Hi,
Does anybody have an idea why an otherwise working MC68HC908JL8 would consume around 100mA? Of course it heats up, of course it sucks the juice out of the power supply...
I am using it with RC oscillator, the freq is correct.
It happens almost 50% of the 40 pieces I tried by now.
Pulling down RST will stop the over-current, even if the (test) program only has configuration.
Changing the controller USUALLY does not solve the problem.
I use them in another product where they perform wonderfully. There I use 3.3V supply; the failing ones get 5V.
It could also be a PCB layout issue, as the 5V regulator is about 7" away from the controller. The supply traces are wide with no loop; everybody got their own .1uF near the supply pins.
Running out of options/time.
Thanks for the short and long shots...
I found an errata showing that the mask set was screwed up. The OSC2 pin is not used in RC mode, and functions only as output (I needed it as input). I had to force it to LO, otherwise it draws extreme amount of current. The errata does not say it though:
I/O Port Function on OSC2/RCCLK/PTA6/KBI6 Pin SE65-IO_PORT
An anomaly has been found with the OSC2/RCCLK/PTA6/KBI6 pin when:
• RC oscillator option is selected (OSCSEL = 0) and the pin is selected as
PTA6 (PTA6EN = 1).
Under this condition, the PTA6 pin can only function as an output pin;
not an input/output pin as stated in the device data sheet.
Because PTA6 is output only:
• The DDRA6 bit has no effect on PTA6.
• The PTAPUE6 bit must not be set to logic 1 to avoid possible current
drain if DDRA6 is logic 0.
• The KBIE6 bit has no effect, as keyboard interrupt, KBI6, is not
available.
To maintain compatibility with future silicon, the DDRA6 bit should be set to a
logic 1 if PTA6 is used as an output port pin.
Strangely during HW reset, the current was minimal.
I'll have to live without this input.
Now I can flag this as "answered". Thank you.
Just a real long shot, but are you using PLL to bump up your clocks? If so, try a simple test program with and without using PLL and check you consumption. We've just discovered PLL uses 3-4 times the published current numbers (but that's on another micro).