New to watercooling.

[monarchco]

So I've always been a person who wants a very robust setup. and its FAR from my first time building a PC. I'd consider myself very experienced at PC building, I just never ventured into liquid cooling, and any time I needed crazy cooling, I just resorted to something like a Cryrorig R1 Ultimate.

 

With that I'm looking at starting off with a bang. I want something robust(currently planning a way to do parallel cpu/gpu cooling, instead of sequential), but I am looking at doing multiple pumps. With that said, my question with the average DDC or D5 pump, is will they inhibit waterflow if they stop? e.g. if 1 fails, can the other pump continue to move the liquid, or will the failed pump prevent any liquid from going pass it? This is assuming they're in a sequential format. If they're in parallel it obviously isn't an issue.

This is like asking if an engine is mechanical(if it breaks it'll seize up), or electromagnetic(if it breaks the turbine/rotor/etc can spin freely)

 

 

tldr; will a failed pump prevent flow through it?

[2SidedPolygon]

Well, water cooling loops tend to be filled with the computer off, so yes, the liquid can pass through a pump that's not running.

[monarchco]

Just now, 2SidedPolygon said:

Well, water cooling loops tend to be filled with the computer off, so yes, the liquid can pass through a pump that's not running.

When a computer is off, the loop isn't running. liquid isn't passing through the pump, its just sitting inside it.

 

My question is not answered.

[JammyCreedog]

I would go with a single res/pump combo to start and plan to install a flow indicator wheel in your loop somewhere, no sense in re-inventing the wheel. If you find down the road that flow is inadequate, add another pump then, but start fairly simple.  

 

And to answer your question, no a static secondary pump will not impede the flow in the loop that much. Water impeller pumps have much more space between the blades and the inside wall of the pump housing than say a case fan's blades.

 

So you plan on doing a cpu block, a gpu block and what size radiator? What case will you be building in?

[monarchco]

3 minutes ago, JammyCreedog said:

I would go with a single res/pump combo to start and plan to install a flow indicator wheel in your loop somewhere, no sense in re-inventing the wheel. If you find down the road that flow is inadequate, add another pump then, but start fairly simple.  

 

And to answer your question, no a static secondary pump will not impede the flow in the loop that much. Water impeller pumps have much more space between the blades and the inside wall of the pump housing than say a case fan's blades.

 

So you plan on doing a cpu block, a gpu block and what size radiator? What case will you be building in?

Yeah but 2 pumps working at lower load not only means safety redundancy, but also reduced load on both(much like how a 4 core CPU thats underclocked 1Ghz will insanely outlive a CPU thats running at factory spec) for longevity?

 

Thanks for the answer though. Glad to know that if 1 fails, the whole cooling solution won't grind to a halt and cause my system to hit TJ

 

 

Evolv X

 

2x 360mm rads

[For Science!]

2 minutes ago, monarchco said:

Yeah but 2 pumps working at lower load not only means safety redundancy, but also reduced load on both(much like how a 4 core CPU thats underclocked 1Ghz will insanely outlive a CPU thats running at factory spec) for longevity?

 

 

 

Evolv X

 

2x 360mm rads

Its actually the other way around. The penalty for a failed pump is fairly neglible in a sequesntial loop but heavy when in a parallel configuration. There is little to ni advantage in having a two pumps in parallel in terms of performance or redundancy.

[monarchco]

Just now, For Science! said:

Its actually the other way around. The penalty for a failed pump is fairly neglible in a sequesntial loop but heavy when in a parallel configuration. There is little to ni advantage in having a two pumps in parallel in terms of performance or redundancy.

Oh no I'm not talking about 1 pump in each line(1 in cpu, 1 in gpu), but single line, then split to parallel, then pumps, then back to single line. That parallel portion in this case would be exclusively for the pumps, then only later on go back to parallel for cpu/gpu. This would prevent CPU or GPU from being starved if 1 pump failed.

The function of this would be very similar to bypass surgery.

 

Thanks for the clarification though. I'm glad I don't have to worry(as much) about a pump failure if there is another in sequential.

[For Science!]

1 minute ago, monarchco said:

Oh no I'm not talking about 1 pump in each line(1 in cpu, 1 in gpu), but single line, then split to parallel, then pumps, then back to single line. That parallel portion in this case would be exclusively for the pumps, then only later on go back to parallel for cpu/gpu. This would prevent CPU or GPU from being starved if 1 pump failed.

The function of this would be very similar to bypass surgery.

 

Thanks for the clarification though. I'm glad I don't have to worry(as much) about a pump failure if there is another in sequential.

Yes, my statement is still correct. These pumps are gravity fed and a failed pump is essentially a path a lower resistance and you will lose pressure. This has already been extensively looked into many times.

 

https://www.google.com/amp/s/martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/amp/

 

[monarchco]

2 minutes ago, For Science! said:

Yes, my statement is still correct. These pumps are gravity fed and a failed pump is essentially a path a lower resistance and you will lose pressure. This has already been extensively looked into many times.

