hardware or software?

[alecto]

In a few words : I build an led cube(I know...original ) 3x3x3
I have 9 common cathode columns of 3 leds, each column is switched by NPN (BC337, 10k on base, load on collector, emitter to ground ) and 3 common anode levels of 9 leds, each level is switched by PNP (BC327, 10k on base, load on collector, emitter to 5v)
the logic is provided by a PIC16F690 and me with PickIt2 MPLAB and ASM...(unexperimented experimenting beginner)
The All is alimented by a 9v wall wart regulated to 5v with an LM7805...

The cube worked all right for 3 weeks ...my program was too heavy, badly implemented but it worked !
So last week I wrote a new program, tried it ...and I was a happy programer...it was way lighter...I added software pwm...made some tests, a few animations...and proud of myself decided to turn it off...but 2 minute later I turned it on(yeah just to watch it once more...so sad) and there it started...

Now when I turn it on, it does strange animations...then a few minutes later it seems to "stabilize" and run my program correctly
If I try to modify the program during this "strange phase", the result is different... but still not coherent
but if I wait for it to stabilize, every modification applies immediately and correctly...

What I did so far :

• checked  voltages...seems ok
• checked for bridges or short...none
• replaced the PIc...no change
• added 10k between MCLR an VCC(even if deactivated by software config)...no change
• added 0.1µf between VDD an VSS...guess what...no change
• uploaded my old badly implemented program...still this strange (but slightly different) behaviour
• today on a breadboard I redid the circuit, except I used 2 ULN2803 to drive the 9 columns...and yes...still strange

While writing this, I realized that it must be software...I just don't get why when I reverted to my ancient code, the bug remained...(or appeared??)

hardware or software...maybe neither, maybe just me....

anyway if you have any suggestions...a tired amateur thanks you in advance

[tautech]

I'm guessing HW although it's hard to confirm without a scope.
Thinking it's a power rail issue from the IC coupling state switching transients into the power rail. 

What bulk capacitance value have you used after the PSU?

[jdraughn]

Sounds like you may have an overflow happening in code.

[Chris Mr]

If the running the old code that worked ok doesn't fix it then I would say hardware

Presumably you have resistors in series with the LEDs?

[danadak]

Post the code and a schematic.


Regards, Dana.

[SL4P]

You mention the new code is 'lighter'... does that mean 'brighter'?
Perhaps your new code has increased the 'on time' of the LEDs - pushing the current consumption of the circuit toward the edge...?