HP Forum Archive 12

[ Return to Index | Top of Index ]

HP9100B Repair Message #1 Posted by Thomas Falk on 18 Apr 2003, 9:27 a.m. Hi everyone, I have a nonfunctional HP9100B calculator here. The symptoms: No display. With the help of the service manual on the museums CD I have found that the analog part is ok. All voltages are there and the deflection circuits seems to be ok. The digital signals starting the horizontal and vertical deflction are missing, but some other signals controlling the character generator are coming from the digital circuits The service manual does stop right at the border to the digital part. The digital circuit works in principle, there are clocks and signals everywhere, but not at the connections to the character generator. Are there any typical errors I could look for or does have somebody some more information on the digital part? Thanks for any hints, Thomas

Re: HP9100B Repair Message #2 Posted by Ellis Easley on 18 Apr 2003, 10:27 a.m., in response to message #1 by Thomas Falk Mine developed an intermittent display at one time. I was able to trace the problem to a transistor on the vertically oriented board on the far right side of the base (not one of the main plug-in boards with ejectors in the card cage). The schematics show a connection directly from the power transformer to the logic - or possibly limited by a resistor in the power supply chassis. The transistor that failed simply squares up the AC signal from the transformer and the squared output signal is what starts every re-paint of the display. The squaring circuit is a common emitter amplifier overdriven by the AC input. The transistor was shorted from collector to emitter by the time I found it. I don't remember exactly what pointed me in that direction or where I heard that the display was synchronized to the line the way it is - possibly it's in the text in the manual? Here you go - on the power supply schematic, it starts at pin 6 of the transformer and goes to 2K resistor "R1" and from there the signal is labeled "60 Hz sync to display J27-K". That's pretty clear! BTW, be careful counting pins on the connectors that use letters to identify the pins - some letters are skipped to avoid ambiguity. The machine had worked fine when I first acquired it and then it was in storage for a number of years. I had just "torn it down" to clean everything and then put it back together and it was working fine. I left it running and went out of the house for a while. When I came back the display was dark and the unit was very warm. I had left it sitting on a sheet of foam that I had used for protection while I was working on it, and had blocked all the air inlets in the base. I turned it off, opened the case and went away to pray for some time. After it cooled off it worked OK but then I found that the display would go blank intermittently and seemed to respond to tapping the case (you've got to try that!). I think I followed the AC signal just because it was something simple that went down into the logic and luckily, found the incorrect behavior of the shorted transistor right away. I replaced the transistor with a 2N3904 or 3906, I don't remember what the polarity of the circuit is but it will be obvious if this is the case with your unit. Since the main logic power supply is -15V, it is probably a PNP - 2N3906. While I had the machine apart I tried to count all the transistors - I stopped at 600! The text for the vertical deflection amplifier says "The cycle starts with ZDRV pulsing negative momentarily." This is not the output of the transistor I replaced. It is DC coupled to the AC signal, it's not a differentiator, its output is a square wave. But if the square wave is missing then there probably won't be a pulse on ZDRV.

Re: HP9100B Repair Message #3 Posted by Tony Duell (UK) on 18 Apr 2003, 7:04 p.m., in response to message #1 by Thomas Falk Jsut about _all_ HP9100 faults result in no display. For the display to come up, the entire digital side must be working. Even a fault in the keyboard or card reader can remove the display!. FWIW, the 9110B will work (assuming no other faults) with the card reader removed, but _not_ with the keyboard unplugged from the logic chassis. So you might want to remove the card reader for the moment, just to eliminate it from the list of possible causes. The bad news is that, AFAIK, there are no official schematics or repair information for the 9100 processor. I've never found such information in a patent, for example (unlike for the HP98x0 machines, where the patents have a lot of useful information). The better news is that I pulled my 9100 to bits ;-), and have been known to give talks about the internals at HPCC meetings, etc. There are a fair number of things you can check inside the 9100 with a 'scope or logic analyser (I would start with the master clock, then the clock divider, then look at the microcode and ROM address registers). If you have a 'scope, then feel free to contact me by e-mail and I'll talk you through the testpoints, etc. One thing to be aware of. The PCBs with red handles contain 20 JK flip-flops each. The top connectos on these boards carry the Q and Q/ signals from each of the flip-flops (and a few extra signals on the corner pins), and are thus useful testpoints. But while the outputs show a voltage change, the J and K inputs to these flip-flops are _current_ operated (they're the base of a common-emitter transistor stage) and can't easily be monitored with normal test equipment.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall