Leak testing

[Beerzerker]

This may not actually apply to the OP's situation based on the exact system setup but with mine I have a spare PSU that powers the waterpump itself and nothing else.

All I have to do is make sure the system's PSU is plugged to the board to provide a ground (Doesn't have to be switched on by the powerswitch on it) and that the green wire in the ATX connector has ground (NOT power - The green wire is always ground) to one of the mounting holes of the board itself.

I use a simple piece of wire to make that connection for the pump's PSU and turn it on or off with it's switch.

 

You don't have to use a mounting hole if you don't want, any ground for this wire to the green wire in the ATX connector will do.

I then just switch it on and let it run, watching for leaks without having to switch on the system's PSU. 

 

Need to point out my pump (MCP-655 Vario) has the older molex 4-pin plug for it's power.

Many pumps like an AIO or newer pump models in general have the smaller fanheader-style power plug so it may not be doable in every situation. In my case I can start the cooling system, check for leaks and so on without having power to the system itself so if it does have a leak, chances are it won't hurt anything.

[ChrisZH]

3 hours ago, Beerzerker said:

This may not actually apply to the OP's situation based on the exact system setup but with mine I have a spare PSU that powers the waterpump itself and nothing else.

All I have to do is make sure the system's PSU is plugged to the board to provide a ground (Doesn't have to be switched on by the powerswitch on it) and that the green wire in the ATX connector has ground (NOT power - The green wire is always ground) to one of the mounting holes of the board itself.

I use a simple piece of wire to make that connection for the pump's PSU and turn it on or off with it's switch.

 

You don't have to use a mounting hole if you don't want, any ground for this wire to the green wire in the ATX connector will do.

I then just switch it on and let it run, watching for leaks without having to switch on the system's PSU. 

 

Need to point out my pump (MCP-655 Vario) has the older molex 4-pin plug for it's power.

Many pumps like an AIO or newer pump models in general have the smaller fanheader-style power plug so it may not be doable in every situation. In my case I can start the cooling system, check for leaks and so on without having power to the system itself so if it does have a leak, chances are it won't hurt anything.

I can power the pump separately no problem. Just when I see I have air pressure leaks, makes me nervous to put fluid in. Did you do hard line cooling or soft?

[Beerzerker]

Soft, that's what I'm setup for here.

[ChrisZH]

10 minutes ago, Beerzerker said:

Soft, that's what I'm setup for here.

Yea all my builds with soft tubing ive never leak tested. I feel it's almost impossible to leak with soft and compression fittings.

[ChrisZH]

6 hours ago, Sakaki_Makio said:

Air is less viscous and smaller particles than water, so water gets more difficulties to pass through a micro gap which air can pass through it.

 

I dry tested my loop with it 4x, and I got almost 1 strip drop on the indicator after 2-3 hours on the leak tester when I air pressured my loop. 1 strip is veeeeeeery small delta pressure that water cant pass through it while air itself which less viscous and smaller particles need 3 hr to lose that tiny amount of pressure.

 

You'll be fine.

So if after like 4 hours I have like a .75 strip drop in pressure. Could this be a margin of error thing? Or should I investigate some more? The only other component I haven't checked is my CPU blocks screws. Maybe there is still some more that needs to be tightened with the screws. But everything else is as tight as it's gonna be. All the fittings and tubes are perfect fits. I guess once I take off the CPU block the only thing I can do left is to run liquid through it and see what happens.

[ChrisZH]

Well, I have officially tightened every screw on every block. Doing one last leak test, and then I guess I'll have to put liquid in to know more if it fails

[Beerzerker]

9 hours ago, ChrisZH said:

Yea all my builds with soft tubing ive never leak tested. I feel it's almost impossible to leak with soft and compression fittings.

Actually it is possible for them to leak, ran into that before here.

 

If for some reason the lines have a bit of downward/sideways "Pull" (Such as the weight of the hose itself pulling down on it), that can mess up the compresssion seal and they will leak. I lost a board and GPU from it for that reason, that's why for my setup I'm using the older style hose clamps.

With hard tubing you really won't have that so much aside from it having to simply hold/support the mass of the tubing and coolant within but it shoudn't flex enough to ruin the seal once it's achieved. 

