Re: Cooling a HP desktop

On Mon, 6/1/2026 4:15 PM, sticks wrote:
> On 5/31/2026 10:19 PM, Paul wrote:
>> On Sun, 5/31/2026 9:31 PM, sticks wrote:
>>> On 5/31/2026 7:07 PM, sticks wrote:
>>>> The video below is the exact model of one I have.  I am worried about overheating in the summer as this is out in my garage.  At the 9:05 mark the vid talks about adding a second fan, an exhaust fan.  There does seem to be the power hookup on the board right there.  Am I correct that just having the fan on the processor now and adding this exhaust fan should really help keeping the temps down on this machine?
>>>>
>>>> <https://www.youtube.com/watch?v=N82dHQJ8TXg>
>>>>
>>>> Any remarks appreciated!
>>>
>>> I want to add one thing.  The reason I am looking into this again is because I've been monitoring the temps with Core Temp, and even though it is still not as hot outside as it will get this summer, I have noticed the temps higher than during the winter in the heated garage. Where it used to idle at about 90-95F, it now runs at about 105-110F, and if I idle for a little when I come back I've seen it as high as the low 130's.  I am guessing this is because the fan on the processor isn't running as high since the processor is not working hard?  I go back to using it and it drops back down, which would seem counterintuitive. Anyway, I need to see if I can get these temps down a bit.
>>>
>>
>> I recognize the aluminum heatsink as a 65W one. The processor is 6 P cores and 4 E cores.
>> It has a 13 TOPS NPU on it, which is not the 50 TOPS that Microsoft is looking for,
>> but any NPU is better than nothing. You should have an extra section in Task Manager,
>> a GPU performance graph and near that should be one for the NPU.
>>
>> https://www.intel.com/content/www/us/en/products/sku/241070/intel-core-ultra-5-processor-225-20m-cache-up-to-4-90-ghz/specifications.html
>>
>> Processor Base Power    65 W
>> Maximum Turbo Power    121 W
>>
>> Like a lot of OEM computers, the cooling solution is a disgrace :-/
>>
>> As you note, the exhaust fan location on the back, is not populated
>> but there is a fan header. The problem I see, is I am not locating
>> the "square hole pattern" for an industry standard square fan! You'll
>> have to check this for me, and see if one of the three hole patterns,
>> the holes happen to be a subset of the four holes on a case cooling fan.
>>
>> The airflow may start at the top of the CPU, flow through the fins,
>> then the cowling redirects the air from the bottom of the heatsink,
>> towards the exhaust vent. That is a right-angle turn.
>>
>> The front cooling vent, the fake woodgrain panel on the front seems
>> to block air movement there. Perhaps there is some secret ventilation path
>> that makes the vent a usable one. The available pictures do not allow
>> me to figure this out.
>>
>> You could cut a circle in the sliding side panel, and fit an external
>> intake fan and "pressurize the case". Remove the cowling for the rear
>> hole, so that the air will be in a rush to escape when pressurized by the side
>> intake, Instead of ruining the side panel, you may be able to
>> make a metal plate to take its place, and mount your external
>> square fan on the metal plate.
>>
>> You can cut away the cheese grater ventilation covers, and
>> go with something that allows more airflow. On my large (EATX) case
>> computer, I stared longingly at my nibbling tool and the cheese
>> grater vent, trying to figure out a way to cut the vent out of the
>> way and increase room for airflow.
>>
>> A larger fan, with lower RPMs, helps with noise. When a machine is
>> a bit smaller like that, any provided fans have to run at higher RPMs
>> because their diameter isn't all that great.
>>
>> The way I see it, you have next to no airflow as it is, and unless you
>> can find a way to pressurize the case (the side area may be sufficient
>> for a large fan), messing with the tiny fan holes will likely leave
>> a bit of a noisy situation.
>>
>> Now, you'll have to get out your metric "ruler", and see if
>> any hole pattern is there for 60mm, 80mm, 92mm, 120mm, 140mm and so on
>> in the series of standard fans. There are a few other sizes which are not
>> common. The 40mm ones are virtually useless... unless they run at 6000RPM
>> and are 1" thick or thicker.
>>
>> When I pressurized my EATX case, I was dismayed to find that the PSU did not
>> pass a lot of air from inside the case and through the vent in the back.
>> The PSU was quite resistant to being helped (the PSU sits in the bottom
>> of the case, and the intake on the PSU is down low). That machine has three intakes on
>> the roof of the case, a large fan on the front of the case as an
>> intake, and a number of other vents are taped off to encourage
>> the air to leave only via back-vents.
>>
>> So part of your problem then, is the air is stale inside the case,
>> the case air is heating up, and there is no movement to bring it
>> back down to about 35C or so.
>>
>> While you could change the computer case, all those holes for I/O would be
>> a pain in the ass to create to finish the job. It would be an endless nightmare,
>> to change cases. I like the idea a bit better, of replacing the side plate,
>> with your own side plate, with an intake fan mounted on the outside of it.
>> My daily driver used to have an external intake, the fan being mounted on
>> pieces of aluminum angle iron. I've since removed that when the latest
>> motherboard was fitted to the case.
> 
> I had a Foxconn model PV902512L which runs at 0.16 A and has the 92x25 hole pattern.
> 
> <https://www.amazon.com/PV902512L-DC12V-0-16A-3-WIRE-92x25mm/dp/B00B877C50
> 
> I removed the two shrouds as the video shows and this fan did fit in this location perfectly and the power plug in was where it showed in the video.  Once I placed the fan in there, the shroud I took off that was either directing air to or away from the CPU didn't fit.  I could try and trim it a bit to get it back, but I am leaning to thinking it is not necessary now that there is a case exhaust fan installed, so I put the covers back on and fired it up.
> 
> Wow, what a difference that makes!  Where before it would spike up every time I asked it to do something, it just hovers now at 97F which is almost exactly that 35C you mentioned.  It is still very quiet, can't really hear either fan at all.
> 
> So, unless you have any words of wisdom, I'm gonna just monitor it for awhile and make sure it keeps working.  This fan is out of a 20 year old box so I probably should get a spare just in case it pukes.  Would the same thing be OK, or should I look for something that spins faster to get even more air flowing?

