HP-IL Printer (odd?) page shifting when disabled.

[Diego Diaz]

Hi all,

I've been asked about why the HP-IL Printer ROM is shifted from page #6 to page #4 when it is Disabled by the bottom switch.

First of all, I must admit I don't have a documented answer; so I provide my best guess in the hope that someone else may disprove or confirm this hypothesis.

Here are my thoughts (at risk of showing my ignorance on the subject :-)

Since the ROMs in the HP modules are built using CMOS technology and they are connected to the CPU by a serial bus composed of clocks, sync and data/address lines; disconnecting a ROM chip from the CPU cannot be achieved by simply removing the Vcc line (and/or GND for that matter).

As long as any line is high and any other line is low the ROM will receive enough power to keep on running and responding to the CPU.

So, the only way to effectively remove this ROM from the bus will require all (or most of) the bus lines to be disconnected. A multipole switch suitable for the task would hardly fit in the scarce real estate on the HP-IL shell.

The (clever IMHO*) trick of switching just the ROM's address mask bit A13 from 1 to 0 (changing its native #6 to #4) solves the problem since page #4 is not (was not until Ángel's LIBRARY4) used by the "normal" HP-41 functions.

*Careful selection of the XROM and FN words allows the ROM to return control to the OS as soon as it's called after wake-up.

As stated, just an educated guess.

Any input will be very welcome.

All the best from the Caribbean.

Diego.

[J-F Garnier]

This is a good question that always intrigued me.

Since the page shift mechanism needs a signal (and a chip pad), wasn't it easier to just use this signal to disable the ROM, instead of moving it to an other address?
As you reminded it, the present solution requires a careful design of the first few ROM words to give up control, since address 4000 is polled at each wakeup from sleep states.

If you look at the "40K ROM" specification at the well-known HP-41 site, you will see that there is a TEST signal to disable the ROM. So IMHO, it would have been better to use it.
But maybe there was a drawback or a potential risk to put the ROM in test mode, so the engineers had to resort to the address selection signals.

Or maybe was the disable feature considered quite lately, something like: "Oh wait, we forgot it, what if a user wants to use his old 82143 printer with the HP-IL module?" and the engineers chose the safest way.

If so I would not call it a clever trick but a workaround.

J-F

[Diego Diaz]

Hi J-F,

Sorry, I may have misled you by (wrongly) referring to the A9 line. It is A13 (now edited) the line that is switched.

I missed a whole nibble when counting from the top of my head... this wouldn't have happened 20 years ago. I'm not getting younger.

A13, and also A14, already have their own pads on the 40K ROM die so no need to further design, added pad or new chip mask.

On a Standard ROM A15 is masked to 1 to trigger the CS/ signal. While A14 and A13 are port dependent and connected to B4 & B3 lines, respectively. B4 & B3 lines were named after the original 16 register D/S chip pads, as they served essentially the same function for the 82106A modules.

On HW addressed ROM's (TIMER, PRINTER, HP-IL, DIAG...) A15 is masked to 0, and A14-13 wired to match desired for that specific ROM.

I didn't think of the TEST pad though, good point. It would indeed be a reasonable alternative.

Again, maybe we will never know... :-)

Cheers

Diego.