Why are blower cooling bad ?

a 3-slot cooler with radial blower would be sick. if the taller heatsink and squirrel cage could bring rpms down to like 1500-1700 that would be a huge acoustic improvement over the 2900-3500 that's needed to cool 250ish watts+ and you get the heat out of the case.
 
Blower type coolers generally do not use heatpipe-based heatsinks; they typically use vapor chambers.

Excellent data, I did not know that.

I was just looking at the TITAN RTX heatsink and it's vapor chamber.
Its fins run top to bottom with 2 "fans" blowing at them, recirculation ick.
Newegg reviews cite poor cooling often enough to make me think about why.

The blower heatsinks, with vape, still have their fins running along the length of the card so the fan end fin portion get pretty much the same portion of HEAT (thanks vape) as the outlet fin area.
But the fins heat the air as it flows along the fins. So the fins see equal heat but uneven air temperature.
 
I've never had a problem with blower-type GPU coolers really, but then again I don't generally mind the noise and tend to set pretty aggressive fan curves with my all of my fans anyway. I like that with a blower-type card you can actually use the PCIE slot right below your videocard without having it impede the airflow.

Before my current RTX 2080 I really milked my previous GPUs for a long time, sticking with my 2 GTX680s and even buying a cheap matching 3rd card toward the end and running triple-SLI for a while rather than paying the cryptocurrency mining inflated prices for a new GPU. The triple-SLI config had the cards sandwiched close together and only worked because all three had blower-type coolers.

I've also never had the fan on a blower-type cooler actually fail whereas I've had numerous instances where the fans on an after-market or premium cooler failed and there is usually no easy way to replace just the fan without replacing the entire cooler. Usually end up having to use zip-ties to attach an 80mm fan to the side of the card or something ghetto like that. It's depressingly difficult to find information on the fans that many of these custom or premium coolers come with, such as if they are even ball-bearing or some cheaper type that tends to get noisy and fail after a few years of 24/7 usage. The fans usually last long enough for the warranty to expire of course, but not long enough for people like me who pass my cards to other computers and like to get 6-10+ years out of each card before I fully retire it.
 
  • Like
Reactions: noko
like this
a 3-slot cooler with radial blower would be sick. if the taller heatsink and squirrel cage could bring rpms down to like 1500-1700 that would be a huge acoustic improvement over the 2900-3500 that's needed to cool 250ish watts+ and you get the heat out of the case.

If a 3-slot blower was made the right way, it wouldn't need to run high fan speeds. The opening at the rear of the card would allow more than 3x the airflow for exhausting.
 
If a 3-slot blower was made the right way, it wouldn't need to run high fan speeds. The opening at the rear of the card would allow more than 3x the airflow for exhausting.
yep, exactly my point. <2000rpm is low for a radial fan and wouldn't be terribly loud. much better than the 3000+ that's needed for 2-slot cards.

anyone remember single-slot blower type cards? I have a (sadly deceased) Quadro FX3700 (G92) with an absolutely tiny fan and the noise was ungodly
 
I was going for a visual.
As the air flows along the line of pipes, the part of the fins associated with each pipe are sequentially seeing hotter air and so will not cool the pinky pipe as well as the first pipe which sees cooler air running along it's fin stack area.

So, I don't like designs that thumb their nose at me, I want each pipe to have the maximum cooling, not just the first one.

Practically impossible in a GPU blower heatsink. Actually, nearly impossible in any computer heatsink design with any meaningful amount of depth. The intake side of the heatsink will always get the coolest air while the exhaust side of the heatsink will get the warmest air. The only way to eliminate that issue you have is to design heatsinks that is hottest at the exhaust side and coolest at the intake side. Essentially, a countercurrent heat exchanger. While not impossible to do, it's extremely complicated keeping it within the confines of a standard computer case and the benefits are negligible at best because all surfaces are warmer than the air and are generally within 2-3 degrees C of each other anyways.
 
