Question Looking for small diameter CPU fan

ptTimeBldr

Great
Nov 13, 2022
135
13
85
Right, the fan that came with the AMD Ryzen 5600G can't be seated or secured because the memory modules and another part of the board are in the way, so I'm looking for a replacement fan that has a smaller diameter, ideally not more than 80mm. Also the fan's mounting assembly can't be much wider than 80mm.
 
Right, the fan that came with the AMD Ryzen 5600G can't be seated or secured because the memory modules and another part of the board are in the way, so I'm looking for a replacement fan that has a smaller diameter, ideally not more than 80mm. Also the fan's mounting assembly can't be much wider than 80mm.
Which MB and case ? Tower type coolers usually have smaller footing but may interfere in tight(narrow cases. Typical 120mm tower coolers are about 150mm tall.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
Which MB and case ? Tower type coolers usually have smaller footing but may interfere in tight(narrow cases. Typical 120mm tower coolers are about 150mm tall.
Thanks for helping. The MB is a Gigabyte A520I AC ITX, but the case is an open-frame type so not too many constraints. Thing is though, that the current fan has a diameter a shade under 100mm, but this is too big. I need one that's no more than 80mm, and just as important, the mounting assembly can't be much more than 80x80mm.
 
Thanks for helping. The MB is a Gigabyte A520I AC ITX, but the case is an open-frame type so not too many constraints. Thing is though, that the current fan has a diameter a shade under 100mm, but this is too big. I need one that's no more than 80mm, and just as important, the mounting assembly can't be much more than 80x80mm.
Yeah, I can see the problem, OEM coolers have 92mm fan and whole cooler a bit wider. You also have to take in account that G version of 5600 runs hotter than non-g so OEM cooler is "Just enough" and would recommend a better one.
Cooler master 212 evo for instance is 120 x 79.6 x 25mm (DxŠxV) whole with base even smaller so that's less than 80mm at widest.
 
  • Like
Reactions: ptTimeBldr

ptTimeBldr

Great
Nov 13, 2022
135
13
85
Yeah, I can see the problem, OEM coolers have 92mm fan and whole cooler a bit wider. You also have to take in account that G version of 5600 runs hotter than non-g so OEM cooler is "Just enough" and would recommend a better one.
Cooler master 212 evo for instance is 120 x 79.6 x 25mm (DxŠxV) whole with base even smaller so that's less than 80mm at widest.
Great, that's really helpful. So just making sure, the diameter of the fan for the Cooler master 212 evo is 79.6mm and the base is less than 80mm?
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85

ptTimeBldr

Great
Nov 13, 2022
135
13
85
maybe this helps

glad you getting somewhere :)
Yes, I'm having fun aren't I :) I saw that vid some hours ago, we must be thinking alike lol! It seems good, but look at the size of the heatsink, even though it might just fit in my ITX mobo because of the vertical fan. In the end I decided to go for this cooler/fan for AMD AM4 processors. It seems fairly good spec to me, will certainly fit and won't add too much extra weight to the mobo, plus I can use the existing mounting plate.
 
Last edited:

Karadjgne

Titan
Ambassador
mITX has always presented its own special challenges, more so with later boards since ppl demand options, vendors include those options in order to 'one-up' the competition, but those options have nowhere to go as the real estate just isn't there on basically an ATX mobo shrunk down half size.

So just like nature, what can't go out, must go up and you end up with a cpu socket completely surrounded by heatsink and extras that are double the height or more of their ATX big brothers.

I forsee issues with that gelid cooler. It's built for a 95w max TDP, which is lower by a bunch than the regular stock coolers, which aren't bad (now) but could be better.
We recorded a maximum of 73C and an average of 65C during an extended period of multi-threaded stress tests while a simultaneous instance of Furmark hammered the overclocked iGPU. However, this is with the Corsair H115i's fans cranking away at full speed.
That's a 280mm, 300w capacity liquid cooler. And the overclock wasn't that severe, less than 100w.

I'd be looking for much larger cooling, something along the lines of a beQuiet DarkRock TF2.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
mITX has always presented its own special challenges, more so with later boards since ppl demand options, vendors include those options in order to 'one-up' the competition, but those options have nowhere to go as the real estate just isn't there on basically an ATX mobo shrunk down half size.

So just like nature, what can't go out, must go up and you end up with a cpu socket completely surrounded by heatsink and extras that are double the height or more of their ATX big brothers.

