Monoblock on 2x 360 rads performing sub-par

[Thick8]

Get one of those cheap brake system hand vacuum pumps from the auto part store. They come with a small plastic reservoir. Fill it (the reservoir) with your coolant. The hose inside the reservoir will suck the fluid into your system when you release the vacuum. Hook it to your system's reservoir and put your system under a vacuum. This is a good time to look and listen for leaks (much better than just running the system and watching). A vacuum will identify a leak in a matter of seconds. If no leaks are present, slowly release the vacuum. This will draw the fluid from the tool's reservoir into you system, filling any voids. You will probably have to do this a few times. Once you have evacuated as much of the air in the system as possible you can run the pump on low speed while continuing with the evacuation procedure until all air is evacuated from the system.
The benefits are:
1. The ability to locate and correct any system leaks prior to filling the system with coolant.
2. Ensure that all air is removed from the pump prior to starting so you don’t run it dry. Manufacturers state that a coolant pump can be damaged by running it dry for just a few seconds. Air bubbles aren’t good for the pump either.
3. Remove all air bubbles and pockets from the system to maximize system cooling capacity and reduce the possibility of cuprosolvency.

Link to pump.
https://www.harborfreight.com/brake-bleeder-and-vacuum-pump-kit-69328.html

And lastly a couple of WARNINGS:
Lower than ambient pressure (vacuum) is a very powerful force. Do not pull more than 20” of vacuum as indicated on the vacuum gauge.

Be careful to make sure that any flexible tubes don’t completely squeeze flat while the computer is running or you will deny coolant flow to your system components

 

[EarthDog]

Yes, I tip my case to and fro as well as power off and on to get air pockets out. It is a method which has not failed me for over a decade of water cooling. I would agree with a vacuum if and only if, he could not get the bubble out and it was causing issues. So far, I haven't seen much of an attempt at doing so the free and easy way. Ive never had to use other methods than that and have built literally dozens in the past for myself, friends, and clients alike.

 

[Lochekey]

I have to agree with ED on this a vacuum seems line a little overkill. I am not saying that it would not or has not worked but on a small scale system like this it is unnecessary. As ED stated, there is plenty of systems out there that got the air out with the shake rattle and tool method. This conversation reminds me of the old argument about adding Water wetter or other surfactant to remove air. While it worth it really is not necessary in this application.

 

[Thick8]

The guys who put water cooling systems on their computers are talking to me about overkill. Really?!? You guys slay me.

 

[Thick8]

Just to share. I redid my system yesterday to add a pump and MB monoblock.
After putting it together I vacuum checked for leaks, filled it with coolant, and had it up and running in about an an hour. No surrounding everything with paper towels and then having to scramble because there was a leak somewhere. No having to let it run on a remote PS for 24 hours to verify there are no leaks. I obviously can’t tip the case to remove any air as it is hooked into the geothermal system.
Take a look at the hose going from the GPU to the CPU. See any air bubbles at all?
I find it hard to believe that you two, more than anyone, would criticize a process that can improve the water cooling effort.

 

[Conumdrum]

I don't see a problem with your methods. It's pretty cheap to do. We didn't send monkeys first to space for any important reasons. We did spend mazzilions to do that before we sent a vacuum pump carrying human.

Thick8, LOL HAHA you have become the monkey man in WCing space! Rock on!!!

One last thing? Make a minor loop of bits etc on a table and make a leak when you build it. See the sounds etc and how we can use your methods. Seriously

Happy new years you folks

 

[corpsejockey]

Out of curiosity, is this a 1700 or a 1700x?

 

[Allubz]

Very interesting reads everyone! In my case it seems doable to flip the system and tune the pump, but a vacuum does get the job done by principle.

I'm running a 1700, but bought it right at introduction where the lottery, I believe, was quite in favour of these chips being on par with the 1800(X). Why do you ask?

I'm guessing that with my current rads I'm able to widthstand about 350-400W? Or is that on the low side. Anyhow, I imagine the Ryzen chip taking up ~150, and the Ti ~225W, so that doesn't leave me with much space for a second Ti, is your thought? When designing the system I looked to an as silent as possible option with as many as reasonable radiators; only the Coolstream SE fits in 2x 360 setup. Is it too miscalculated to hope for ~2x120mm surface per component (CPU + 2x GPU)?

 

[Blaylock]

It honestly depends more on the CPU/GPU. Some just have much higher TDP so they make blanket statements like that impossible. That's not to say you're wrong, there was a time when some people here would suggest a 120mm per component, but over time it was determined to not be accurate enough.

Radiator thickness, fin density, and fan speed are also major contributors.