Why blowers are bad:

The best headphones are gaming- whether talking US$50 or US$1500 (or more)- are open back. Which means you get to hear those blowers scream while you game!

Best compromise: a blower with an AIO for the GPU die or HBM package as may be the case. Corsair does this for various companies and it is the very best stock cooling you'll get on a GPU, both in terms of performance and noise.

Any better and you'd have to do a custom loop.
 
Practically impossible

Confusion rises around my feet like the fog that arose in Denver the day it was legalised.

Either we are looking at different things that behave differently to one another, or you are totally incorrect in your understanding of the physics of heat transfer and fluid flow.

T.l.d.r:
wtf y'all talking about ?
 
Confusion rises around my feet like the fog that arose in Denver the day it was legalised.

Either we are looking at different things that behave differently to one another, or you are totally incorrect in your understanding of the physics of heat transfer and fluid flow.

T.l.d.r:
wtf y'all talking about ?

You want all heatpipes to get equally cool air. That's practically impossible and even if it were, would serve no meaningful benefit.
 
Nah, in the Design R3 case I used to have, the reference cooler worked fine and it was not loud, since the case was designed with sound deadening material on it. Now, when I tried 2 cards at once, well, lets just say that was an experiment that did not last long at all.
I tried an r9 290 in that same case and it was audible from the next room with 50pct fan speed. It was loud and whiney.
 
Once the heat gets up into the pipes where the fins are, each pipe acts as a point source* of heat to each fin that contacts it.
As the air flows along the fin surface the heat flows into the cooler air, lowering the temperature of the fin and raising the temperature of the air, which lowers the ability of the air to absorb heat from the downstream portion of the fin.

So the downstream portion of the fin will be hotter than the leading edge of the fin because the heat input from the pipe is removed by the air at a lower rate because of the lowering of the temperature difference between the two areas of the fin and the air.

So there is a temperature gradient in the fin along the air flow path.
the downstream portion of the fin is hotter than than the upstream area of fin.

* ok, NOT a point source.
A circle with arrows sticking out from the rim of the circle in all directions (in the plane of the circle).
 
The heatsink on my i3, ducted to a Noctua 120mm exaust fan. note that every bit of every fin gets the same airflow.
Which is/will be(when I do the wiring and get the shape of the duct correct) exausting all the CPU heat directly out of the case.

0329192206_HDR.jpg


Note that I still need to duct the 120 to suck air through the 1050ti HS and expell the air out the vent holes in the side of the case.

And I swapped out the .50cent pwm fan for a Noctua 40mm to cool at the Nvme boot drive.
 
Last edited:
No problem with blower type cards, in small cases and confines they can be about the only solution that works. If your case has good to great ventilation then a blower type card maybe not the best choice due to getting better performance with fan type cooling or liquid. I like to have choices. I've found dual card setups usually work better with blower type cards, otherwise the top card is significantly hotter than the bottom no matter what I've tried. If noise is a concern then blower type cards are not the best choice, if not then they can be the best choice. In other words no right or wrong answer - pick what works best for you. Not everyone has to conform to someone else thoughts, ideas or what works best for them. Seen this argument so many times.
 
I like to have choices.

Yup.

I like to get data so I can see options relative to functionality and objectively base my choices on the relationship of different data points.
Hence inquiring about blowers since my collection of data about them was pitifully sparse.

Not here to argue but to trade data.
 
The heatsink on my i3, ducted to a Noctua 120mm exaust fan. note that every bit of every fin gets the same airflow.
Which is/will be(when I do the wiring and get the shape of the duct correct) exausting all the CPU heat directly out of the case.

View attachment 178951

Note that I still need to duct the 120 to suck air through the 1050ti HS and expell the air out the vent holes in the side of the case.

And I swapped out the .50cent pwm fan for a Noctua 40mm to cool at the Nvme boot drive.

looks good. nice to see a fellow cardboard duct enthusiast (y)
 
Di georno pizza boxes are excellent as raw material ;)

And plentiful too ( my belt whimpers )
 
I tried an r9 290 in that same case and it was audible from the next room with 50pct fan speed. It was loud and whiney.

