HP 9845 C with 280 option startup failure

[Tequi Lizer]

>> There is another difference, the 9845C were shipped with colored 
>> functions keys ;-)
>>     
>
> What, all different colours, to indicate the colour that the text (or 
> whatever) will appear on the monitor? The HH Tiger did that.
>
>   

Exactly. Makes the C look a little bite more colorful.

>> Thanks, I'm glad my PSU is still working. I've had enough trouble with 
>> the 9845A PSU, which is only as half as complex, although the functional 
>>     
>
> When you say 'half as somplex', do you mean it only has one main chopper 
> circuit, not the 2 that are used in the 9845B Opt 200?
>
> I have never seen a 9845A, but from what I've heard it's very different 
> to the B or C. To the extent that both processor modules are on one PCB, 
> and there's a bus switching board alongside it.
>   
Yes and no. The mainboards connectors and the output voltages of the PSU 
are the same. Even the mechanical PSU assembly is the same, with the 
difference that it doesn't have LED power-good indicators (there's no 
built-in voltage testing circuit) but dedicated measure points instead 
which can be accessed during operation. The A model PSU is identical to 
that of the low-end B-models, as described in the B/C service manual, 
with one simple internal power supply, one single chopper or 
"pre-regulator", which has a single loop-back from the 5V output for 
pwm-style regulation. The output from the chopper is fed into two 
separated regulators and from there to one assembly each, one for the 
higher voltages, an another for the lower voltages/higher currents. 
Protection circuits exist but are not as elaborated as for the newer 
PSUs. And be aware of those yellowish noise suppression capacitors, one 
of those was in my HP floppy drive and exploded shortly after power-up. 
Did not make severe damage, but I've exchanged them all in the PSUs, 
just in case.

Another difference towards the low-end B models is indeed the combined 
PPU/LPU CPU board, but there are lots of other "enhancements" of the B 
model (all boards but the printer circuits and the I/O backplane have 
been reengineered). The integrated peripherals and the keyboard have not 
changed.

> There are 2 expansion slots in the monitor. One is for the lightpen 
> controller. THe other is for an arbitrary video expansion board. I have 
> neither option, so there are no schematics.
>
>   

The color monitor has just enough space for one additional board, the 
light pen controller. Since I don't have the controller board, I 
currently don't care much about a light pen, although it would be nice 
to see how it works. However I got a working graphics 9111A tablet, 
thats ok for now. