamps, I really hate compression fittings but I did not have much choice as I am overseas and away from all my old watercooling stuff and also all my tools.

I'm waiting patiently to buy some new external radiators and perhaps the new 12900k when all the new motherboards have released. I really would like to wait for another generation or two before upgrading, but the 8700k bottlenecks my 3090s way too much to make full use of them.

 

[jae]

20 years ago, when I was into overclocking and building custom watercooling rigs, I had a friend who dabbled in full immersion cooling. I can't remember which dialectric fluid he was using, but I remember it was expensive - over £50 for 500ml. Anyway, full immersion cooling is a lot more common in DCs now... https://www.engineeredfluids.com/

Almost decided to go this route with electrocool to cool a number of GPUs that do simulation workloads and crypto mining all day. I'm still keen on doing it, but maybe when I go back to Canada where I can really benefit from having a heat exchanger outside in winter and cheaper shipping from the US. Shipping all that fluid to Oceania was also quite expensive which deterred me. I was thinking about doing a 15-20L test run with my own personal machine just a few weeks ago.

 

[maverickronin]

Pedant alert!

Firstly, basically all consumer CPUs after the Pentium 4 era have integral thermal throttling. They slow down rather than kill themselves. It's fairly common to see laptop chips bouncing off their limiters at 105 degrees, because the cooling is not intended for sustained peak power. In a desktop setting you can and should do much better!

Secondly, the phosphoric acid in cola is less acidic than your typical stomach content. The mucous lining is intended to allow high acidity to sterilise and hydrolise food. Then your gall neutralises the acidity and pushes it alkaline as it transits to the intestines. As such I think the stomach can handle some cola, unlike your dentine!

Lastly, the conductivity of deionised water is very low, but it doesn't take much material at all to turn it conductive. As such it's a poor choice for immersion cooling.

Do carry on!

Username check out.

 

[Tks]

Almost decided to go this route with electrocool to cool a number of GPUs that do simulation workloads and crypto mining all day. I'm still keen on doing it, but maybe when I go back to Canada where I can really benefit from having a heat exchanger outside in winter and cheaper shipping from the US. Shipping all that fluid to Oceania was also quite expensive which deterred me. I was thinking about doing a 15-20L test run with my own personal machine just a few weeks ago.

Idk about whatever this dialectric liquid is, but if it's anything like mineral water - don't bother, it's more trouble than it's worth, and the performance simply isn't there once the entire solution reaches steady state. Unless of course you plan on having a large tank big enough to never reach a high thermal soak quantity. But even then, there's the issue of circulation not being adequate. You also have maintenance issues (the liquid somehow naturally gets grimey eventually)

Also, kiss all those parts goodbye if second-hand sales are ever a consideration.

 

[JustAnandaDourEyedDude]

TL;DR : Water-cooling costs more, is noisier, does not lower the max CPU temp by much more than what the best air-coolers achieve, while introducing some small risk of catastrophic electronics failure due to liquid leakage. The diminishing returns are due to the limitation that the lowest CPU surface temperature that these convective cooling methods can produce is that of the ambient air temperature. This limitation could be overcome by using a refrigeration cycle device instead of liquid-cooling.

---

Yes, the best water-cooling is better than the best air-cooling, but not by much. The reason for the latter is simple to understand, based on a lumped steady-state heat-transfer model that shows the essential features. The model says that when temperatures have settled, the heat conduction rate within the solid CPU (or GPU) is equal to the heat transfer rate from the CPU to the air in the case of air-cooling, while in the case of water-cooling the heat conduction rate within the CPU is equal to the heat transfer rate from the CPU to the water, which in turn is equal to the heat transfer rate from the water to the air at the radiator, all of which have to be able to balance the heat load of Joule heating.

The model for air-cooling says that q1 = k/d*A1*(T1-T2) = h3*A3*(T2-T3)

while that for water-cooling says that q1 = k/d*A1*(T1-T2) = h4*A4*(T2-T4) = h5*A5*(T5-T3).