I forsee issues with that gelid cooler. It's built for a 95w max TDP, which is lower by a bunch than the regular stock coolers, which aren't bad (now) but could be better.

That's a 280mm, 300w capacity liquid cooler. And the overclock wasn't that severe, less than 100w.

I'd be looking for much larger cooling, something along the lines of a beQuiet DarkRock TF2.
But ITX systems have been around for many years already, so there can't be excuses any more on the part of motherboard/cpu/etc manufacturers. Either they're talking/listening to each other or they're not, and it's always been a case of supply and demand, not demand and supply.
Why larger cooling? Why more watts? As long as the heatsink/fan delivers enough cooling, I'd rather have a less bulky unit that takes less Watts.
 

Karadjgne

Titan
Ambassador
The wattage rating on a cooler is its saturation point. Everything prior to that affects efficiency. Consequently, the efficiency curve is non-linear, it's logarithmic. You'll have a somewhat flat curve upto the 50-70% point, and then it starts climbing, rapidly.

So to keep max temp lower on the curve, you'll need higher capacity, generally 1.5x-2x cooling capacity vs wattage output. On a 1:1 ratio, 100w = 100°C

Something like this.

That tiny gelid would be absolutely fine for an older i3 that was hitting 50-60w maximum output, as it's rated max is 95w.
But yours is a 65w TDP that easily can hit 80w+ after boosts are applied and under heavy igpu usage.

Unless you have very good airflow to remove that radiated heat from the surrounding area and lower the proportional ambient temp, you'll be idling in the 50's on average, general use will be closer to 60-70° and heavier use with igpu will be closer to high 80's +, which for a Ryzen means it looses performance as a Ryzen boosts according to voltages, workload and Temps.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
The wattage rating on a cooler is its saturation point. Everything prior to that affects efficiency. Consequently, the efficiency curve is non-linear, it's logarithmic. You'll have a somewhat flat curve upto the 50-70% point, and then it starts climbing, rapidly.

So to keep max temp lower on the curve, you'll need higher capacity, generally 1.5x-2x cooling capacity vs wattage output. On a 1:1 ratio, 100w = 100°C

Something like this.

That tiny gelid would be absolutely fine for an older i3 that was hitting 50-60w maximum output, as it's rated max is 95w.
But yours is a 65w TDP that easily can hit 80w+ after boosts are applied and under heavy igpu usage.

Unless you have very good airflow to remove that radiated heat from the surrounding area and lower the proportional ambient temp, you'll be idling in the 50's on average, general use will be closer to 60-70° and heavier use with igpu will be closer to high 80's +, which for a Ryzen means it looses performance as a Ryzen boosts according to voltages, workload and Temps.
I understand what you're saying, I think - although I'm no technician. However I remember like it was yesterday when I had a SFF build in 2009 that I used for everything, including gaming. It had an AMD Athlon dual-core cpu, an after market 70mm heatsink/fan unit (the one supplied with the Athlon was a bit too wide), 2 x 512MB SIMM's and a 1GB graphics card by MSI. There was just one 80mm case fan. Yet I was able to play on a regular basis games like Second Like and Entropia Universe without any real issues. there were no overheating problems and graphics rendering was generally ok.
Fair enough, 70mm is bigger than 65mm but not by much, but the way the fan for the gelid is embedded in the heatsink suggests that sufficient airflow would be directed onto and over the cpu.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
could you post a picture of the part that block cpu cooler install and full name and model of ram .
As you can see from the images, the memory modules and the rear of the main I/O panel blocks the heatsink/fan unit whichever of two possible ways I orientate it. The ram modules are Corsair.
20221119-054551.jpg

20221119-054516.jpg

20221119-055718.jpg
 
Last edited:

Karadjgne

Titan
Ambassador
The lid on that fan can be 'spun' 90° so the logo is at the top of the board and doesn't infringe on the heatsink or ram
View: https://youtu.be/p8qeLXjiVms
Just remove the ram, mount the heatsink with logo over ram slots, move the fan around 90°, replace the ram. It's a considerably better cooler than the gelid, having an extra @ 50w capacity.

There's no such thing as 'over-cooling' a cpu, any extra capacity simply means fans don't have to work as hard, but it's all to easy to 'under-cool' a cpu, which with a Ryzen means a considerable loss in performance.
 
