|
Thanks for rightly pointing this out. And this clearly explains why I am experiencing entirely different thermal behaviour when a same script.bin is tried with our openlec as well as with a regular distro.
But does that mean that the following are not used too? I did't change anything other than the script.bin ddr_freq settings
OpenELEC-opipc3:~ # cat /sys/devices/platform/sunxi-ddrfreq/devfreq/sunxi-ddrfreq/max_freq
480000
OpenELEC-opipc3:~ # cat /sys/devices/platform/sunxi-ddrfreq/devfreq/sunxi-ddrfreq/cur_freq
480000
OpenELEC-opipc3:~ # cat /sys/devices/platform/sunxi-ddrfreq/devfreq/sunxi-ddrfreq/scaling_max_freq
480000
OpenELEC-opipc3:~ # cat /sys/devices/platform/sunxi-ddrfreq/devfreq/sunxi-ddrfreq/scaling_min_freq
408000
I completely agree and accept the method and findings by the linux-sunxi guys about the cpu/memory clock reliabilty. That is very valid for finding a data point considering only the operation of SoC and RAM. But in our case we don't need to push that hard. Our priorities can be slightly different with additional parameters like Low SoC temperature, Lower power consumption, Reliable Board and peripheral operation etc.
Even though the usb ports are wired straight out of the 5v bus I don't think the connect is really that strong and isolated.
With yesterday's test I'm seeing consistently reliable operation during high bitrate play back using WiFi and I don't think it is a placebo effect. Also I have seen the current consumption with just a wifi dongle and usb remote fluctuating north of 1000mA with the default script.bin which is not a good sign.
Also I remember vaguely reading a blog post by olimex guys about their finding about the H3 heating issues in their blog post about the ddr_freq being one of the main contributing factor.
I can't find that exact blog post now, probably deleted.
But few info is there in their first blogpost and it's comments.
https://olimex.wordpress.com/201 ... ototypes-are-ready/
|
|