HP 98203 keyboard "knob" debugging

Given the flamage last time I talked about HP machines here, I am not sure I do want to post anything. On the other hand, I guess I should try to help...

I have a 9816, it's aan interesting little box. But I have the original 'compact' keyboard on mine, I've never seen the one you have, and thus don't have a shcematic of it. The HP service manuals for all HP9000/200 machines (AFAIK) were boardswapper guides with no schematics, although there can be usefukl inforamtion in them (the 9826/36 manaul does have pinouts of all the internal connectors, for example). I don't have the 9816 manual, is it on the web somewhere?

What I can tell you is a little about the 'compact' keybaord, in the hope that yours is related. At least the interface to the host should be much the same. The 'compact' keyboard is stuffed with 4000-series CMOS ICs. Whether HP did the sensible thing (for them, and a pain for us now) and used a microcontroller in the later keybaord I do not know. What does your keybaord contain in ternms of ICs?

The knob is read out as 8 (I think) consecutive 'keys'. 7 bits are number of pulses from the knob, one bit is the direction. There is a misfeature here, in that the counter used is not bidriectionsal. So if you turn the know 3 pulses one way, 4 pulses the other, the counter will say 7 pulses, the direction bit will be the last direction the knob moved. Not the correct reuslt of '-1 pulses moved'. Fortunately the keyboard is scanned sufficiently often that this is not a real problem. I would have expected HP to get it right though!. The actual knob is an HP optical shaft encoder unit. It has 4 conenctions. Ground, power (I think 5V, maybe 12V) and 2 outputs. THese are square waves in quadrature, in the normal fashion. The outputs may be open-collector, if so they will be pulled up on the logic board. I think what I would have done, before pulling the encoder apart, is to stick a 'scope or logic analyser on those signals and turn the knob to see if anything is coming out.

As you have discovered these enocders do come apart. The light source is an IR LED (so if you've put 5V striaght acorss it it's probably dead by now!).

In some cases one of the 'wires' in the housing that fits into a pin socket on the light source assembly is acutally a resistor. The encoder disk comes out by springing the circlip off the spindle (outside the encoder, on the knob end of the fixing bush) and sliding the spindle/disk out backawards. Inside the housign is the detector module with a little slotted metal grating on top. Repairing this is not easy. When a similar enocder in an Olivettie JP101 Sparkjet printer died, I fixed it using the sensor from an optomechanical mouse (such things were easy to find back then...)( and a bit of simple circuitry (I think a 74LS14). I think, though, that opto-mechncial shaft encorders as used in these HP keyboards are still available, if not exactly cheap. if it has failed, it may be simplest to just replace it. But first, does it do anything at all?

-tony