 

https://www.google.com/amp/s/martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/amp/

 

Oh I'll definitely read through that. I'm not familiar with how these pumps work, but I would have assumed the working pump, being active, would have ended up creating a low pressure area right behind it(at the intake), drawing water towards it, while the dead pump "lane" would effectively turn into an eddy.

 

Now on another question, how can I best go about creating a parallel system, without 1 line seeing lower cfm(I guess itd be L/hr for this smaller use case?), than the other?

[For Science!]

6 minutes ago, monarchco said:

Oh I'll definitely read through that. I'm not familiar with how these pumps work, but I would have assumed the working pump, being active, would have ended up creating a low pressure area right behind it(at the intake), drawing water towards it, while the dead pump "lane" would effectively turn into an eddy.

 

Now on another question, how can I best go about creating a parallel system, without 1 line seeing lower cfm(I guess itd be L/hr for this smaller use case?), than the other?

Dont do it.

 

Think about it, if you have a split in your loop, then a pump on either side, and they rejoin afterwards, this is the parallel pump configuration you describe.

 

If a pump fails, you form a mini loop which goes pump to failed pump (in reverse) and back to the working pump, essentially bypassing the entirety of your actual watercooling loop since the cpu block/gpu block is so much more restrictive than a failed pump, the majoriry of the water will just be going around the pump parallel circuit ina futile cycle.

 

Edit: if it helps the understanding, pumps and generally all watercooling pc stuff are bidirectional, there are no one way valves. The major directionality is provided by a working pump.

 

Parallel pumps literally only hurt your loop.

[monarchco]

2 minutes ago, For Science! said:

Dont do it.

 

Think about it, if you have a split in your loop, then a pump on either side, and they rejoin afterwards, this is the parallel pump configuration you describe.

 

If a pump fails, you form a mini loop which goes pump to failed pump and back to the pump, essentially bypassing the entirety of your actual watercooling loop since the cpu block/gpu block is so much more restrictive than a failed pump, the majoriry of the water will just be going around the pump parallel circuit ina futile cycle.

 

Parallel pumps literally only hurt your loop.

let me clarify.

 

I'm not talking about parallel pumps anymore. I'm talking about parallel loops haha. Thats my bad I should have specified. You clarified for me I shouldn't do parallel pumps, which I wasn't planning on in the first place, I just wanted to make sure that sequential wouldn't be a problem if 1 died.

 

That being said I do want to parallel loops, both for looks(I like the look), and to ensure the coolest water possible hits the gpu instead of having been warmed by the cpu first(or vice versa).

[For Science!]

1 minute ago, monarchco said:

let me clarify.

 

I'm not talking about parallel pumps anymore. I'm talking about parallel loops haha. Thats my bad I should have specified. You clarified for me I shouldn't do parallel pumps, which I wasn't planning on in the first place, I just wanted to make sure that sequential wouldn't be a problem if 1 died.

 

That being said I do want to parallel loops, both for looks(I like the look), and to ensure the coolest water possible hits the gpu instead of having been warmed by the cpu first(or vice versa).

I would say just do it and keep the fingers crossed. I wouldnt bother about trying to make the gpu block more restrictive for example.

 

A typical D5 pump has adequate flow and pressure that the discrepancy of flow will not impact thermal performance.

 

Additionally there is also no provlem with plumbing everything in series either since the water is cycled so quickly through a typical loop that you do not get localised heating.

 

At 1 gpm, any segment of water is cycled through the system a couple times a minute and so a correctly setup waterloop heats up and cools down as a whole. You dont get "hot water afyer the cpu" or "cold water after the radiator" it doesnt work like this.

[cashan]

9 hours ago, monarchco said:

When a computer is off, the loop isn't running. liquid isn't passing through the pump, its just sitting inside it.

 

My question is not answered.

I think he meant that you can fill a computer with the pump off and the liquid will pass through it a little bit

[Limecat86]

10 hours ago, monarchco said:

So I've always been a person who wants a very robust setup. and its FAR from my first time building a PC. I'd consider myself very experienced at PC building, I just never ventured into liquid cooling, and any time I needed crazy cooling, I just resorted to something like a Cryrorig R1 Ultimate.

 

With that I'm looking at starting off with a bang. I want something robust(currently planning a way to do parallel cpu/gpu cooling, instead of sequential), but I am looking at doing multiple pumps. With that said, my question with the average DDC or D5 pump, is will they inhibit waterflow if they stop? e.g. if 1 fails, can the other pump continue to move the liquid, or will the failed pump prevent any liquid from going pass it? This is assuming they're in a sequential format. If they're in parallel it obviously isn't an issue.

This is like asking if an engine is mechanical(if it breaks it'll seize up), or electromagnetic(if it breaks the turbine/rotor/etc can spin freely)

 

 

tldr; will a failed pump prevent flow through it?

Yes, liquid can pass through a pump that is not running. The whole idea of running two pumps in series is that, if one fails you can still use your system while looking for a new pump to replace the broken one.

I'm using 2 DDC 3.2 pumps (serial) in my system and it still works fine if I pull the power from one of the pumps.

Of course you will lose about half of the head pressure when running 1 pump instead of 2.

 

Running two pumps in parallel is generally a bad idea...