Custom loop exploded

[wonderboy8]

So. This is wowza crazy. So crazy that I am using the word wowza. My loop. Literally exploded. A pic of my loop (before it exploded) is attached. The pipe that exploded is the bottom one from the gpu to the pump. It was the first long load that I put on the system. I was playing fallen order for about an hour with my headphones on so I did not hear an preemptive noises. But I did hear the explosion. The weird thing is that my temps were fantastic. I had Hwin open and was looking up every once and a while. The last temp I saw was mid 60's for Celsius. Now there is no where in my loop for air to escape. I am using bykski water cooling parts and ek cryofuel coolant. Does this coolant seriously increase pressure like this? 

 

Cryfuel is supposed to be non conductive so technically everything should be fine. 

 

As I was pulling everything apart, I did notice that all of the water parts were decently warm. Mostly the metal fittings. 

 

The only thing I can think is that the coolant built up pressure and that did it. What I don't get is how I was getting wonderful temps if the coolant was overheating. Is it always necessary to have an air outlet fitting?

 

Don't know if everything is working yet. 

 

And yes. When I say explode. I mean

Gonna let everything dry off.

All that really got wet was

bottom half of Mobo 

The acrylic Side of gpu

WiFi card 

And of course power supply is kinda soaked...

 

[Guest]

No kidding! That's crazy. Could that be just from the liquid expansion from heating up??!!

[asquirrel]

Steam is 1600x the volume of water. "But it was 60C!" Doesn't matter. Physics is weird and you'll still get evaporation below 100C. Jayz2cents did a test where he over-pressurized a water cooling loop until it failed. I forget how many PSI it took, but it was a considerable amount. Looking at that bend, probably what gave out was the O-ring seal on the compression fitting (that looks like a weird angle, may just be the picture). If you're lucky that's all that broke and the radiator isn't bent internally.

 

Leave a gap at the top of the res for air to form. Steam may be 1600x the volume of water, but unlike water, steam can be compressed. Even 1/8" is enough to keep you from exploding. More is better though. I'd still consider 1/8" dangerous.

[Eigenvektor]

14 minutes ago, ShrimpBrime said:

No kidding! That's crazy. Could that be just from the liquid expansion from heating up??!!

Absolutely. Water will expand as it heats up and with nowhere to go, it'll rupture the weakest point. Unlike air, water doesn't compress easily or much at all. So it will make room for itself as it expands In other words, do leave some room in there (e.g. air) so the water has some room to expand into without causing fatal damage.

[Guest]

21 minutes ago, Eigenvektor said:

Absolutely. Water will expand as it heats up and with nowhere to go, it'll rupture the weakest point. Unlike air, water doesn't compress easily or much at all. So it will make room for itself as it expands In other words, do leave some room in there (e.g. air) so the water has some room to expand into without causing fatal damage.

I know my car has a 15lb pressure release cap...... might not be a bad idea to vent a loop!!

[For Science!]

6 hours ago, wonderboy8 said:

 

 

 

 

You say your temps were "wonderful" but do not specify beyond saying mid 60's. What component is this, or is this fluid temperature? What really counts for liquid cooling is fluid temperature and this should be kept below 60 at all times as a bear minimum, beyond this pumps start having issues and more importantly coolant stability starts getting affect in very negative ways.

 

I can't quite tell if you are using hard tubing, or soft tubing, kinda looks like hardline to me. My gut guess would be that your tube just wasn't sitting in the fitting correctly and would have popped out anyway. I would guess this is down to user error and you should simply "try again" assuming there are no alumnium based parts as it looks like you are using opaque fluids (which I also generally wouldn't recommend anyway).

 

Not leaving air gaps is actually my preferred way of setting up custom loops and the 6 systems I have put together so far have never had anything close to an explosion

[jeppouza]

It looks like you didn't leave enough space for air in your WC system, your reservoir seems to be fully to the top, that's no good.

 

I had a similar issue but in my case it was the radiators' Delta fans that turned off due to a power failure (low juice/amperage) and then the WC liquid boiled till it exploded the reservoir, a tragedy...that time I've lost a X99S Gaming 9 AC MSI motherboard and a 5960x....after that I use only Air Coolers on my setups 

 

Today even a 3970x@PBO can be handled with an aircooler, IMHO the 10º to 15ºC in full load difference with custom WC does not justify the exotic cooling, the only justifiable is a 3990x@PBO CPU, with this one things get a little toasty and exotic cooling is required.

