Some personal stuff - and prototypes building!


The issue is not of fitting cooling solution in a smartphone or tablet but industry focus on making devices slimmer with larger touchscreen which leads to sacrifice of ports, keyboards, cooling solutions, audio jack etc.,
That's the actual problem. I'm not into Smartphones or Tablets so I must ask. But from Laptops I know that the Companies also usualy don't focus much onto propper and SILENT (!) cooling. They only build in what's really close to the minimum requirements. Geting a silent and gaming capable Laptop is almost impossible.

@fusion_power You'd only need to fit a heat exchanger and two hoses/pipes. ;)
The Pyra has a modular CPU socket, I guess that's the main issue in our case which makes a "professional" cooling solution (heatpipes) so dfficult.
Exactly, because what most people do on smartphones and tablets is browsing websites. It's important that these load up fast, hence the high-power CPU.
Once they're loaded, the power is not needed anymore.

Thanks to the overpowered SoC, the smartphone feels a lot faster and snappier when navigating or loading / rendering websites.
Of course this makes sense for the average Smartphone User. I asked because I think onto the Pyra, many users may run more demanding programs that require longer and higher max load like games, emulators, compiling, whatever.

Normal PCs can't even, at least with the stock cooler. I updated the cooler on my six year old 65W Pentium server with a modern day 110W cooler and now it mostly doesn't throttle, but it does still momentarily drop the clock to half, according to my script that checks temps and clocks once per second.
Since my Watercooling AIO broke, I use a normal 90mm CPU heatsink/Fan in my Mini-ITX case. Since it's only a 65W i5 CPU, the fan and some case fans (80 and 120 mm) will do the job. Not as good as the AIO but at least I have decent Noctua Fans installed, they do a great job actualy and are dead silent too.

Hm, hard to tell, as this depends on multiple things.
...
There are quite a few factors playing to the whole heat section :)
Wow, Wall'o text, thx for your detailed explanations, ED. I didn't want to steal all you time in writing here so sorry for that. ^^"
 
I think I might be getting lost in translation here. I just mean turning the existing aluminium 90 degrees (a quarter turn) either way so that the ends of the fins sit further from the adjacent plastic stop, allowing more air in and out. No new part needed.
I guess, the length to what he cuts them to fit is shorter than the width. So rotated by 450 degrees they woulnd't. ? They could still be cut in two directions, though.

So it managed to reach thermal equilibrium self radiating to the room without a heat spreader, heat sink or radiator - just a naked SoC to the room? I would not have expected that.
That's what I meant earlier. No matter what's your design, they heat flow goes up with the difference in temperature between yout hottie and the environment.
[doublepost=1555527767,1555527648][/doublepost]
The Pyra has a modular CPU socket, I guess that's the main issue in our case which makes a "professional" cooling solution (heatpipes) so dfficult.
I meant pumping water through. Also, your were asking about tablets. ;)
 
That sounds like a job for a cpu-freq governer to me. You also want the on-demand behaviour there I assume, such that if the CPU is running idle, it drops the clock speed to save battery life, but when it is running it monitors the CPU core temperature so that it runs as fast as it can within thermal limits.

I'm not sure downclocking an idle CPU makes a lot of difference, since (I assume) it will be clock-gated (⇒ basically no dynamic power) when idle (maybe even power-gated). I'm pretty sure SoC CPUs don't run an infinite loop when idle, they just don't run at all. Correct me if I'm wrong.
 
I'm not sure downclocking an idle CPU makes a lot of difference, since (I assume) it will be clock-gated (⇒ basically no dynamic power) when idle (maybe even power-gated). I'm pretty sure SoC CPUs don't run an infinite loop when idle, they just don't run at all. Correct me if I'm wrong.
Well, it's how it worked on the Pandora with its OMAP 3 chip. Maybe assuming the OMAP 5 works the same is a failure on my part, but until shown otherwise I'm assuming it's basically the same with a few headline differences.
 
I'm pretty sure SoC CPUs don't run an infinite loop when idle, they just don't run at all.
The difference is not as significant as you might expect. The wasted energy and therefore generated heat depends on which parts of the pipeline are active, the more complex the pipeline components you currently use are the more transistors will waste energy and generate heat. If you're only executing NOPs, you hardly use anything - you are already really close to not using any energy at all, especially if the microcode of the CPU is able to detect what you're trying to do.

A CPU never stops when idle, it only stops if you actively shut it down, e.g. for a standby mode.
 
I mean, the prototypes I had the chance to play with over the years were alright with the thermals (though, only running at 1 GHz iirc), but reading this reminds me how I really wouldn't mind having a fan on this thing.
It doesn't need to run all times, maybe not even much at all. But it sounds way better than throttling on time constrained tasks (games).

You don't need to tell me how this won't affect me at all. There will be many games and apps that'll unthrottled for sure. And yet we'll likely also have enough stuff hitting the performance ceiling too.
 
While we're straying into the realm of silly humorous cooling solutions...

There are tiny piezo fans that could even work at ultrasonic speeds at really tiny scales.

And fingers might get numb very quickly due to the imperceptible vibration. But it's something that'd fit in that tiny space :D
Sell more replacement parts due to people dropping theirs after their hands go numb :D


Remember the USB power capacitor that made noise? That was acting as a piezo element.

Hey, @EvilDragon let's bring back that capacitor and attach a tiny flap to it! We got our cooling solution. :D
 
Last edited:
We just need to start playing some games, processor will heat --> air will get hot --> air will start moving. If you know this no cooling solution is needed. VOILA :D
 
While we're straying into the realm of silly humorous cooling solutions...

There are tiny piezo fans that could even work at ultrasonic speeds at really tiny scales.

And fingers might get numb very quickly due to the imperceptible vibration. But it's something that'd fit in that tiny space :D
Sell more replacement parts due to people dropping theirs after their hands go numb :D


Remember the USB power capacitor that made noise? That was acting as a piezo element.

Hey, @EvilDragon let's bring back that capacitor and attach a tiny flap to it! We got our cooling solution. :D

This looks rather noisy! For sure a robust solution for really dirty/wet/greasy environments.
 
You'll always hear the airflow itself...
Yeah, from turbulences, but the sound it produces directly - as an oscilating membrane is ultrasonic.
I'm more astonished, that it can get to that amplitute or that the air doesn't restrict the amplitute more at that speed. Does it maybe rather bend in waves, but a single curve?
 
The 500MHz was removed from the kernel.org's vanilla kernel OMAP5 device tree by Ti.
One board member commented out that it may be linked to stability, or something like being in the right tolerances.

---Edit, source:
https://pyra-handheld.com/boards/threads/omap5432-at-home.75874/page-6#post-1341191

Yeah, although I've ran it for a long time on 500Mhz without issues... In theory if we can come up with a proper Frequency to voltage table, we could have a setup like the Pandora and have a full range of frequencies we can set.
 
Last edited:
Back
Top