Here, T1 is the maximum (interior) temperature in the CPU or GPU or whatever solid we are trying to cool, while T2 is the temperature at the surface of the CPU where the air-cooler or water-cooler makes contact with the CPU. Since metals are much better conductors than the other substances here, the temperature drop across the metal blocks are relatively small, and so we can with small error take the entire air-cooler block or water-cooler block to also be T2. T3 is the temperature of the ambient air to which we are rejecting heat, and T4 is the temperature of the water at the CPU while T5 is the temperature of the water at the radiator; we can take T5 to be equal to T4 with negligible error. Edit: T5 and T4 are not actually equal, as their difference accounts for the heat carried away in the water: q1 = mdot*Cp*(T5-T4), but with large water flow rate mdot and large heat transfer contact areas, the difference T5-T4 will be small relative to the difference T1-T2, and will become smaller as the cooling is made more extreme by increasing the water flow rate and heat transfer areas.

q1 is the heat load generated by electrical operations, and depends on the overclock or not and the stress test such as Prime95, etc. To understand the cooling system trends, we think of q1 as fixed at the desired operation level. Similarly, while the ambient temperature T3 depends on the weather and whether the room is air-conditioned or has sun-facing windows, we think of it as fixed for purposes of cooling system analysis. Similarly k/d*A1 is an as-manufactured characteristic of heat conduction properties of the particular model CPU or GPU we examine (the user selects which model CPU/GPU they buy), and we consider it as constant for our analysis. h3 is the convective heat transfer coefficient between the air-cooler block and the surrounding air (radiative heat transfer does not play a large role at these temperatures), while A3 is the area of that interface. Similarly, h4 and A4 are the corresponding quantities at the CPU water-block to water interface, and h5 and A5 are the corresponding quantities at the water-radiator to air interface. The values of the convective heat transfer coefficient h3, h4, h5 depend on the velocities of the fluids and their thermophysical properties.

The goal of cooling is to reduce T1, the max temp experienced by the CPU. The first equality in each of the two models tells us that T1-T2 is fixed, and so in order to lower T1, we have to lower T2. This is done with the cooling system, but if T2 is too high due to poor cooling, then T1 will be too high and thermal throttling of the CPU will kick in. The air-cooling system we pick determines h3 and A3, while the water-cooling system we pick determines h4 and A4 (as well as h5 and A5, but these latter are not as important except that to beef up h5 and A5 sufficiently, the radiators are big and their fans noisy). We can raise h3 and h4 by increasing the air or water speed, and we can increase the surface areas A3 and A4 used for convection; raising h*A lowers the temperature difference dT required to drive the heat transfer. All of that comes at a cost (noise, power, size, $$$).

The key observations are (1) that to dissipate a given load q1 using standard air- or water-cooling, the best you can do when beefing up h3, A3, h4, A4 is to lower T2 to the ambient air temperature T3, which in turn sets a limit on the most we can lower T1; and (2) the better air-coolers already do a good job lowering T2 to be fairly close to the ambient T3 (relative to their distance from T1). The better air-coolers already take us far into the region of diminishing returns, and water-cooling can bump up h*A by an order of magnitude at greater cost, but gain us only single-digit temperature improvements at most. And going more extreme with convective cooling would cost far more while bringing much smaller further lowering of T1 as we approach the limit q1 = k/d*A1*(T1-T3) of the best possible air- or water-cooling. Even raising h*A to infinity (at infinite cost) would only lower T2 to equal T3 exactly, with T2=T4=T5-T3. The bottleneck with the best air-cooling is already the heat conduction limitations within the CPU and the fixed ambient temperature to which the heat is rejected, and water-cooling does not change that (and nor does direct contact convective heat transfer by immersion cooling in any other fluid, unless you are able to keep that coolant at a lower temperature than that of the ambient air). Heat flows only from higher to lower temperatures.

While water-cooling may not be able to improve much on lowering T1, as compared with the best air-cooling, and it brings a risk of disaster due to leakage onto electronics, it does bring a benefit. Water has a headstart on air due to h4 being much larger than h3, owing in large part to the much larger thermal capacity of water. This allows the water-block to be much smaller while achieving the same cooling capacity as an air-block. The use of water effectively moves the transfer to the air away from the CPU to where larger radiators are not a problem. So while liquid-cooling does not bring much benefit for consumer PCs (except for overclockers trying to eke out another few MHz), it is important for rack-mount servers and particularly supercomputer applications, where liquid-cooling enables high compute-density, and the liquid in turn is cooled externally by refrigeration.

