IME that error usually indicates that the USBAsp is working fine, but connections to the target are wrong (and that the problem is not a missing clock source, which gets slightly further into the process)... could there have been a connection problem that was incidentally fixed when switching to the second usbasp?