Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > alt.comp.hardware > #20852

Re: Cooling a HP desktop

Show all headers | View raw

On Sat, 6/6/2026 1:19 PM, Carlos E.R. wrote:
> On 2026-06-06 17:38, sticks wrote:
>> On 6/4/2026 12:26 PM, Paul wrote:
>>> On Thu, 6/4/2026 10:23 AM, sticks wrote:
> 
> 
>>> In fact, the other day, I had to make myself a drawing
>>> of how the fans are wired in the machine across from me,
>>> as every time it does something annoying, I can't remember
>>> which wire or function controls it. Now I have the drawing
>>> to look at, before I walk over there. The computer case
>>> comes with a "fan manifold" PCB, the CPU fan header goes
>>> into the board, and six headers come out of the board.
>>> And then you have to remember which fans are running off
>>> CPU-sensitive control and which fans run off a separate
>>> header. And the drawing helps with that. If I re-do the fans,
>>> then the drawing will get updated.
>>
>> It really is a little sad that they've made it this difficult to control the temperature of your systems.  I'm sure I'm not the only one who has a box in an area without summer climate control where it is gonna get hot and humid.  For now, the combination of FanControl and Mouse Jiggler is working nicely and I'm going to leave it alone and continue monitoring.
> 
> The only way to run computers in a humid place for a  long time, is to seal the computers or use AC.
> 

You can buy computers that are hardened, with the
expectation of an exotic working environment.

Some HP machines, have had a 50C ambient spec, and
having spec in hand, we placed a HP machine in our
walk-in thermal chamber. Some of the other test
equipment in that chamber was rated for 50C as well.
And then we would run the whole shebang up to 50C,
including the DUT, and the HP computer worked just
fine under those conditions. That was not a "humidity
test run" because we didn't want to die in the chamber.

Using test chambers, you can absolutely destroy equipment,
but then you're violating the "95% humidity, non-condensing"
limitation, and allowing liquid to condense on the surface
of the DUT. A rugged computer would stop that
moist air from getting to the PCB.

Using heat pipes, you can seal a device up, and
still have a conduction path for the heat. Zalman made
a couple computer cases, one costing $1000, and there
were heat pipes in there compatible with various
graphics card designs. And the heat would flow to
heatsink fins on the outside of the "box". This was
intended for applications such as recording studios.
There may still be a need to swirl a bit of air around
the internal parts of the cabinet, to prevent hot spots
(on things not having heatpipes fitted). While the box
may have had a 400W ATX supply in it at the time (fanless),
it wasn't really a good idea with that kit, to push it that
hard. And that would not be suitable for 95% humidity,
as it may not be sealed well enough to separate the
inside from the outside.

Circuits have been "potted" in the past. The Ethernet
transceivers we used to use, those were potted and quite heavy.
If you were wearing sneakers and dropped one of those
on your foot, that would sting. Whether that kept the
environment totally at bay, is hard to say. To do that today,
would likely cost a lot more than it cost back in the
potting era.

At one time, integrated circuits shipped in hermetically
sealed packages, with a glass frit seal between the two
halves. And then at some point, they used plastic instead
and nobody gave a damn any more about the environmental part
of it. The ceramic ICs were "MIL spec" and worked from -55C
to 125C, and things like moisture should not be getting
through the ceramic. That kind of packaging was quite popular
on the ECL logic boards we made -- when you used an oscilloscope
on such boards, you would lay your hand on a hundred ECL chips
running at 55-60C, and that takes, um, "discipline" to do.
You get a bit used to it, after a while [he said, wincing a bit].
ECL just loves the heat, and really, it's not happy until
it is quite warm. If you have a cooling failure on some
of these designs, you may have to wait ten minutes for the
thing to be cool enough to take apart. The chips in that
case, were still running in mission mode just fine and
they were not complaining.

CMOS, which is what your computer is filled with, does not
like the heat, and the logic runs slower the hotter it gets.

The guy at the fab told me that for our CMOS process,
above 130C die temperature, there could be parametric
shift after 100,000 hours, and that's one of the
definitions of lifetime. And you can see in that case,
where a MIL spec figure would come from. When you push
ICs to 300C, they last around 1000 hours, for borehole
microcontrollers. Let us hope the garage isn't 300C.

   Paul

Back to alt.comp.hardware | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-05-31 19:07 -0500
  Re: Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-05-31 20:31 -0500
    Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-05-31 23:19 -0400
      Re: Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-06-01 15:15 -0500
        Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-01 22:31 -0400
          Re: Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-06-03 17:18 -0500
            Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-03 19:45 -0400
              Re: Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-06-04 09:23 -0500
                Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-04 13:26 -0400
                Re: Cooling a HP desktop sticks <wolverine01@charter.net> - 2026-06-06 10:38 -0500
                Re: Cooling a HP desktop "Carlos E.R." <robin_listas@es.invalid> - 2026-06-06 19:19 +0200
                Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-07 03:09 -0400
                Re: Cooling a HP desktop "Carlos E.R." <robin_listas@es.invalid> - 2026-06-07 14:55 +0200
                Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-07 14:46 -0400
                Re: Cooling a HP desktop "Carlos E.R." <robin_listas@es.invalid> - 2026-06-07 23:18 +0200
                Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-07 18:56 -0400
                Re: Cooling a HP desktop Paul <nospam@needed.invalid> - 2026-06-07 22:09 -0400

csiph-web