16F872 funny issue

[lexter]

The funniest part is that now I've replaced that PIC with PIC18F252 and it works nicely with the same board, when powered from external power supply (not through ICD). To my mind, it is so that in my ICD2 implementation, I have a voltage measurement circuit, that is connected to !MCLR pin of the target, which is a simple voltage devider (6.8K and 2.2K = 9KOhm in total, going to GND, obviously), however, my taget board has 10K pull-up at MCLR pin, which means, that instead of getting 5V, !MCLR is getting roughly 2.5V, which doesn't let it run (keeps RESETting?). So, this might be the reason for it not to let me to go to debugging mode. Thinking about it currently ...

PS!

Apparently, I've changed the code as well ...

 

[lexter]

Well, I was correct ... (nice, me knowz smth!!!)

The problem was that voltage divider at target !MCLR pin was interfering with 10K pull-up resistor on the same pin. Thus the issue with entering debug mode, thus the issue with not being able to run the code when programmer is connected.
However, the issue with previous PIC not clocking is still open ... Perhaps bad PIC or smth ...

 

[blueroomelectronics]

Why did you have a voltage divider on mclr?
Microchip has a document on ICD restrictions. Do you have the mandatory 0.1uF bypass caps near the PICs power pins?

 

[lexter]

The idea of voltage divider was to measure the voltage @ !MCLR pin (now, I had to put it to stable 13V point, so that MPLab will not cry about failed tests), apparently (design taken from one of the websites arround). Bypass was there (pins 20 & 19).