I've spent a few hours tinkering with my code and an oscilloscope, and I think I understand my protocol problem! vassilip(a)dsl.cis.upenn.edu wrote: Ahhh, here we have it! The HP 12821A 'hi-speed' HPIB interface board has very different pull-up values! Only 240 ohms for the pull up (as compared to 3.1K) and the pull down values are 680 ohms (compated to 6.2K). Looking at NRFD with the analyzer in slow-mode, I see clean edges and valid levels. NRFD will pulse 'high' for random durations from ~23usec to over 1 ms, depending on what instruction is being sent to the HP1350. NDAC is always about 500ns, positive going pulse. So NRFD is trying to tell the bus how fast to go, it looks like the HP1350 is working correctly, with the analyzer in slow mode. In fast mode the levels and edges change for the worse. NRFD's edges look awful, and it does not pull the signal all that 'low', the pull-up values do indeed appear to be much too low. But is it possible that my HPIB driver code is malfunctioning, and not waiting for NRFD high before starting a transfer? I looked at this, and NRFD goes directly into the 12821A's transmit handshaking logic, this is not the programmers responsibility, nor is there a real-time status signal available to check the state of this line prior to sending a byte. So the code cannot be screwing this up, besides, I can talk to other HPIB devices at full speed, its just the HP1350 that I'm having problems with. NDAC appears to be doing ok, but NRFD is having problems. If I cannot get the programming info for the HP59310B instrument control version of the HPIB controller, I'll try to 'alter' a 12821A by removing the pull-up and pull down networks and replacing them with more reasonable values. The 12821A schematics refer to these resistors as a '7 unit load'. This sounds reasonable, if each device has its own termination resistors. I thought only the system controller had these terminators, so why does it appear each device has a parallel set of termination resistors?