Possible cavitation?

[jumper2high]

I'm wondering about one thing - after my loop is fully bled, the bubbles from most of the loop are totally gone, but i see a small string-like stream of fast moving tiny bubbles on the GPU block outlet (EK FC460GTX).

There are no bubbles entering the GPU block, and it's been running for several days now so it's unlikely the block didn't bleed fully yet.


Could this be cavitation from water at higher temperatures, exiting the block (from thin channels inside the block to a big tube 3/4" tube) with a pressure drop?

[theseeker]

I doubt it, but what is your loop layout?

[DarthBeavis]

Do you see any bubbles when the system is turned off?

[SNiiPE_DoGG]

nope

your block just still has small amounts of air in it or there are air bubbles going in that you just dont see.

[penguins]

+1, sometimes there is just that one spot in the block bubbles love to hang out for a while.

[jumper2high]

Additional info:

Loop details:

Tube size - 1/2" - 3/4"
Fittings - Bitspower Compression fittings
Pump: DCP 4.0
Reservoir: Iandh 275
CPU Block: Supreme HF
GPU Block: EK FC460GTX
Radiator: Swiftech MCR320

Loop order (without fittings)
PUMP - CPU - GPU - RAD - RES - PUMP

Loop order (with fittings - [direct] means a regular compression fitting, everything else is an adapter with a compression fitting in the end)

PUMP(direct) - (dual rotary)CPU(direct) - (barb)GPU(dual rotary) - (direct)RAD(direct) - (45deg rotary)RES(45deg rotary) - (direct) PUMP

The bubbles coming out form a small string-like stream that is UNINTERRUPTED for as long as the system is running under load. It is more noticeable when the water temp is higher (like when I run AVX LinX or something that loads both CPU & GPU). It's actually more of a mist-like stream (1mm in diameter) throughout the entire length of that tube (once it enters the radiator, I can't see it anymore).


When the system is turned off, obviously there are no bubbles - but moments immediately after the flow stops, it looks like someone stopped stirring a glass full of water (the tiny bubbles slowly stop twisting around each other and disappear). I wish there was a way to film it but it's too hard to notice on tape.

The system exhibits this behavior after months of running.

[eth0s]

I'm wondering about one thing - after my loop is fully bled, the bubbles from most of the loop are totally gone, but i see a small string-like stream of fast moving tiny bubbles on the GPU block outlet (EK FC460GTX).

There are no bubbles entering the GPU block, and it's been running for several days now so it's unlikely the block didn't bleed fully yet.


Could this be cavitation from water at higher temperatures, exiting the block (from thin channels inside the block to a big tube 3/4" tube) with a pressure drop?

If you are asking if it is possible that you are getting "cavitation" inside your GPU block, I would say no. Cavitation is a pump related phenomenon, and will not occur inside your GPU block. You simply have trapped air bubbles inside your GPU block. The only other explanation is that your block is so hot that it is boiling the water inside, which also defies credulity, especially if the water is moving.

[penguins]

actually the only other explanation is a small leak that isn't dropping water, but rather sucking air in.

as boiling water with bubbles isn't going to happen before tubes pop off.

but i'm going to go with what i think and everyone else thinks it's just air trapped brother. calm down. how long has it been since you started the loop up? a few days? week? have you done the strong man routine? tip shake shake tip turn shake shake bob shake weave, shake dodge.

[jumper2high]

Maybe cavitation isn't the best name for what's going on, but I doubt it's just residual bubbles.

I assembled the loop in January and I saw it all the time since then. I recently re-assembled the loop (2-3 days ago) and once again I can see it once the water warms up (to approx 33*C). I'm thinking a combination of warm water and a pressure drop (lowering the water boiling point) makes a very small amount of it evaporate. Strange thing is it only starts showing when the water gets a little warmer than idle.

The 'sucking air in' theory is interesting, but unlikely as my PC spends about 16 hours running and 8 hours off, I guess I would have seen a drop or two during the off-time.

[theseeker]

Every EK GPU block that I have owned has this issue. Maybe a reason for the new design in the GPU blocks.
The reason I asked about your loop configuration? If you have a highly reduced flow into your pump inlet, the pump would be spinning faster than the water could make into the inlet. I hope that makes sense, but in your case, it is not possible.

[penguins]

not necessarily, the tubing would be fine under non pump pressure, then with the warming of tubing, + pressure of pumps ( you account that for an impossible option, why not this one? ) pulling air in, but the vacuum isn't enough to drop water out.

I build a watercooling system 4 years ago, it still had air bubbles in it 3 years later. Time doesn't matter, only life matter.. wait wrong movie.

Time doesn't matter, only Shaking the case to get everything out and not stopping till you're done. Got Acyclic tops on your blocks?


Edit: ever heard of a radiator for cars? the cooling system is under pressure, usually ~10-13 psi it RAISES the boiling point. a Vacuum Decreases it
you would need ~28 Vacuum Inches of Hg(Mercury) ( VERY ROUGH ESTIMATE ) to bring the boing point down to a normal Loop temperature. aka Impossible. you're raising the boiling point with pressure.

[jumper2high]

Got Acyclic tops on your blocks?

Acetal....otherwise I would have looked in

[Martinm210]

Yeah, not cavitation. Cavitation is actually the formation and collapse of air bubbles in the pump itself cause by negative pressure to the point where water will boil at room temperature. This can only happen on the inlet side of the impeller where pressures are negative. The bubbles collapse back to water by the time they leave the impeller.

Added air could occur from a tiny leak at the pump inlet side though where the negative pressure is sucking air rather than leaking water. Pretty rare though..

[meanmoe]

Yeah, not cavitation.

+1, it's not going to be cavitation. There's air still in your loop.

Also, s/negative pressure/low static pressure/e

[gmat]

My current loop took a few weeks to fully get rid of last pockets of air... Depending on how your reservoir / air trap is set up you might be recirculating bubbles a bit, nothing to worry about (unless it lasts for months)

actually the only other explanation is a small leak that isn't dropping water, but rather sucking air in.

That is the next best possibility.

[jumper2high]

I'll see what happens...

[Conumdrum]

OP? This your problem?

http://www.overclockers.com/forums/s...3&postcount=26
http://www.overclockers.com/forums/s...0&postcount=30

Full thread.......
http://www.overclockers.com/forums/s...d.php?t=643159

[Rogue Warrior]

I say tip that thing around, BUT not upside down