Then you had a problem with your specific card, which I can see happening. The only real issue I had is that I did not have good enough airflow to improve the thermals of the card but, it was still a beast, nonetheless.
 
How would the two impellers be mounted ?

I can see how triple slot thickness would allow the impeller to be taller for more air volume, and allow the fins to be taller both allowing lower pressure, AND allowing more fin surface.
Also allowing the exaust area to be larger, reducing pressure requirement.

But how would the second impeller work ?

Correct, a much deeper impeller for the blower and a secondary fan assisting like you would a rad. Twin-fan design.
 
a secondary fan assisting like you would a rad. Twin-fan design.

I fear I am adrift on a sea of unknown physics.

A unshrouded centripital impeller has a flow pattern similar to the pipe heat flow pattern I described above.
the circle represents the outer rim of the impeller, the arrows however are angled in the direction of impeller rotation.

That angle is why shrouds on stand alone blowers are snail shell shaped.
That shape uses the angled airflow to allow the airflow to be directed into a linear flow pattern with maximum efficiency.

In the case of a blower card I wonder if there is a internal snail shroud to direct the airflow into the heatsink ?

Now enters the second impeller, arrayed similarly to a twin fan card.
The only way I can see that working is the impeller furthest from the exaust would require ducting Around the impeller closer to the exaust.

Without such ducting the airflow pattern inside the shroud would be horrific.

With correct shrouding & ducting the fan pair would need to rotate in opposit directions so the snails could mirror each other. Both pointing at the heatsink.

The major problem I see is the total airflow of the upstream impeller must go around the downstream impeller and it's shroud.
Which would require the downstream impeller to be smaller in diameter to allow the total airflow of the upstream impeller to have enough flow area to be least restrictive at the choke point.

Ugly.
 
  • Like
Reactions: Auer
like this
Then you had a problem with your specific card, which I can see happening. The only real issue I had is that I did not have good enough airflow to improve the thermals of the card but, it was still a beast, nonetheless.
Doubtful seeing the reviews all had it too.
 
Doubtful seeing the reviews all had it too.

I know what a noisy fan is, some of those old cpu coolers from the athlon days were intentionally noisy. The reviews did not all have it too, just that it was the popular thing to bash on it. If you say you had a Fractal Design Define R3, then you would know it has sound deadening material throughout and you would not have heard the card a long way away from the computer. I had the card for 2 years before I sold it and only at the very end, when I installed it in a much more open case, did I hear the fan but, it was not terrible, just noticeable.
 
Blower cards are just fine, but if you want to overclock and push the card then the fan speed needs to come up and thats when you can run into noise issues. I ran a 290x blower cooled cards for a year then I switched over to water cooling it. The sound of the blower never really bothered me but was noticeable if the fan speed was beyond 60% and a few games could push it up there to that point. But it was not annoying or deafening by any means, just kind of a dull rumble. Now if you do what they do in reviews with open air setup and mic next to it then yeah expect some noise. Nice thing about a reference model is if you get the itch to water cool it then you wont have any issues finding a block for it.
 
Doubtful seeing the reviews all had it too.

You dont even know loud unless you have been near a delta fan at full tilt, a 290x was silent next to one of those. And yeah your card was defective if you could hear it in the next room at 50%.
 
You dont even know loud unless you have been near a delta fan at full tilt

Been there- how about the ass of an F-16 being ragged out by a test pilot? Asshats hover them over us. Stationary, nose to the sky, us in their afterburner. You don't bother trying to talk.

And yeah your card was defective if you could hear it in the next room at 50%.

You want airflow in an enclosure, air needs space, and sound escapes through that space. There's no getting around it with the enclosure itself; you just have to put more surfaces and more density between the enclosure and your ears without suffocating the system.
 
