6 pin works, 10 pin doesn't

[stkmtd]

I swear to god I've checked the wiring a hundred times. I can successfully program chips with the 6 pin connector. When I try to use the 10 pin connector though, I get the dreaded "rc=-1"

Edit: I just got it working on a breadboard. My problem is that I can't get this to work on a fixture I've created to connect a 10-pin header to the atmega324P. Quick question:

Are you supposed to connect all the GND pins on a 10-pin header together? Would connecting them be the cause of the error I'm getting? I didn't connect the GND pins when I got the 10 pin header to work on the breadboard

[charliex]

All the grounds need to be connected.

[stkmtd]

well, grounds connected or not, it doesn't want to work on anything but a breadboard.

I may just end up posting some pics to show how many times I've gone over the wiring

[stkmtd]

Now I've hit an interesting twist. Pin 30 (AVCC) and 31 (GND) of my atmega324P have 55.5k-ohm of resistance between them. I have a spare atmega324P and this is not the case on it.

Now I could see how this might pose a problem for programming, but it's strange that it will work on the breadboard, and not my devboard (which only shows a short if the chip is plugged in, due to the anomoly with pin 30 and 31)

Anyone ever experienced this before?

[stkmtd]

I tested my homemade dev board with the spare chip I had (without pin 30 and 31 somehow shorted internally), and my dev board works. I was right about the wiring, but I failed to consider that the chip itself might have been shorted (I thought to check when I tested the continuity on the chip socket itself, and realized that the BOARD didn't have any shorts).

But my question still remains. Why would the "bad" chip (pin 30 and 31 shorted) work with the programmer through a breadboard, but not a dev board wired up in an equivalent fashion? The only thing my breadboard has that the dev board doesn't is decoupling caps (but these are only needed for high speed operation, and not for programming).

I am perplexed and curious at the same time.

[adafruit]

you -always- need decoupling caps, speed has absolutely nothing to do with power supply caps

[stkmtd]

you -always- need decoupling caps, speed has absolutely nothing to do with power supply caps

Might be handy to include a note on this page then:
http://www.ladyada.net/learn/avr/programming.html

I only found this surprising because I had programmed chips without decoupling caps in the past. I guess I was lucky. Very well, religiously assimilated into my mind: "ALWAYS DECOUPLE"

When I had decoupling explained to me, I was told you need the caps to smooth out the noise created by rapidly switching ICs, however when I think about it, the default speed of an AVR is 1MHz, which is still pretty blazing fast.