...

The combination of ATN and EOI comes out on a separate pin on the device.

Example trace

I connected my HP 5385A frequency counter to my HP 9825A calculator and wrote a simple loop to read the ID of the counter:

dev "meas",703
wrt "meas","ID"
red "mes",X
dsp X
gto 1

A similar program for the HP 85A would be:

10 output 703,"ID"
20 enter 703;V
30 disp "type: ";V
40 goto 10

An example trace from the 1st program looks as follows:

...

It starts by sending a COMMAND LAG_03. This makes device 3 the LISTENER (the 5384).

It then sends the ID<CR><LF> string as data to that device. This will make that device want to answer its type number. It now sends the COMMAND LAG_21 making device 21 the listener (the 9825A) and device 3 the TALKER (by sending the TAG_03) command. Device 3 now answers HP5385A<CR><LF>. The final command is UNLISTEN which cleans the listener. The set then repeats.

Problem solving HPDISK communication

I built the HPDISK device (see http://www.dalton.ax/hpdisk/). But it would not work on neither my HP 85A nor my 9825A. It looked like something was wrong with HPIB communications.

I made a HPIB trace of the HP85A command:

cat ":D700"

which looked like this:

...

This repeatedly does a parallel poll. That parallel poll should have been answered by the HPDISK device by pulling bit 0 of the data lines LOW. This does not happen, and that does seem to be at least the first cause.

After the poll it UNLISTENs and UNTALKs; the other commands I do not yet know

More info on HP-IB:

• http://www.interfacebus.com/Design_Connector_GPIB.html

• https://gist.github.com/asanza/4e2719744fb0a41c8828e21dc2b7f012 (hpib commands)