I noticed my system was running a little sluggish and slow after long usage. I put this little cooling fan on this thingiverse mount and it did improve performance.
Then, I overclocked my raspberry pi to 1750 MHz (normal is 700 MHz). The running temp is still at 37-40C.
This is a good idea. We need to think about when v0.4 of the Sangaboard is a hat if we will have even more cooling problems. We may need to funnel air in once we move to v0.4
I really won’t recommend putting any source of vibration on the scope. Even tho the effect, depending on the fan, might be negligible, it just should be avoided out of principle. Raspberry Pi 4 overheats without cooling and adjusts clock, yes, but performance is still better than with Pi 3. I guess for basic operation of the microscope, the performance of the Pi 3 should be sufficient and if it isn’t I would say there is rather a problem with the software.
Depending on the type of operation, a scope might have a lot of uptime and so would the fan. Cheap as fans wear out quick and you get even more vibrations. So you end up with with a part which has to be serviced and in the current housing stack, this takes some time. So what would be a economic solution in that case? Well, I would buy at least the best fan there is, which has minimal vibrations and runs virtually forever without any degradation. Just sets you back 13€.
What else do we get for 13€? Passive cooling for the Raspberry 4! [1]
So if I would run into the problem of overheating I would design a housing with more air convection. The current one is rather tight in that regard. And modify the board mount so it supports an available passive heat sink.
If a version using such a heat sink would become the recommended design version of the Scope, someone may argue that this takes away some openness, or at least, we need a rather propriety part to build it, compared to a standard 40mm fan and a 3D printed mount. However, I guess that’s just the issue of using a non-open board like Raspberry Pi in the first place. There are many applications where someone would use a Raspberry Pi, but where a fan is just not possible. So I suspect that these passive cooling heat sinks for the Pi will get quasi-standard and as available as the Pi itself.
So far I dont experience any motion artifact. My overclocked RPi 4 is working better than ever.
I suppose it depends on the resolution of your optics. We are about to do MTF measurements with a 100x objective. It would be really interesting to see how this was affected by a fan.
And 72 hrs later I do see the vibration. 
it is imperceptible but enough to generate motion Artifact and make the images blurry. it was only $9 for 6 of them.
I think one option for the PCB design for option 4 would be to add pads on the lower side of the board to interface with the heat generating components on the Pi board with vias to conduct heat to the top side of the board. Then a suitable aluminium heat sink could be mounted there, Addiitional cooling airflow could be achieved by a fan external and physically unconnected to the microscope but blowing into the Pi housing. This could be any fan one has to hand.
Regards,
Steve
Just an idea to throw out there. A flexure stage for the fan could negate the vibration going to the rest of the scope. Some diy 3d printers ive seen use them for example to keep movements from the rods to be tranfered to the bed. Might be a bit over engineering though for this purpose.
Hi Steve: You are right. After several tries I have to admit that the fan introduces too much vibration to the microscope. Even though they are imperceptible, the images are blurry due to the motion artifact. I still have my Rpi overclocked and the temperature stays within working range
I think @aurel is right, as a microscopist the idea of bolting a fan to my microscope is horrifying! That said, it does seem to be the case that there’s a consensus that the Pi 4 does better with cooling, and I can’t imagine this is going to be anything other than way worse running it in Tanzania rather than the UK! For sure, passive cooling is a nice option if it can be made to work, though it could get tricky with the Sangaboard v4.
If active cooling turns out to be needed, I think the best solution would be to isolate the microscope from the electronics enclosure - now v7 alpha is out, we have Julian’s redesigned electronics box. Currently the electronics and the microscope are rigidly bolted to the same base, but I reckon the time is ripe for some alternative box designs, e.g. using laser cutting and/or moulded boxes. There are one or two posts on here about that already. It would be a very good idea to have the electronics (especially the Pi, which may need to have cables plugged into it) bolted to the outer box, but have the microscope “floating” inside. If that can be made to work nicely, I suspect adding a fan (particularly if it’s a decent quality one, on a vibration-reducing mount) will then be OK. Ultimately, if you’re running a 100x objective, some vibration isolation is needed even without a fan on the Pi, so this is a problem that it would be nice to solve.
I have semi-deliberately avoided doing too much work on an “official” case beyond the printed “bucket” on the grounds that it’s something that probably wants to be tailored to each particular region/use case/manufacturer, and also that it’s one of the nicest places for others to contribute (because you don’t need to deal with my impenetrable OpenSCAD, though Julian has done a power of work making it more accessible for v7). So, if anyone is keen to put together a vibration-busting, ruggedised carry case, I’ll be excited to see it, and will help where I can! We may also put some effort into this at Bath, but don’t currently have the person-hours available…
