PC-IB card hp-150

[brouhaha]

One of the "features" of PC-IB is that the actual values/commands/responses read and written to the registers in the device are not documented anywhere. The only supported interfaces to control PC-IB are their GUI software, and a GW-BASIC library. Want to use C, C++, Pascal, Fortran, assembly language, etc.? You're probably S.O.L.

This is unlike e.g. HP-IB and HP-IL, where the manuals for almost every device documented the commands and responses actually sent over the bus, even if the user ultimately uses some library API rather than crafting the commands and parsing the responses.

They justified this by claiming that it saved having to put more processing power and memory into the devices, as would be needed for a user-facing command set (e.g., the later SCPI language). However, this excuse falls falt because these instruments were designed after the time that microprocessors and memory necessary to parse a command lanugage had gotten cheap. Being generous, that might have saved them almost a dollar in manufacturing cost for an instrument sold for $650.

A general overview of the bus protocol was in the May 1986 issue of HP Journal. The interface carries both a parallel 8-bit bus for devices needing higher speed, and a serial bus for lower speed devices. Each instrument used one or the other, but the host software hid that detail from the user. They made a custom NMOS bus controller chip for the host interface cards (HP-150 and IBM PC bus). The low-speed devices big-banged the serial interface (which was weird, not just normal async serial). The instruments using the parallel interface used custom silicon, which might or might not have been the same chip used on the host adapter.

There might have been depot service manuals, but unlike almsot all other HP T&M equipment of that era, there were no service manuals available to the customer.