Is there a standard CPU fan that will work on the x15? When I operate both A15 cores over 50%, the processor get hot (147 degrees F).
Regards,
John
This one should fit:
https://www.digikey.com/product-search/en?keywords=%20COPAL%20F251R-05LB
You can connect it to the j1 connector, haven't worked out the
connector.. (it should pwm the fan based on load)
i'm just using a personal fan across the board..
Regards,
Thanks Robert. Seems pretty quiet so that should work great.
BTW, lshw shows CPU1 as disabled:
*-cpu:0
description: CPU
product: cpu
physical id: 1
bus info: cpu@0
size: 1500MHz
capacity: 1500MHz
capabilities: cpufreq
*-cpu:1 DISABLED
description: CPU
product: cpu
physical id: 2
bus info: cpu@1
size: 1500MHz
capacity: 1500MHz
capabilities: cpufreq
That is strand because htop shows both A15 cores working.
Regards,
John
Yeah, that's lshw, don't worry about it. 
a6-imx6q-wandboard-2gb
description: Computer
product: Wandboard i.MX6 Quad Board
width: 32 bits
*-core
description: Motherboard
physical id: 0
capabilities: wand_imx6q-wandboard fsl_imx6q
*-cpu:0
description: CPU
product: cpu
physical id: 0
bus info: cpu@0
size: 996MHz
capacity: 996MHz
capabilities: cpufreq
*-cpu:1 DISABLED
description: CPU
product: cpu
physical id: 1
bus info: cpu@1
size: 996MHz
capacity: 996MHz
capabilities: cpufreq
*-cpu:2 DISABLED
description: CPU
product: cpu
physical id: 2
bus info: cpu@2
size: 996MHz
capacity: 996MHz
capabilities: cpufreq
*-cpu:3 DISABLED
description: CPU
product: cpu
physical id: 3
bus info: cpu@3
size: 996MHz
capacity: 996MHz
capabilities: cpufreq
It doesn't know what to do..
Regards,
Is 145°F (64°C) considered hot for this CPU?
Regards,
Leonid.
Not by TI. And they designed it.
Gerald
Yes, 64C is toasty, but normal. The AM572x processors are high-performance CPUs and will run considerably warmer than our previous ARM Cortex A8 and A9 devices. To achieve high performance, certain tradeoffs must be made in the design and silicon process, which unfortunately translates into higher heat.
Depending on your application you may face thermal issues. Provisions have been made on the board to attach a 5V fan. TI suggests using Sunon Fans MC25100V1-000U-A99 (http://www.digikey.com/product-detail/en/MC25100V1-000U-A99/259-1557-ND/2757805) or equivalent with this board. For silent operation, a larger heatsink can also be used.
TI maintains a wiki with thermal characteristics of the AM572x processor: http://processors.wiki.ti.com/index.php/AM57x_Thermal_Considerations. The AM572x thermal considerations wiki will be updated in the near future to include our test data with, and without a fan, under high CPU loading usecases.
If using the X15 as an evaluation platform for a new product, it is imperative to thermally profile your system, and design for adequate heat dissipation, whether passive and/ or active cooling devices are chosen.
For comparison, thermal data for the AM335x used on the beaglebone family can be found here: http://processors.wiki.ti.com/index.php/AM335x_Thermal_Considerations
Hi Robert,
I installed a CPU fan, but it seems that the fan is either on or off. No PWM. Also, the bit-bank control lacks hysteresis. The fan is on for a few seconds and then switches off for a second or two and then turns back on again. Sometimes the fan turns on/off for 1/2 second and then off/on again.
Regards,
John
That is the way it is designed. There is no speed control, just on or off based on the SW set point that determines when it comes on and when it turns off…
Gerald
Hi Gerald,
That’s fine, but then we need to use some hysteresis because these fans won’t like switching on/off 36 times per minute. Sometimes the on/off/on or off/on/off transition is less than 400mS.
Regards,
John
That is up to the software.
I suggest a trip point to turn on and drive it several degrees past that point before turning off.
Gerald
Agreed. Robert, do you know who worked on this feature? Perhaps they can fix this or when I have some time, I’ll take a stab at fixing this.
Regards,
John
Is it even designed to work with a fan in "normal" circumstances?
Which temperature range should be considered too hot?
I think it's one of those things we enabled, but never really tested..
There is a tmp102 sensor near by we just need to tweak the settings.
Regards,
oh btw, with the fcc testing issue, we have time to fix this and roll
out a newer default image .
Regards,
The TMP102 is the ambient sensor. The one I believe that is used for the fan is the on chip temp sensor.
Gerald
Agreed. Robert, do you know who worked on this feature? Perhaps they can fix
this or when I have some time, I’ll take a stab at fixing this.
The culprit who had worked on this is me.
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=d723cfeafc7b4c73e89ed3d4b1a4d747e990872c
temperatures selected were empirical ofcourse. 
I do invite folks to see if they can improve this and post fixes
upstream - I can easily pick them back up.
in general, the idea was to tweak them as needed later on..
I think it's one of those things we enabled, but never really tested..
hmmm... this is probably one of the first set of platforms that
actually had ability for active cooling... so yep, tweaking will
probably be needed.
There is a tmp102 sensor near by we just need to tweak the settings.
actually the fan trigger is based on two things: on-SoC temp sensors
and tmp102.
Looking at the patch, there is no way temperature is changing 2C in fractions of seconds. Something else is making this fan switch on/off/on/off that quickly. Seems to be hovering about 46C measured with a Fluke IR thermometer.
Regards,
John
OK, I put an oscilloscope on the CPU fan 5V and I see on/off timing as follows:
off 501ms
on 2497ms
off 1381ms
on 1497ms
off 119ms
on 499ms
off 1133ms
Seems these times don’t square with the poll timing.
Regards,
John
How about on the enable pin?
Gerlad