Depending on whether the 92mm fan was a three pin (voltage control) or
a four pin (PWM 5V logic signal control), that will tell us whether
the fan is likely to be adjustable or not. If you put a three pin
fan on a four pin header, it tends to run at 100% speed. That's because
when the computer provides four pin control, it does not have a linear
regulator to make three pin control in its place. In modern systems,
you get one control method, not two. Part of the reason for this,
is Intel has the fan spec for four pin, and they suggest VCC stay
at 12V, as the PWM pin indicates what fraction of 12V the PCB
inside the fan should be applying to the motor. On the four pin fan,
the fan provides its own infrastructure (regulator), and the control
signal is lightweight.

   GND VCC RPM PWM    <=== PWM speed control, fixed VCC 12V input voltage.

   GND VCC RPM        <=== Variable VCC, 7V to 12V should work well.
                           If plugged to four pin header, it gets 12V always.

If you're handy with electronics, you can make your own speed controls.
I've checked the odd time, and I can no longer find a "fan rheobus controller"
to fit into a 5.25" drive bay. That's a kind of speed control with knobs,
for custom fan control. For my 110CFM Ultra can, I used a number of 0.7V silicon
diodes (1N4004) in series, to drop some voltage and run the fan at 7V to 8V or so.
And that's the one exhaust fan on my build, an exhaust fan that hauls the
warm air away from the giant cooler so the warm air will not spin around
inside the case. In that way, you adjust the exhaust fan on the back,
so that "most" of the warm CPU air exits by the short path to the back.

And you've already done that.

Now you can afford to play with the setup a bit, figure out whether
it needs "moar fan" or not.

for example, let us say you have a 92mm x 25mm thick fan. If
you want moar air, you can buy a 92mm x 37.5mm and that will move
the air with slightly more authority (for the same RPMs).

Draw the line at say, 1800RPM. A 1200RPM fan will be loud enough.
A 3000RPM fan, running full speed (they make a few that annoying),
you won't put up with that for very long. It's only the smallest
of fans that will do 6000RPM, and the air those move while doing
that, is quite impressive for such a small thing, but the noise
is not something you could tolerate.

Your fan types...

    GND VCC (NC)      Two wire fans used to be computer case common-materials.
                      There is no RPM pin so you cannot monitor the speed.

    GND VCC RPM       The three pin fan was an improvement, and it was the "primo"
                      fan for a while. Both the two and three pin, you could
                      vary VCC safely, from 7V to 12V for the 12V fan.

    GND VCC RPM PWM   These are the fancy fans now, and maybe all the headers in
                      a modern PC use this four pin pattern. You can plug a three pin
                      fan to a four pin header, but then it runs 100% RPMs. When you
                      mix four pin fan with four pin header, now you have to go into
                      the BIOS and define the "fan curve". A piecewise linear response
                      curve from sensor (CPU or Mobo temp) to fan speed.

But at least at the moment, you've corrected the worst part
of what those idiots did. It should have had that fan in there.

   Paul