 

Do know my setup is a competition cooling system I use for benchmarking, using it right now to do some runs in Super PI and yes, I can and have beaten results from more exotic forms of cooling.

For me it's not about looks, it's all about function.

 

BTW I've never really known a typical system to build much pressure, it's not like you'll have x amount of PSI going on with it like I've dealt with at the job I held working on cooling towers and so on.

Maybe a couple of pounds (If that) dead-max is all I'd ever expect to see from a typical PC cooling setup.

[ChrisZH]

6 hours ago, Beerzerker said:

Actually it is possible for them to leak, ran into that before here.

 

If for some reason the lines have a bit of downward/sideways "Pull" (Such as the weight of the hose itself pulling down on it), that can mess up the compresssion seal and they will leak. I lost a board and GPU from it for that reason, that's why for my setup I'm using the older style hose clamps.

With hard tubing you really won't have that so much aside from it having to simply hold/support the mass of the tubing and coolant within but it shoudn't flex enough to ruin the seal once it's achieved. 

 

Do know my setup is a competition cooling system I use for benchmarking, using it right now to do some runs in Super PI and yes, I can and have beaten results from more exotic forms of cooling.

For me it's not about looks, it's all about function.

 

BTW I've never really known a typical system to build much pressure, it's not like you'll have x amount of PSI going on with it like I've dealt with at the job I held working on cooling towers and so on.

Maybe a couple of pounds (If that) dead-max is all I'd ever expect to see from a typical PC cooling setup.

I mean none of my tubes are really crazy bends. It's all just 90's and straight lines. But I'm like 99% sure that all my runs have great contact with the fittings and o-rings.

[Tonberry]

On 2/15/2020 at 1:06 PM, ChrisZH said:

So if after like 4 hours I have like a .75 strip drop in pressure. Could this be a margin of error thing? Or should I investigate some more? The only other component I haven't checked is my CPU blocks screws. Maybe there is still some more that needs to be tightened with the screws. But everything else is as tight as it's gonna be. All the fittings and tubes are perfect fits. I guess once I take off the CPU block the only thing I can do left is to run liquid through it and see what happens.

Here is mine.

 

I didnt see any drop after 30 minutes, but when I let it pressurized more than 2 hours, I saw 1 strip drop.

Between 0.5 to 0.75 bar there are 10 strips, so that's 0.025 bar / 0.36 psi for each strip. That is a veeeeeeeeery tiny amount of pressure for air that water will get difficulty to leak out. My loop was dropped 1 strip for over 2 hours when I was still unsure.

But I work with alot of pressurized equipment in my real life job so I still feel comfortable with that small of pressure drop in my loop after all.

 

For now, my loop is running for around 2 weeks currently, and I didnt see any liquid decrease in my reservoir, and no indication of leak inside my PC.

[ChrisZH]

1 hour ago, Sakaki_Makio said:

Here is mine.

 

I didnt see any drop after 30 minutes, but when I let it pressurized more than 2 hours, I saw 1 strip drop.

Between 0.5 to 0.75 bar there are 10 strips, so that's 0.025 bar / 0.36 psi for each strip. That is a veeeeeeeeery tiny amount of pressure for air that water will get difficulty to leak out. My loop was dropped 1 strip for over 2 hours when I was still unsure.

But I work with alot of pressurized equipment in my real life job so I still feel comfortable with that small of pressure drop in my loop after all.

 

For now, my loop is running for around 2 weeks currently, and I didnt see any liquid decrease in my reservoir, and no indication of leak inside my PC.

That's what I've been hearing. I'm doing one last test after work. I did it yesterday after I tightened every single screw in all my blocks. Didn't see any drop after 3 hours. Gonna test again for an hour or two, then fill it up and hope for the best I guess. But if there is a drop it's like .25 - .5 of a strip drop

[Tonberry]

14 minutes ago, ChrisZH said:

But if there is a drop it's like .25 - .5 of a strip drop

That drop is less than mine. You'll be fine.

[ChrisZH]

9 minutes ago, Sakaki_Makio said:

That drop is less than mine. You'll be fine.

I always considered that like margin of error. Like something moved or idk lol

[ChrisZH]

So I have been leak testing at 0,75 bar. Not a single movement of the dial. Looks like there was a loose screw on a block somewhere.