The problem you will forever have till you throw your pegboard away is that you will never know if a problem is your crapboard or your design. Have you considered one simple explanation: chip a from foundry x has slightly different pins than chip b from foundry y and using a crapboard that can make all the difference.

As Jan points out you have no decoupling and, on a pegboard, that will be even more important, the stray capacitances are horrendous, also USE A SUPERVISOR!!!.

One more issue with the crapboard is that if chip b has faster rise/fall times than chip a it is possible that chip a will 'work' and chip b not even if you are lucky enough that all contacts make.

at fast rise/fall times, a wire is NOT 'a wire' it is inductors and capacitors and - for that reason - only a carefully laid out multilayer PCB will guarantee success. Note 'guarantee' some may have made something work by other means but, again, the situation of "is it my design or my board" is the bane of all progress.

One way to get a glimpse of the rise/fall time characteristics of a uC is to look at the max frequency it will run at, the higher it is the faster the fall/rise times will be. This is NOT a perfect way, but very fast to see if the problem is there. It may say "the problem is there" but never "it is not there". The frequency you run the uC at does not matter, only the rise and fall times.

Erik