Thats really what I was getting at with my original statement. The bottom line is the AIO pumps are very bad at moving only air. So if you end up with ONLY air trapped there then you’re in for a bad time. It would be better in theory to have the pump somewhat below the radiator but I’m fairly certain these pumps dont produce enough flow to cause damage to themselves in a sub atmospheric pressure condition.
1 Like
I agree.
The problem is that cases or hardware configuration do not always allow for the optimal installation. And the question becomes: “which is the least compromising.”
Heres what was said in the GN AIO video:
(While speaking about “front mounted and with tubes down”.)
As far as I understand, they therefore stated that this installation is an equally big compromise as having the radiator flat at the bottom with pump on top. (And I agree that radiator flat at the bottom with pump on top is real bad.)
I am (kind of) responsible for a build with an AIO installation that has to have the pump above the radiator. The reason for this post is to find out how big the compromise is (I did not believe/understand GNs compromise evaluation), and to decide whether a different cooling solution should be installed.
It looks like (till now) most agree that it is OK.
I think the GN video is meant to cover people from worst case scenario and optimal operation. I personally do not compromise on cases where this has ever been an issue.
To add to this, I think GN video was also meant to raise awareness for the pre-built manufacturers. They were also first main stream tech channel to point this out.
I had my radiator mounted the wrong way and ended up changing my case so that I can get it mounted properly. I noticed immediate improvement in performance and noise, just as they mentioned, but also feel much more comfortable using this system in the longer run.
1 Like
I hope this post is not considered to be off-topic, as this can be considered the result of the discussion that happened within this thread.
I have been thinking about creating a (graphical) “AIO mount” tier list.
The goal would be the creation of a wiki article (in this forum) containing the tier list.
Current ranking criteria are:
- Cooling Performance
- Noise
- Longevity
I came up with the following tiers:
Tier A
- Perfect
Tier B
- Still OK.
(*)
(*) This configuration was (falsely(?)) discouraged in the GN AIO video
Tier C
- May result in constant bubbling noise.
Tier F
- Increased stress on pump likely
- Reduced or total failure of cooling performance possible
- May result in constant bubbling noise
Legend
- Case:
- Radiator with input/output:
- Pump:
Need Feedback
- Other opinions on the order/assignment of tiers?
- Are the graphics readable/understandable?
- Any important cases I missed?
- Is tubing layout important?/In which cases is tubing important?
- (I couldn’t come up with a case where the tubing layout is important; which is why I omitted it).
Questions about Level1Techs wiki:
- How are wiki threads handled?
- I assume I just need to tag it as “wiki”
- Can other people edit the post?
1 Like
Tier E and Tier F are functionally identical.
This configuration SHOULD be discouraged as its still possible to trap an air pocket at the pump, though not that likely.
I can move your post to a new thread and make it a wiki if you would like. Theres no real standard. Other people will be able to edit the post if it is a wiki post.
I think this is a great idea and kudos to you for making the template.
2 Likes
The example shown in Tier F is more likely to be circulating 100% air.
The radiator could be mostly filled with water and only air would be circulated by the pump.
The same amount of water in Tier E will result in a loop that needs to circulate at least some water. (If the pump is strong enough(?))
I can see that both cases are really bad (though in a different way). I would be ok with putting both of them into the F tier.
I disagree (which is why I created this thread 
)
See for example:
So my understanding is:
Both pictures in Tier B have the exact same chance of trapping an air pocket at the pump!
I disagree. You have 0 head at the inlet as the pump is above everything. Any air in the system will accumulate at that high point then assuming the tubes are more or less straight.
You can disagree, but that doesnt make it reality.
You have at least some amount of water column providing pressure to the inlet of the pump with the first tier B photo.
You have 0 water column on inlet of the pump in the second picture of tier B.
Both would have a similar chance of having the air pocket there from the start, but only one has 0 water column on the inlet.
1 Like
So you’ve got to take into account the water in the pipes. Any water above the level of the top of the rad will be effectively exerting weight on all the water in the rad (including in the pipes and pump/block), and this will take any air out of the radiator.
If the pump’s the highest point, this is where the air pocket will form. It won’t necessarily be a constant thing, and the movement of the pump impeller will increase how much air is mixed in the water and broken into smaller bubbles, but there’s nowhere in the loop for it to not be forced around.
In those two, the dual rad would be fine, but the single rad wouldn’t be. Saying this, if the manufacturer specifies the air volume as a % of total loop volume, there would be a smaller air pocket in that AIO, so the sploshy sounds would possibly be less pronounced, assuming the same pump was used.
I think ultimately we’re splitting hairs here. I fundamentally disagree with that one instance being OK and think that last picture in Tier C should take its place. I also think it isnt the end of the world if your configuration is like this.
My argument assumes 2 things.
- theres an air pocket large enough to cause the pump to have issues pumping in the first place.
- the air in the system has somehow only accumulated there and not in the radiator at all.
I’m not sure on the volume of the reservoir they put on the ends of AIO radiators, but I’d assume if the pump couldn’t function properly being at the top, an air bubble large enough to do that would also prevent flow in the reservoir, if it was higher. In which case time to use that warranty lol
If the air is trapped in the radiator endcaps as is common, then theres really no issue there. Theres still flow but not enough velocity there to move that air from those end caps. In that case you’re fine, which is probably the most likely scenario.