The way to overcome the limitation regarding the lowest value of T2 attainable by air- and liquid-cooling being T3, is to introduce a refrigeration cycle instead of the standard liquid-cooling. This is a reversed heat engine and involves a gas or a phase-change fluid as the working fluid, as well as a compressor, and works essentially like a refrigerator or air-conditioner. These astounding technological advances (of nearly a couple of centuries ago) are based on the divergence of constant-property lines on the p-v thermodynamic property diagram regions of gases and phase-change fluids. So instead of T5=T4, refrigeration enables T4 (and thus T2) to be much smaller than T3, while T5 can be much larger than T3; of course this all comes at a cost so it is unclear whether it is commercially viable in the consumer PC space. Thus T1 can be lowered dramatically below what air-cooling and liquid-cooling can achieve. Lowering T3 hugely by replacing the ambient air with liquid nitrogen is also an alternative. Thermo-electric cooling is another alternative, though I recall reading that it cannot handle large heat loads q1, due to the Peltier effect being a weak one. The Peltier and Seebeck and similar effects are usually weak, being secondary transport phenomena that arise as cross-modal transport, with the transport coefficients subject in the linear regime to the Onsager reciprocity relations.

 

[jae]

Idk about whatever this dialectric liquid is, but if it's anything like mineral water - don't bother, it's more trouble than it's worth, and the performance simply isn't there once the entire solution reaches steady state. Unless of course you plan on having a large tank big enough to never reach a high thermal soak quantity. But even then, there's the issue of circulation not being adequate. You also have maintenance issues (the liquid somehow naturally gets grimey eventually)

Also, kiss all those parts goodbye if second-hand sales are ever a consideration.

These newer specialty dielectric fluids are quite a leap from the typical mineral oil immersion in terms of longevity and other characteristics. They can easily last 5-10 years or longer and come with warranty for that long if desired. Performance wise, it may not be that much greater than a typical high performance water/glycol cooling loop but it offers lots of distinct benefits on top of marginally better thermals, namely cost savings. For example, I plan on always buying at least 1-2 of the latest/greatest GPU available every generation, and spending money on water blocks is not necessary with an immersion setup. Buying waterblocks and a few sets of good thermal pads over the life of the fluid will easily cost $2000+ for a couple pairs of gpus over that time period. There may be unforseen CPU socket changes which now needs a new CPU water block which further adds more to the cost. The fittings and hosing/piping for all those blocks also have a cost, and blocks should require semi-regular maintenance which also takes time and effort, coolant swaps etc.

Another huge advantage is that immersing in this fluid allows for subambient cooling with zero risk of condensation. Subambient cooling is already quite inefficient and can rarely be done unattended or without elaborate insulated and controlled setups, but this makes it effortless and viable for continuous 24/7 operation. So if I opt for this I further increase performance and component longevity.

A large tank isn't really necessary- it only needs to be large enough to submerge the motherboard and gpu(s) so for my specific setup with traditionally horizontally mounted gpus is only 15L (10" x 12" x 8") which is smaller than most small form factor cases. If one only has one GPU and mounts it vertically, it can be done with nearly half that volume- 7.5L. You can even fill in the void space with things like pieces of acrylic to use even less volume of fluid. 20L of this fluid ranges from about $300-600 depending on quality/warranty option, typically about $400-500. Other than the one off cost of building a small custom tank/case, this is already significantly cheaper than buying a number of water blocks.

Also regarding cleanup- it is an extra step but it is trivial all things considered. I used to be into LN2/extreme overclocking so I am no stranger to motherboards coated in vaseline and craft putty. Dunking a motherboard or gpu in an ultrasonic cleaner with some isopropyl alcohol and leaving it for an hour or so basically has it coming out newer than it looked coming out of the retail box. Mineral oil is simply too volatile and takes up ions too easily/poor materials compatibility which is what causes so many problems for people. These diaelectric fluids also have a specialised solvent for them which costs $100 (and can be reused multiple times) that you can wash all the PCBs with and leave to air/blow dry and there is no way of knowing it ever touched this fluid. Second hand sales are not a problem at all, and the components basically look like they have never been used if you do the process properly.

Honestly, the biggest issue with this is that component swaps are a bit more involved to do them properly because of the cleanup. This also scales extremely well. If I wanted to add more gpus lets say, I would at most have to add maybe another 0.5-1 L of fluid each, and in many cases I would have to use even less fluid per component as the component itself displaces volume, depending on the configuration of the vessel.