Cyclone III vs USB Blaster Rev. B

I had the same errors with the USB Blaster Rev C but only if I download large code. An hello world will work fine, Got a terasic Blaster and works fine ever since. Due to time constraint, I really cannot investigate what's going on

If it's a custom board, I would expect a JTAG hardware problem, e.g. TCK overshoot/ringing edge, presence of interfering signals, too high resistance TCK pull-down resistor. The different USB Blaster models have different driver hardware and behave partly different.

I do suspect signal integrity problems too. It is a custum board, but not my first with CYCIII

I work for a contract manufacturer, I am programming and testing hundreds of boards, and I use a shell file created by our customer and NIOS shell. The errors that I've seen range from minor failures like a bad A to D reading or the eithernet port not communicating properly to the unit under test behaving as if it hadn't been programmed at all. How it can "sort of" program a unit, I'll never know. Since I've gotten the Rev C USB Blaster, I've programmed exactly 99 units with no errors.

We had the same problem with RevB ByteBlaster. We removed the ribbon cable and replaced with a much shorter one and haven't had an issue since. The same board has no problem with RevC ByteBlaster either.

Yes, boards that have problems with USB Blaster Rev.B and the yellow flex print cable often run better with a short (or even a long) IDC cable. I tried to give an explanation in a previous thread: http://www.alteraforum.com/forum/showpost.php?p=9490&postcount=24