Last edited:

ptTimeBldr

Great
Nov 13, 2022
135
13
85
The lid on that fan can be 'spun' 90° so the logo is at the top of the board and doesn't infringe on the heatsink or ram
View: https://youtu.be/p8qeLXjiVms
Just remove the ram, mount the heatsink with logo over ram slots, move the fan around 90°, replace the ram. It's a considerably better cooler than the gelid, having an extra @ 50w capacity.

There's no such thing as 'over-cooling' a cpu, any extra capacity simply means fans don't have to work as hard, but it's all to easy to 'under-cool' a cpu, which with a Ryzen means a considerable loss in performance.
Yes I could remove and rotate the fan lid (I did think of this), but this does not solve the problem of the heatsink resting on the rear of the I/O panel as clearly shown in the images I uploaded. So it is still not possible to insert the heatsink.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
Problem with OEM coolers like that is that they are almost or just as wide at bottom so they can't clear nearby elements. If you want a horizontal cooler look at one like this
https://noctua.at/en/products/cpu-cooler-retail/nh-c14s
Heat pipes raise it above the level of RAM and other elements but body and fan can be quite large and that helps cooling a lot,
I agree, this is exactly the problem, at least with ITX motherboards anyway. I've never had this issue with mATX. Thanks for the link, but I've already ordered the gelid heatsink/fan, so must go with it now.
 

Karadjgne

Titan
Ambassador
Doubtful. Most of the 'they leak and will kill your stuff' facts and theory are leftover 300th hand stories from 20 years ago when AIO's first showed up and had a corrugated plastic tube. That tubing was flexible, to a point, but installers generally managed to twist it unduly or lift the rad by the pump and stress the connectors etc. Which then sprung a leak. As is, the coolant used is not electrically conductive in today's AIO's, and the failure rate is like 0.01%.

If, for instance, Corsair sold 1M aios last year, at a 0.01% failure rate that's 100 ppl doing an RMA. Out of those, 90 ppl rma'd for dead pumps or other unknown reasons like yours, guesstimates. That leaves @ 10 ppl with a leak. Out of those 10,1 person had a catastrophic leak and made a youtube video about the damages and safety concerns about using an aio, and here's the proof! That 1 video goes viral, and gets 1M hits, so now there's a million ppl who are sure that aios are the devil and will destroy your pc. And they get on reddit, here, other forums and inform 10 more million ppl.

Now you have half the population absolutely positive they'll get a leak, aios are bad news, stay away from them, and no mention other than Amazon reviews about 'it works, fast shipping' goes to the 999,900 ppl who did buy a Corsair aio last year and are perfectly happy with the product, or the 90 others who rma'd their defunct aio and got a refund or replacement with an apology.

All it takes is 1 video. Hate to break it to the world in general, but a full custom loop almost always leaks on first fill, if it doesn't, it's somewhat of a fluke, so an AIO's failure rate is somewhat laughable in comparison.
 

ptTimeBldr

Great
Nov 13, 2022
135
13
85
Doubtful. Most of the 'they leak and will kill your stuff' facts and theory are leftover 300th hand stories from 20 years ago when AIO's first showed up and had a corrugated plastic tube. That tubing was flexible, to a point, but installers generally managed to twist it unduly or lift the rad by the pump and stress the connectors etc. Which then sprung a leak. As is, the coolant used is not electrically conductive in today's AIO's, and the failure rate is like 0.01%.

If, for instance, Corsair sold 1M aios last year, at a 0.01% failure rate that's 100 ppl doing an RMA. Out of those, 90 ppl rma'd for dead pumps or other unknown reasons like yours, guesstimates. That leaves @ 10 ppl with a leak. Out of those 10,1 person had a catastrophic leak and made a youtube video about the damages and safety concerns about using an aio, and here's the proof! That 1 video goes viral, and gets 1M hits, so now there's a million ppl who are sure that aios are the devil and will destroy your pc. And they get on reddit, here, other forums and inform 10 more million ppl.

Now you have half the population absolutely positive they'll get a leak, aios are bad news, stay away from them, and no mention other than Amazon reviews about 'it works, fast shipping' goes to the 999,900 ppl who did buy a Corsair aio last year and are perfectly happy with the product, or the 90 others who rma'd their defunct aio and got a refund or replacement with an apology.

All it takes is 1 video. Hate to break it to the world in general, but a full custom loop almost always leaks on first fill, if it doesn't, it's somewhat of a fluke, so an AIO's failure rate is somewhat laughable in comparison.
If you're trying to give me an even bigger headache you've just succeeded ;) But don't worry, I have no intention of going for an aio build.
 

Latest posts