[Deli]

8 hours ago, asquirrel said:

Steam is 1600x the volume of water. "But it was 60C!" Doesn't matter. Physics is weird and you'll still get evaporation below 100C. Jayz2cents did a test where he over-pressurized a water cooling loop until it failed. I forget how many PSI it took, but it was a considerable amount. Looking at that bend, probably what gave out was the O-ring seal on the compression fitting (that looks like a weird angle, may just be the picture). If you're lucky that's all that broke and the radiator isn't bent internally.

 

Leave a gap at the top of the res for air to form. Steam may be 1600x the volume of water, but unlike water, steam can be compressed. Even 1/8" is enough to keep you from exploding. More is better though. I'd still consider 1/8" dangerous.

Ha, maybe he lives very high up in the mountain. Boiling point will drop to ~80C at 6000m.

[SolarNova]

Is that soft tubing with compression fittings ?. Assuming ur sufficiently tightening the fittings, they are near impossible to remove especially if u use tubing with an outer diameter that is right on the limit to what can fit the fittings. Or hard tubing ? ..Soft tubing is incredibly difficult to 'pop off' ..but hard tubing can easily be mismatched due to manufacturing tolerance between manufacturers. In that instance its conceivable that the relatively mild pressure buildup caused by the heating fluid could cause it to pop off.

Warming liquid does increase pressure after all, thats why u usually dont fill a res to the point of overflow.

[wonderboy8]

2 hours ago, Deli said:

Ha, maybe he lives very high up in the mountain. Boiling point will drop to ~80C at 6000m.

Yes for sure that's it

[wonderboy8]

1 hour ago, SolarNova said:

Is that soft tubing with compression fittings ?. Assuming ur sufficiently tightening the fittings, they are near impossible to remove especially if u use tubing with an outer diameter that is right on the limit to what can fit the fittings. Or hard tubing ? ..Soft tubing is incredibly difficult to 'pop off' ..but hard tubing can easily be mismatched due to manufacturing tolerance between manufacturers. In that instance its conceivable that the relatively mild pressure buildup caused by the heating fluid could cause it to pop off.

Warming liquid does increase pressure after all, thats why u usually dont fill a res to the point of overflow.

Hardline tubing. 

[wonderboy8]

4 hours ago, For Science! said:

You say your temps were "wonderful" but do not specify beyond saying mid 60's. What component is this, or is this fluid temperature? What really counts for liquid cooling is fluid temperature and this should be kept below 60 at all times as a bear minimum, beyond this pumps start having issues and more importantly coolant stability starts getting affect in very negative ways.

 

I can't quite tell if you are using hard tubing, or soft tubing, kinda looks like hardline to me. My gut guess would be that your tube just wasn't sitting in the fitting correctly and would have popped out anyway. I would guess this is down to user error and you should simply "try again" assuming there are no alumnium based parts as it looks like you are using opaque fluids (which I also generally wouldn't recommend anyway).

 

Not leaving air gaps is actually my preferred way of setting up custom loops and the 6 systems I have put together so far have never had anything close to an explosion

Gpu and CPU both were mid 60s 

[wonderboy8]

10 hours ago, Eigenvektor said:

Absolutely. Water will expand as it heats up and with nowhere to go, it'll rupture the weakest point. Unlike air, water doesn't compress easily or much at all. So it will make room for itself as it expands In other words, do leave some room in there (e.g. air) so the water has some room to expand into without causing fatal damage.

Should I just untwist one of the caps on my reservoir to let air escape then? 

[wonderboy8]

4 hours ago, For Science! said:

You say your temps were "wonderful" but do not specify beyond saying mid 60's. What component is this, or is this fluid temperature? What really counts for liquid cooling is fluid temperature and this should be kept below 60 at all times as a bear minimum, beyond this pumps start having issues and more importantly coolant stability starts getting affect in very negative ways.

