You may want to go back to the latest Non-beta version of the firmware. I don't like to run beta firmware. We have enough proven versions available to us. I run my PWM rates at 12 and I have been doing that for a long time without issues. I don't know if changing that will make any difference. I am not one to change what's been working for me very fast.

I have read somewhere but cannot confirm that Castle says not to use the "Outrunner" selection for our control liners.  I run 12Khz myself v4.02

I have an old log showing 60°C ESC temp when running 4.02 and 8 Khz. Another log from a test run with a yet older firmware (I have not found its version) shows 40°C only under elsewise identical configuration.

Castle says "Outrunner" is for high pole count motors at high RPM and I do not know if my 14-pole AXI2826/12 at 9'000 RPM meets those criteria...

Trying to find the influence of PWR rate / mode on controller temperature and energy consumption I did a couple of static run tests today:

Motor AXI 2826/12
Prop:APC 13 x 5.5 EP
ESC:ICE lite 75
Timer: Hubin FM-9
Throttle-In signal: 1.393 ms
RPM: 9'092 governed high
Run time: 60 seconds

Here is what I have found:

Firmware V 4.19 (Beta),  PWR 8 Khz   
Max. Controller temp: 59 °C  re-charge: 510 mAh

V 4.19 (Beta), PWR 12 Khz 
Max. Controller temp: 62 °C  re-charge: 494 mAh

V 4.19 (Beta),  PWR Outrunner Mode   
Max. Controller temp: 60 °C  re-charge: 526 mAh

V 3.27 logging & log download error, no result

V 3.26 (Beta)   PWR   8 Khz   
Max. Controller temp: 60 °C  re-charge: 500 mAh

As it seems in static run tests, firmware Version and PWR rate / mode remain without significant influence on controller temperature and energy consumption. Could it possibly be that higher controller temperature is caused by continous governed RMP operation at low power? (Throttle-In 1.393 ms only)

Peter,
 I look at your static run tests and I don't see a wide temperature swing when going from 8 khz to 12 khz. Your battery consumption is strange, I would expect it to go higher at 12 khz but your test shows it didn't. I reread Norm Whittles cookbook describing PWM rate and understand it to say the temperature will go up along with eddy current losses(more battery consumed?). I would fly the plane for a minute instead of a static run to get a better idea of what is actually happening. The throttle setting you stated of 1.393ms is slightly above a throttle setting I see on my plane(1.386),the ESC temp read at 140 deg F with outside air of 70 deg F. So in my opinion your throttle setting didn't cause the higher temperature. Lots of good data you have here I will be following this to see the outcome.
Bob
                           
                                  Taken from Norm's Cookbook
“PWM Rate” set to 8 KHz but others have used 12 KHz successfully. 8 KHz is the most efficient
settings for a medium size outrunner. The general rule is; the lower the motor’s inductivity, the
higher the switching frequency should be. Increasing the switching frequency reduces the
ripple in the current flow in part-load mode; but at the same time causes a rise in eddy current
losses in the motor and switching losses in the controller. These losses do create heat in both
the motor and the ESC.

Increasing the switching frequency ... causes a rise in eddy current
losses in the motor and switching losses in the controller.

This is not necessary to be true ... if the PWM rate is high enough, the current in motor does not change too much. PWM does not make alternating current in motor (commutation does), PWM only switchs on and off force which keeps current on wanted average value, current flows via winding also when transistors are closed, because of its induction, and that induction is key for proper PWM setting. The higher the PWN is, the higher are switching loses. But you do not want current completaly stop (at low PWN frequency), because ohmic loses on motor winding are different at constant current and different at current switched on and off, because loses are = I^2 * R so if you have constant current 1A on resistance 1 Ohm you have 1 Watt, but if you have current 10A for 1/10 of time what gives the same average current (the same torque on shaft), loses are 10 ^2 * 1 /10 = 10W ... so that is reason why we want higher PWM frequency - it will make little higher switching loses but much smaller ohmic loses .. up to some optimal value

Eddy current at higher PWM will be higher if current really jumps between 0 and 10A. Twice higher frewquency will make twice higher eddy loses ... but also only to some value of PWM frequency, if coil induction has enough power to keep current almost constant, PWM caused eddy loses will disapear.

So there are 3 effect determining proper PWM frequency. Peter's values shows 8kHz and 12 kHz are close to optimum.

However one note is important here, higer battery voltage will need higher PWM frequency, it is because also ohmic and also eddy loses will be higher because higher voltage will be able to make higher ripple current. It means that if we compare static tests and in flight tests, we can get different results as battery has reatively higher voltage with unloaded motor during flight, si it will be good to make static tests with smaller prop by aproximately 1 inch. If battery voltage is equivalent to voltage necessary for wanted RPM, PWM can even disapear and thus also its loses. That is also answer why even higher motor load (especially lower pitch at higher RPM) can lead to coller ESC.   

Hello Peter,
In this two-run comparison, the higher current that accompanied the slightly lower voltage (in the second run) should have caused more ESC heating.
Perhaps your comparisons have been muddied by the difference between battery packs?
regards,
  Dean P.