HP 9000/200 (9816S) keyboard knob repair

[Martin Hepperle]

Hello,

recently I obtained a HP 9000/216 box with a big and heavy "Nimitz" keyboard. Like the smaller keyboards this has a nifty rotary knob which should control the cursor movement.
While I got the (German) keyboard to work the knob does not work.
After dismantling I found that the small lamp inside has burnt out.
I guess this is a similar setup as on the HP 85 tape drive and other HP desktop machines.
When powering the knob assembly with 5 V I measure 4.4 V across the (broken) lamp.
Before hunting for a replacement I would like to know whether someone has recommendations for this lamp or whether may even has it replaced by a LED-Resistor combination. Not sure whether this works, as the light should be emitted to the side and not to the front. Are there any other specific requirements concerning this lamp?
Currently I am inclined to search for a 6V bulb (3.2mm diameter) with wires as a replacement. Any other comments?

Besides this I am also looking for the cover plate and external PCB with screw terminals for the 98640A A/D converter (98640A-38). My machine came with this A/D converter board but the terminal end is missing.

Thank you,
Martin





Missing terminal block with cover plate ans thumb screws (photo stolen from HP-Museum):

[Martin Hepperle]

... just to answer my question and as a help for others with this problem:

I have now replaced the broken light bulb with a 6V 40 mA miniature light bulb (diameter 3.2mm, length 6.5mm, available in Europe e.g. from Conrad Electronics).
After reasembly the knob works as intended.

Martin

[Paul Berger (Canada)]

I should check that bulb on my Nimitz keyboard since the dial does not work on it. I wonder if an LED would work, maybe an IR LED.

Did you happen to get any software with your 9000? HP produced a software package for the 98640A called "HP98645A Measurement Library" that would make using these cards much easier as there is no support for them built into BASIC or Pascal, which makes using the card more difficult.

For a connector I found that a Cinch 252-22-30-160 fits over the pins securely but the two outside pins, which are redundant grounds are not connected. It appears that there is also a 24 pin version.

[Martin Hepperle]

(06-08-2015 03:00 PM)Paul Berger (Canada) Wrote:  I should check that bulb on my Nimitz keyboard since the dial does not work on it. I wonder if an LED would work, maybe an IR LED.

I also thought of using an LED with a resistor, but then I decided to stick with the bulb. It is important that enough light is shed to the side of the bulb and an LED would concentrate its light more to the front, where it is not needed. The sideways emission is fed into two parallel "channels" to project two thin beams of light onto the rotating disk. The whole system looks very simular to the HP-85 tape drive speed control.
Maybe it works with an LED, but I would not be surprised if the light emission to the side is too weak.
The wheel can be tested in the BASIC editor environment or outside with an oscilloscope.

(06-08-2015 03:00 PM)Paul Berger (Canada) Wrote:  Did you happen to get any software with your 9000? HP produced a software package for the 98640A called "HP98645A Measurement Library" that would make using these cards much easier as there is no support for them built into BASIC or Pascal, which makes using the card more difficult.

Unfortunately I only got a disk with the testing software, which can also be found on the HP Museums web site. I think this software needs the extra test assembly to connect some pins and to supply the right voltage levels.
I have not yet dug deep enough, but I hope that it should be possible to access the card's control registers etc. It would be nice to have the "CSUB Utility for BASIC" software, though, to create the corresponding functions in assembler in the Pascal environment and then make them avaiable to the BASIC environment.

(06-08-2015 03:00 PM)Paul Berger (Canada) Wrote:  For a connector I found that a Cinch 252-22-30-160 fits over the pins securely but the two outside pins, which are redundant grounds are not connected. It appears that there is also a 24 pin version.

Thank you for this tip, I have to see whether I can find this connector here in Europe. However this seems to be a card edge connector - my card has pins...

Martin

[Paul Berger (Canada)]

(06-12-2015 11:14 AM)Martin Hepperle Wrote:  I also thought of using an LED with a resistor, but then I decided to stick with the bulb. It is important that enough light is shed to the side of the bulb and an LED would concentrate its light more to the front, where it is not needed. The sideways emission is fed into two parallel "channels" to project two thin beams of light onto the rotating disk. The whole system looks very simular to the HP-85 tape drive speed control.
Maybe it works with an LED, but I would not be surprised if the light emission to the side is too weak.
The wheel can be tested in the BASIC editor environment or outside with an oscilloscope.

I have never looked inside this particular one but I have a rotary encoder from a 16500A that uses a punched metal disk with two tracks of holes different sizes. It appears to use LED light, but it appears there may be separate emitters for each track. The dial on my small keyboard works and I was tempted to swap the two, but it seems it may be easy to fix too.

(06-12-2015 11:14 AM)Martin Hepperle Wrote:  Unfortunately I only got a disk with the testing software, which can also be found on the HP Museums web site. I think this software needs the extra test assembly to connect some pins and to supply the right voltage levels.
I have not yet dug deep enough, but I hope that it should be possible to access the card's control registers etc. It would be nice to have the "CSUB Utility for BASIC" software, though, to create the corresponding functions in assembler in the Pascal environment and then make them avaiable to the BASIC environment.

I too am hunting for the CSUB software and documentation but it seems to be rare.

The documentation of the card in the manual 98640-90001 is quite good, it has a section on programming that gives you information on how to talk to the card, you can even do it from BASIC using direct I/O commands but it would not be very fast.

(06-12-2015 11:14 AM)Martin Hepperle Wrote:  Thank you for this tip, I have to see wheter I can find this connector here in Europe. However this seems to be a card edge connector - my card has pins...
Martin

Yes Mine has pins too, but even though this connector is designed to be an edge connector, it has the same connector pitch as the pins and fits snugly over the pins.

[Martin Hepperle]

(06-12-2015 03:21 PM)Paul Berger (Canada) Wrote:  The documentation of the card in the manual 98640-90001 is quite good, it has a section on programming that gives you information on how to talk to the card, you can even do it from BASIC using direct I/O commands but it would not be very fast.

Following the instructions in the manual I wrote some BASIC functions to setup and calibrate the board and to perform measurements. All this works very well, but I am not into "high speed" data acquisition (maybe a battery capacity tester program would be my next goal).
I will try and see how fast I can go with the pace timer and pure BASIC.
As far as I understood there is even a way to call assembly byte sequences directly from BASIC without the need to go through Pascal/Assembly/CSUB. But I am still learning.

Martin

[Paul Berger (Canada)]

(06-17-2015 12:28 PM)Martin Hepperle Wrote:  As far as I understood there is even a way to call assembly byte sequences directly from BASIC without the need to go through Pascal/Asembly/CSUB. But I am still learning.

Martin

My understanding of CSUBs is it is just formal documentation of the call linkage between BASIC and assembler / Pascal routines, and provides a library of code to make developing the programs easier. I would be happy if I could just find documentation of how to call and pass data between BASIC and assembler or Pascal routines.

[Martin Hepperle]

(06-17-2015 01:32 PM)Paul Berger (Canada) Wrote:  My understanding of CSUBs is it is just formal documentation of the call linkage between BASIC and assembler / Pascal routines, and provides a library of code to make developing the programs easier. I would be happy if I could just find documentation of how to call and pass data between BASIC and assembler or Pascal routines.

Paul,

as I see, the CSUB utilities are still available for $340 - which is a bit high for my hobby purposes.
So: do we have some example CSUBs on some of the HP-Museums TD0 disk images?
Maybe it would be possible to reverse engineer the interface by analyzing the files.
Then one could write a program to add the the appropriate wrapper around a Pascal/Assembly module.

Martin