 

I can't quite tell if you are using hard tubing, or soft tubing, kinda looks like hardline to me. My gut guess would be that your tube just wasn't sitting in the fitting correctly and would have popped out anyway. I would guess this is down to user error and you should simply "try again" assuming there are no alumnium based parts as it looks like you are using opaque fluids (which I also generally wouldn't recommend anyway).

 

Not leaving air gaps is actually my preferred way of setting up custom loops and the 6 systems I have put together so far have never had anything close to an explosion

Also doesn't everything equalize? So by the time the system stops rising in temps that means everything is about the same temperature. Its not like my fluid was any higher than mid 60s if it was then my components would have to be higher than 60. 

[wonderboy8]

11 hours ago, asquirrel said:

Steam is 1600x the volume of water. "But it was 60C!" Doesn't matter. Physics is weird and you'll still get evaporation below 100C. Jayz2cents did a test where he over-pressurized a water cooling loop until it failed. I forget how many PSI it took, but it was a considerable amount. Looking at that bend, probably what gave out was the O-ring seal on the compression fitting (that looks like a weird angle, may just be the picture). If you're lucky that's all that broke and the radiator isn't bent internally.

 

Leave a gap at the top of the res for air to form. Steam may be 1600x the volume of water, but unlike water, steam can be compressed. Even 1/8" is enough to keep you from exploding. More is better though. I'd still consider 1/8" dangerous.

I think I might just leave a cap off for the system to breath next time... Haha this was not fun

[wonderboy8]

10 hours ago, ShrimpBrime said:

I know my car has a 15lb pressure release cap...... might not be a bad idea to vent a loop!!

I think I will take one of my reservoirs caps of when the system is under decent load

[asquirrel]

7 hours ago, wonderboy8 said:

I think I might just leave a cap off for the system to breath next time... Haha this was not fun

That's not a good idea. If you do this, have a long tube (like, one foot long, vertical) to make 100% sure that any 'burps' (gas pockets getting into the res) can't shoot water all over creation. Nuclear Reactors (PWRs anyway) have a thing called a pressurizer that does the same job as the 1/8" air gap I talked about leaving.

 

One other thing for cleaning: Your best bet for cleaning is going to be ultra-sonically cleaning the parts in alcohol (isopropyl). Core concept:

 (Note: I did not actually watch this video, but if Louis doesn't at least mention *why* you ultrasonically clean parts in alcohol I'd be amazed). Failing access to an ultrasonic cleaner, I don't know what the next best thing would be. The reason being: the ultrasonic will remove the corrosion that has already started and help rebuild the oxide layer that prevents corrosion from happening.

 

It's not the end of the world if you don't do this. But if, 2-3 years from now, your PC randomly turns off, shit just dies, etc, it may be corrosion from this water spill that finally did enough damage to prevent normal operation.

[wonderboy8]

UPDATE:

dried everything and cleaned with a electronic quick drying spray to help evaporate whatever coolant was left. 

Keep in mind I wasn't using water. I was using non conductive coolant. 

The health of everything is perfect. No problems at all with any parts. No disrespect to anyone who disagrees with this but I will be going with a venting fitting that will release only air. These are the fittings I'm getting. 

 

So for those of you who are scared or are marred from a previous experience like this. There's no reason to be. While you should only custom loop if you're an enthusiast who loves it for what it is, (because air cooling accomplishes the same temps now) you shouldn't be scared about accidents completely destroying your PC. This is my takeaway from all of this. While I do hope that this is the last time I will have to panick, I would be ok and at peace with a small leak happening because technology of coolant has come a long way. 

[SolarNova]

Chances are since ur running a hardline loop, ur tubing run simply wasnt a snug fit.

[AnonymousGuy]

There's no chance your loop got so hot that it built up any meaningful pressure.  My portable LAN box is absurdly hot because of space constraints and the water only gets to about 45C, and it's not blowing tubing off barbs.  Half the barbs don't have any clamps on them.

 

By comparison in my desktop the water never goes above 30C.

 

The only time I've blown tubing off: when I put 3x D5 pumps in series it made so much pressure that it blew off tubing even with a hose clamp on it.  Also when I had 155F (dunno..65C?) water [not in a desktop...] where it bulged out a radiator.