Hi
I expect the M52235 value with disabled PHY will be close to the M52223.
I did measure the M52235EVB at 40MHz (but always with active PHY, where connecting a LAN cable makes a big difference). Also ran the test using STOP mode support but didn't get great improvements but this was measuring the complete boards.
The results are here:
http://www.utasker.com/docs/uTasker/uTaskerBenchmarks.PDFIn any case the stop mode should be considered sicne it does save current consumption and allows the processor to run without any performance reductions. The uTasker project supports this by adding an optional task which checks for system activity (other tasks scheduled to run or messages waiting to wake then etc.) and if there is presently nothing to actively be perfomred it enters STOP mode.
...
low power task - has just checked that nothing pending and has disabled interrupts
// WAIT or DOZE mode - commanded by STOP instruction (STOP # - writes data to SR and enters stop mode with enabled interrupts)
iInterruptLevel = 0; // mark that no nested interrupts after wake up
asm { stop #0x2000 } // enter STOP mode with interrupt enabled
...
When executing the STOP instruction the processor is frozen but all peripherals work as before. The value 0x2000 is set to the status register when entering STOP mode to enable interrupts again. Any interrupt in the system will immediately wake the processor out of the stop mode and it will continue running from this location without any noticable delays. In a typical network application it can be seen that the processor remains in the STOP mode for maybe >95% of the time. When running with low power support there is no performance decrease since STOP mode is only entered when the processor is idling.
On a Coldfire without Ethernet the current consumption reduction using this method will of course be much more noticable - the PHY is the main mA eater!