 

[corpsejockey]

I'm running a 1700, but bought it right at introduction where the lottery, I believe, was quite in favour of these chips being on par with the 1800(X). Why do you ask?

Just making sure the reported temps weren't the result of the 20 deg offset on the x chips.

 

[Blaylock]

Correct me if I'm mistaken here but didn't they correct that 20° offset?

 

[corpsejockey]

Yeah, I believe it has been corrected in the ryzen master software maybe 6 months ago. Did they release micro code for global fix or did all hardware monitor software have to patch for the offset?

 

[Blaylock]

I thought they did but TBH I don't own one so I didn't pay that much attention. I just remember reading it a while back.

 

[GTXJackBauer]

OP, you have more than enough rad for your loop. Heck even for 2 GPUs in SLI/Xfire and still be able to run your fans at low speed for quiet operations while maintaining great deltas.

I would compare those rads to the Alphacool ST 30s while these rads are only 2mm smaller.

As for this whole vacuum procedure, that's just a fancy way of doing things but not the norm. Sure you could do all that extra work but if you wanna do it the norm, run your pumps @ 100% in the beginning during leak testing to help get those bubbles out and than bring them down to normal operations. It could take days, weeks or even month for most of the air to make its way to the reservoir but it will suffice. Just top off your res and all will be well.

 

[EarthDog]

Just to share. I redid my system yesterday to add a pump and MB monoblock.
After putting it together I vacuum checked for leaks, filled it with coolant, and had it up and running in about an an hour. No surrounding everything with paper towels and then having to scramble because there was a leak somewhere. No having to let it run on a remote PS for 24 hours to verify there are no leaks. I obviously can’t tip the case to remove any air as it is hooked into the geothermal system.
Take a look at the hose going from the GPU to the CPU. See any air bubbles at all?
I find it hard to believe that you two, more than anyone, would criticize a process that can improve the water cooling effort.

View attachment 195690

Its the level of effort for the returns Thick. If i dont have to spend $30 and used methods that work 9/10 times, please understand where i am coming from. Its a good idea, do not misunderstand me, but not needed in the overwhelming majority of cases.

8t can improve things, but only if there is a major bubble stuck. Otherwise, nothing to negligible performances increases. AKA, good luck at measuring it and not correlating a cause.

PS - i dont have air in my system either.

 

[maxfly]

sadly i have air in my tubing that will not get out of my gpu blocks. its been there for a couple of months now. no manner of tossing and turning will expel the little sobs. my case is to friggin heavy to have to do all that! ill give the brake pump a try when i tear my rig down to install a new pump later this week. i haaaaate bubbles.

 

[Conumdrum]

Are temps okay? If so you got a nice feature! Not a big deal.

 

[Allubz]

Alright, the replacement reservoir came in, and I'm out of water to top off. Not sure if it matters, but here we go:

https://imgur.com/a/wTrkV

Pictures 1-3 show the system, and two spots that still have bubbles. I can't seem to move these bubbles, as I've been moving around my system for a good couple of times. I'm especially surprised by the b ubble in the downward tube aligned with the reservoir.

Pictures 4-6 show a few things:
4: Lowest idle temperature
5: "higher" idle temperature - The system jumps in between roughly these two temps every 5-10 seconds, to come down again. Not sure what forces this
6: System after 1:30 minute load. 75 Celcius just seems too much to me.

Water temperature is quite stable, roughly at GPU idle temps (28-29C). Room temperature should be around 24C I imagine.

So, what do you think? Should I just redo the CPU block to see if I accidentally forgot TIM, or that it is somehow badly aligned? Would getting rid of the last bubbles somehow solve my load temperatures?

My expectations were roughly that my 8-core Ryzen could do 4GHz at about 50-65C load. Now I'm looking at 3.80GHz and 75C+

 

[maxfly]

thats a whopper you have in your res line. hard to believe that your pump cant push that out. i would try cycling your pump on and off. that should help move it through. if that doesnt work your going to have to work your case around in a circle. basically following the routing of your loop until it gets worked around into the res.
the only thing i can think of with your temps is that somehow your block isnt seating properly or the tim is getting pushed around the outside of the cpu? how did it look when you remounted? tim spread look good? anything unusual when you pulled the block? i dont have a ryzen so i cant compare temps unfortunately. maybe someone who has one will show up and do a comparison.

 

[Allubz]

I'll get to turning the system a bit more tomorrow. Last night wasn't my best, so I barely had the strength to lift the case today

If I look at AIO temps people seem to be able to get <1.4V 3.9-4.0GHz, around 70-75C load. I'm looking at worse, but perhaps my secondary voltages need more work.

https://www.reddit.com/r/Amd/comments/6ahc51

I'll see that I get those bubbles out soon.