Blower cards are just fine, but if you want to overclock and push the card then the fan speed needs to come up and thats when you can run into noise issues. I ran a 290x blower cooled cards for a year then I switched over to water cooling it. The sound of the blower never really bothered me but was noticeable if the fan speed was beyond 60% and a few games could push it up there to that point. But it was not annoying or deafening by any means, just kind of a dull rumble. Now if you do what they do in reviews with open air setup and mic next to it then yeah expect some noise. Nice thing about a reference model is if you get the itch to water cool it then you wont have any issues finding a block for it.

My issue was, I had the R9 290, was able to flash it with the 290x bios and it was fully unlocked. However, since it was a reference card, overclocking was just not going to happen, even with upping the fan speed. At least with my RX5700 reference, I get a good bump from overclocking and it does not run as hot as the R9 290X.
 
I've used an A/C cooled case, closed system and the blower cards ran colder than any fan/heatsink out there. 290x/290 (yep worked great) with blowers and really no audible noise. Case was insulated (reduced noise as well as keeping the cold in) and the cards exhausted to return ducting to the A/C unit, noise channeled out and they ran mostly less than 40% speed if that much. So depending upon a number of factors blower cards can be the best choice and not the best choice.
 
I've used an A/C cooled case, closed system and the blower cards ran colder than any fan/heatsink out there. 290x/290 (yep worked great) with blowers and really no audible noise

...but then you have A/C noise ;)

No free lunch... open-air coolers are the cheapest way to cool a card quietly, and AIOs are the cheapest way to max out a card. Beyond that, you're going to spend and the solutions will get more cumbersome.
 
TANSTAAFL , Aye.

But this is not about defying entropy, it is about managing heat flow.

Phase change (heat pipes/vapor chambers) tech were a dramatic improvement over a chunk of metal for control of heat flow.
Simple, very effective, and totally out of the question for an individual to fabricate in the garage cheaply.
Cheap and readily available now.

Blower cards are very good at getting hot air out of the case but are bad at getting a LOT of hot air out of the case.

Because of design compromises required by the absolute need for the assembly to integrate seamlessly into every computer on the planet.

Being able to disregard that constraint, and optimize the setup for only the computer before me allows me to directly address the problems as I see fit.

For example, ducting a 120 fan to blow into the inside end of the shroud with another 120 on the outside drawing the air directly out of the case.
The heatsink would act as a flow straightener between the two fans.
Which would almost double the pressure across the fins, which would directly increase airflow volume through the HS.

At the cost of figuring out where to put the fans and the shape of the ductwork.

In this case before me and, no others.
 
Another problem with exhaust fans is where you place your case. Most of us put it near a wall next to our desk where that sound is going to reflect off the wall.
 
that sound is going to reflect off the wall.

Audiophiles solved that one.
A 12'x12' damping panel would be all fat and happy soaking up those annoying high frequency whines.

I am in impressed with those "new" Noctua 120s, my HSF is indeed pointed at the wall. the only time I hear anything is the woosh when my CPU goes over 39°c and the fan goes to 100%.
 
I've used an A/C cooled case, closed system and the blower cards ran colder than any fan/heatsink out there. 290x/290 (yep worked great) with blowers and really no audible noise. Case was insulated (reduced noise as well as keeping the cold in) and the cards exhausted to return ducting to the A/C unit, noise channeled out and they ran mostly less than 40% speed if that much. So depending upon a number of factors blower cards can be the best choice and not the best choice.
Hey, ;) in Florida you always have A/C noise.

Plus in this design noise is channeled into a insulated computer. Less noise then if just blowing in the room. Yes noise from an A/C unit is present. Pick your poison maybe.

I have not done this in years, tempted to do some experiments now with whole different easier method.
 
