HC908QY4A A very peculiar bug

[bigmac]

Hello all,

Background -

I am using a QY4A device, in conjunction with assembly code, to decode an on-off keyed (OOK) signal that is output from a wireless receiver.  This signal is connected to the external IRQ pin.  The decode algorithm makes use of a pair of sub-routines that will exit when a change of state occurs, or after a timeout period has elapsed.  The sub-routines are designed to ignore short noise pulses from the receiver.

The assembly code for one of the sub-routines is attached.  The other sub-routine is similar, except that it handles signal transitions of the opposite polarity.  The attached code operates as expected, and the decoding process is successful.

However, when I attempted to replace the pair of instructions

    DEC  1,SP

    BNE  WL0

with the single instruction

    DBNZ  1,SP,WL0

the decoding process failed.  I tried preceeding the line with a couple of NOPs, so as to retain the same number of cycles, but to no avail.  It is as though the branch is always taken, or never taken - I can't be certain which one.  Full chip simulation gives the expected result.

I wonder if anyone else has experienced a similar problem.  It has certainly involved some head scratching for the last few days - perhaps I am missing something obvious.

Regards,

Mac

[JimDon]

The only difference I see between DEC and DBNZ is that DEC affects the conditions code and DBNZ does not.


You comments says:

; On exit, ZF = 1 if timeout occurred.

So perhaps this is it...

[bigmac]

Hello Tony P and Jim D,

Thank you for pointing out the problem - I think a question of not seeing the wood for the trees.  Yes I do test the Z-flag to determine whether timout had occurred, or not.

Regards,

Mac

[tonyp]

Your code says:

; On exit, ZF = 1 if timeout occurred.

DBNZ instructions do NOT affect the Z flag.  Could this be it?