That’s right? or that’s wrong? 
You’re using a translator, correct?
Anyway, if my results are correct, it looks like the maximum throttling is around 80%. Which to me sounds quite acceptable.
These SoCs are built for small form factors anyway, throttling is in their “DNA”.
Did the dart board fall during your tests?
It does seem like it did! The truth is the Edge-V decided to rotate the screen upside down, so the last 2 photos are inverted. 
Yes, I wondered about it, as the thing to the right of the monitor is later on its left, also the wire/cables appeared to defy gravity. I for one am glad we got that straightened out.
Does anyone have any good recommendations for a “ubuntu stress tester” app?
The phoronix-test-suite has some very lengthy tests that could be used as a stress tester I assume.
Sorry for my radio silence, been busy with some designing! Anyway @kenny made a new passive heatsink that fits to the upper-side of the SoC, and lo and behold, it works better! (of course you geniuses already knew that)
I’ve only had the chance to really stress out the heatsink on an Edge-V at 30C, will attempt to make more measurements during midday when the temperature is higher.
Edge-V sees about a 10% decrease in throttling with the new heatsink. VIM3L experiences no throttling at 26C. The tester did report “momentary throttling”, but looking at the overall results, it looks like performance is stable.
You can compare the new results with the old results here https://www.khadas.com/post/passive-cooling-a-vim3l. I’m not sure why the latest results are choppier.
tsangyoujun, I was thinking, what if the screws made the transition from the top of the metal cover to the fully metal bottom? 
You mean you’re thinking that the bottom cover should be made of metal too? That could be a little expensive - but technically could be optional for those who want it.
@hyphop had the slightly crazy idea of attaching the passive heatsink to both sides the of the case, and then letting the case rest vertically on the table - it would look like a porcupine.
maybe he hiphop correctly says if it concerns Ubuntu 
Well, maybe a flat metal plate from the bottom?
I would love a full metal VIM3 case. Probably need to rework the design to transfer heat from the lid to the base to go in line with the vim diy case, metal plate and thermal pad, following the idea of transferring heat not just from the radiator but surface it sits on.
then you have to bring the antennas out, in this case.
Metal Top/Heatsink and Base with thermal pad
The Middle would be as it is
Just as much heat can be found on the base,
Make a Metal Vim sandwich if you need more cooling attach fan externally on top or base
Small fins on the base enough to lock a screw thread if a fan is attached maybe same on top
It will work Very Well.
This is an old post from 2 years ago, this guy made a full metal case for VIM2 using CNC, I hope he could make a new version for VIM3, I’d order it in a heartbeat:
True a good solution would be to include two u.fl to sma connectors with antennas and have two holes in the side to mount them much like premium PC motherboards with a wifi slot.
Yeah that was what I was thinking of in terms of wifi\BT solution. though would would like a more sleek version that looked more like it was designed for the vim3 rather than looking like rectangles cut out of the back of a aluminum bar. I know it seems petty, and that nobody will likely see it but that’s just me 2 cents in terms of design choices. Also might want to thin out and shape the rear plate to fit around the IO port as to avoid issues like the VIM DIY case preventing good connections with the power cable.
Love the anti-vibration binding.
I dub this one, The Pragmatist’s Cooler.
Worth emulation.
Just sharing some hard behind-the-scenes work, sweating.
I’ve been running a whole series of CFD simulations in an attempt to determine the optimal geometry, really hard work! I should learn how to automate a solution for this - future work.
Anyway, I found that the uglier solution works better. Would you guys prefer:
A) Completely functional, but awesome cooling, ugly
B) More beautiful, less awesome cooling, pretty
C) Try to make the ugly pretty, somehow, would require copious amounts of imagination
CFD was run at 35C ambient, 7W TDP RK3399.
I’ll also have to reconfirm the cost with @kenny. Cost would be a factor too.
Hello! will depend on the scenarios, everyone has a different scenario 
