FlightAware Discussions

Keeping the RPi cool does make a difference

I use that case, @abcd567. They are really nice and give a lot of thermal smoothing for peaky loads. There were some early versions where the machined steps to thermally contact the chips weren’t quite the right height, but that’s probably been fixed even in the clones by now too. Fantastic case though and not much impact that I can tell on WiFi or Bluetooth.

I just finished up a design and 3D print last night to add the Argon40 fan hat to that heat sink. This print uses that case specifically. It’s the best of both worlds in my opinion.

Link to the .stl files if anyone else wants to print one: https://www.thingiverse.com/thing:4388677


I’ve got the fan and I’ve got it mounted to the adapter but we’ve got high winds today so I’m not lowering the mast. I think it’ll be Monday before I can do this.

1 Like

This looks very good, but limits the use of the expansion slot, unless one uses a ribbon cable. The reason I say this is because I just installed an RTC (real time clock) on one of the Pis. By the looks of it, this case would not ‘allow’ that, unless an extension ribbon cable is used.

Well this is disappointing.

I’ve got the replacement fan mounted to the PoE hat but the clearance between the fan and the top of the box is really small so I’m guessing that it’s hardly able to drag any air through. As a result, my temperature is currently 10°C higher than it was before and that’s with the Noctua fan at full speed compared to the original PoE hat fan at slow speed!


Fortunately, it’s possible to buy different depth lids for the box so I’ll be picking one up on Tuesday and swapping it over.

1 Like

Just mount it so it blows between the two pcbs?

That would involve finding some other method of mounting it which I’m not really keen on. I do electrical stuff, not mechanical!

I’ll just get a deeper lid which will give it clearance it needs.


1 Like

Many many moons ago, I started cooling an AMD processor to do some CPU overclocking (1995 or so). I used a peltier device over the CPU then cooled the peltier with a water cooling rig. good insulation around the cooling parts, but the mother board had frost all around the processor. Then a power bump, and the system had to reboot. Residual heat melted the frost nicely. The now wet mother board, and power reapplied yields nice sparks and the magic smoke came out of the CPU and adjoining parts. Net result: Fried CPU, scorched motherboard, shorted power supply. Experiment terminated, lesson learned.



I’m very sorry for the final outcome, but thanks for posting about it. Isn’t experimenting fun?

Good education costs money, usually. :wink:


I still haven’t found out how to put the magic smoke back in the components to make them work again…

I’ve swapped the case lid today and it doesn’t appear to have made any difference. I don’t understand why this should be - I’ve fitted what appears to be a much better fan in both terms of sound and air throughput but it’s running hotter. The only thing I can think is that perhaps it’s struggling to push the air through the adapter which is effectively tunnelling the airflow. Before anyone asks, I do have the fan mounted the correct way!

Looks like I’m going to have to do something like this. I can fit some mounting posts and attach a fan to those and it will blow across the Pi. It won’t look pretty but it’ll hopefully work.

The new fan being larger might also create more heat from running.
I would have it pull air through the pi (blowing out) not sure if that is how you have it.

The fan with the adapter is blowing onto the Pi, the same as the original horrible noisy one.

I took it down again this morning after I’d put a deeper lid on the box and heath-robinsoned a second 80mm fan into the box blowing across the PoE board and the Pi. I’m not posting a picture, it looks dreadful but it’s secure and it works!

It’s 17°C outside now and the Pi is sitting at 35°C, the only time it’s been lower than that in the last six months is when the outside air temperature was in the low single digits. The forecast for tonight says it’s going to drop to a mild 11°C so it’ll be very interesting to have a look in the morning and see what it went down to in the box.

/edit - It’s dropped down to around 32°C but because of how I’ve configured the small fan, I’d be surprised if it ever goes below 30°C. When it hits 30°C the small fan will go off and come on again at 35°C.

1 Like

i’d never heard of “heath-robinson” before. i just googled…now i understand (thanks for the humor)


I was thinking on this for awhile as I was perplexed as well until I browsed up and looked at the adapter picture again.

If I were a betting man, I’d put some money on the fact that the adapter design is creating enough (stack effect) upwards air pressure, or pressure delta, that the fan is having a tough time overcoming since you’ve effectively created a velocity stack. Got me wondering what would happen if you flipped that fan over if it would move enough air over the board from the underneath outer edges. it would also lend to better air dispersion from around the board instead of being focused more or less directly under the stack…

1 Like

I think that’s just a more technical way of saying what I said :smiley:

I’m not too worried about reversing the fan, it would be difficult to do because the anti vibration mount things are rubber pull throughs that are really single use only and would have to be destroyed to remove. The cooling I have now is perfectly good and I’m happy with it.

1 Like

You can just swap the power leads to the fan. :wink:

Just as much of a faff really, I’d need to take it all apart, cut off the heat shrink, cut the wires, resolder, heat shrink, take it back outside. Nah, unless it overheats, it’s fine.
(yes, I’m a lazy git sometimes)

Unless it is a really low tech fan, the efficiency will drop if the rotation is reversed.


Yeah it was sort of nerdy wasn’t it. Hey… as long as I believe my own BS, all is good. :upside_down_face:

1 Like