Any configuration where the pump is at the highest point is bad.
Air locks are bad, air in the pump is also bad.
With the magnetic bearings most pumps use, you can easily end up with air being traped in the “bearing” wich will then not dampen vibration causing more wear and noise.
They greatly increased my understanding of your arguments.
(My assumptions were:
- “reasonable / still good water air mix” and
- “same absolute amount of air in all cases”)
OK, i unified Tier E and Tier F: both cases are just really bad.
About moving third of Tier C and second of Tier B: I believe I understand how you rank the examples: “Percentage of air necessary till pump failure.” (Please correct me if I am wrong).
Based on (what I think is) your ranking criteria, here is what I am wondering:
- When considering a small and a large diameter for the tubing to the pump:
Would the ranking of third Tier C and second Tier B always be in the same order?- For a small diameter second Tier B seams to me like it would be better.
- For a larger diameter the third Tier C seams to me like it would be better.
I added my current ranking criteria:
Especially the noise is (currently) separating Tier C from Tier B
My understanding is that any high point is susceptible to air locks, not just the highest.
(I am only arguing about the differences of mounting configurations that are currently both shown in Tier B.) Both configurations have a high point at the radiator and the pump; therefore air locks are created at both the top of the radiator and the pump in both configurations.
My hypothesis is that the difference between both configurations in Tier B is negligible for a reasonable air/liquid mixture.
(“reasonable air/liquid mixture” is of course very vague…)
Tubing diameter is really irrelevant in most cases. It can matter but for the most part doesnt due to the low flow in this application.
Air locks arent really a thing in AIOs unless they have very little fluid in them compared to what they can hold. They arent under any real amount of pressure and the flow is already pretty low in the first place that it never inhibits anything. The only issue arises when the pump is trying to pump air around. It sucks (or literally doesnt) at doing that. The speed of the fluid in the tubes is enough to provide the kick air needs to move along, but once it gets to the radiator its going to sit at the end cap due to the increase in diameter and therefore drop in velocity.
The following was written late at night… 
I am agreeing with you.
I was disproving the following assumption:
"
[because of air locks]
AND I assumed that the following was said additionally: “First figure of Tier B is good” (Please correct me if I was incorrectly assuming this!)
"
My argumentation in different words:
Assumption: “air locks are bad”
=> all high points (not just the highest) would be bad
=> both figures in Tier B would be bad.
However, we agree that the first figure in tier B is definitely OK
=> contradiction
=> our assumption “air locks are bad” must be wrong.
So, yes, I agree with “Air locks aren’t really a thing in AIOs”. If air locks would be a thing the first figure in tier B would not be an OK situation.
Now about the tier list:
Here are my ranking criteria:
- Cooling Performance
- Noise
- Longevity
2A = “second figure in Tier A”
2B = “second figure in Tier B”
2C = “second figure in Tier C”
3C = “third figure in Tier C”
I believe 2B and 3C are fairly similar in cooling performance as well as longevity; I could see a case for 3C being a little bit better (as you already described).
- My reasoning for having 2B a tier above 3C is the possible additional noise with 3C.
- In addition it should be fairly obvious that 2A, if possible, would be better than 2B. So people that do not need to mount it like in 2B are going to mount it in 2A. And people that do need to mount it like in 2B are going to be ok with it (because it is ok).
- If 2C and 3C are in different tiers there kind of needs to be a clearer distinction point between them.
- I just realized: If you look at 3C it is extremely likely that 2A is going to be a valid option. As 2A is the perfect mount I don’t think we should encourage 3C by moving it up to Tier B.
Because the top of the radiator is at the same hight as the pump. Not idea, but okay.
Air locks are bad, that is not an assumption.
In case you need proof for that, have some:
http://ceeserver.cee.cornell.edu/mw24/cee332/readings/AirInPipesManual.pdf
The assumption on my part is the airlock being big enough (or becoming big enough) to cause problems in the sense of noise /increased wear.
Your last drawing in Tier C is also fine. Yes, the outlets are the hight point, however the hoses still provide enough reserve for the pump to not fall dry.
Ok let’s say air locks are bad. Why would air locks be more bad if the pump is above the top of the radiator?
When looking at the Wikipedia page (or the YT video or article about pipes you posted) about air locks: I can see an air lock forming on each high point. The highest point is not getting a different treatment. In other words: Air is not able to find and wander to the highest point when trapped in a high point.
I think the relevant question to answer is:
Arguing about it seems to have failed. I do not have the experience to be sure about the outcome nor the ability to do the experiment. (I do not know where to go from here…)
I am fairly sure, however, that once the air is trapped in the radiator it won’t be a problem anymore. (Air is at a high point -> Air lock tells us air is getting stuck there)
I agree that it is fine. The question is: Is it significantly better than the other options in Tier C to be warranted to be at a higher Tier?
I am no expert on pump design, but I don’t think power has anything to do with this. It all depends on the pump design. Some are self priming, others are not.
This can be seen demonstrated by GN at 15:50:
Because not all cases made sense to test.
You need to test these 4:
(Blue line indicates pump level)