My favorite solution is to select a residence that has a room suitable for conversion to a nest of computers and networking equipment where you can't hear that crap, turn all the fans on 100% fulltime, run cables through the wall to your office monitors, keyboard, mouse, etc. This way also keeps the cat and dog hair from accumulating in the computers or them walking behind backplanes and unplugging things etc. The monitors and HIDs are still vulnerable so I use a locking sliding tray for keyboard and mouse and http://www.vizomax.com/ for the monitors. They market it for anti-children purposes but for me it's anti-cat armor for my screens. When I got the last of the nice plasma TV's the cats leaped at it and that was terrible so ever since then they don't come out of the box until that crap comes in the mail. Also handy for this you aren't overheating your room, I use a one ton portable a/c two hose with insulated hoses and a custom window adapter to keep it no higher than 78f in the room that must have been an in-home hospice room with 9 120vac single gang outlet boxes on 3 circuits. Idk why else it would be 106 ft^2 with that amount of electric capacity but I am happy regarding it.
 
Last edited:
  • Like
Reactions: noko
like this
Please unpack that a bit.
Sure, 18,000 BTU or 1.5 Ton 240w A/C unit, big box, sitting inside of big box, little box (PC) + A/C unit cooling the big box, closed system. Does that answer you unpacking question? :D
 
Hey, ;) in Florida you always have A/C noise.

Plus in this design noise is channeled into a insulated computer. Less noise then if just blowing in the room. Yes noise from an A/C unit is present. Pick your poison maybe.

I have not done this in years, tempted to do some experiments now with whole different easier method.

The problem with using home AC to cool your computer, is that it fires up intermittently and no necessarily when you want to cool your computer, and it is seasonal as well in many places.
 
Yeah I take it down during the Winter but generally April-September is air conditioning season in AR. You don't have to use it to directly cool your computer, you can use it to control the ambient temperature in the room. The computers are heating up the room, when it gets hot enough to deviate sufficiently from the central air conditioning's set point the portable a/c in the room kicks on and cools it down to the central's set point before turning off. When the ambient air in the room is not hot it helps cool the computers with their fans. All that heat has to go somewhere, it is exhausted from the computers and cool air replaces it continously. When the hot air raises the ambient in the room a few degrees F the a/c kicks on until it's back down to normal. It works great because when the a/c is running it is exhausting hot air from the room outside with the second hose. IT's also nice for humidity, idk if your region tries to imitate a swamp for most of the year but this one sure does and the air conditioning keeps the mold down. 4 tons central downstairs, 4 tons upstairs, plus the one ton portable upstairs to deal with all the equipment in such a confined space. One day I'll probably put a rack in there. The central a/c has mechanical ventilation/ inlet and outlet in the room so it works fine for the beginning and end of air conditioning season and the upstairs 100k btu gas furnace keeps it from freezing, identical one downstairs for that, backed up by some wood stoves. Anyone around here recall the Ice Storm of 2000? Winter before last it got down to 6 degrees and stayed below freezing for a week in January and it was awesome for benchmarks.
 
Last edited:
The problem with using home AC to cool your computer, is that it fires up intermittently and no necessarily when you want to cool your computer, and it is seasonal as well in many places.

He should put the AC cold inlet/outlet and PC into a storage container. That way he can set it to 60F and it’ll maintain it.
 
The problem with using home AC to cool your computer, is that it fires up intermittently and no necessarily when you want to cool your computer, and it is seasonal as well in many places.
Separate dedicated A/C unit (Window) to PC directly (what I did previously) or box which PC sits in (may experiment here some).

Anyways you can control the A/C unit, bypass the current controls, basically power to the blower/fan motor and compressor and control the temperature how ever you like. Using a temperature controller to operate motor starters or solid state relays is not that hard, add in a few protective measures as in a power relay that opens if temperature gets too high that shuts off your computer automatically can be added while keeping it simple and relatively cheap. Previously I insulated a case to keep it from sweating and cold in, hooked up ducting to and from that case with snap on and off ducting to an A/C unit. Everything gets cooled, even stock HSF end up performing well when cooled to around 0 c. I ended up running around 5c-10c for the most part and test runs at about -10c (basically ICE would not melt inside the case). Case just got very obsolete and never redid that. Now I am more thinking any case will work if you just enclosed the case with a cool enclosure which will make everything much more simpler.
 
Back
Top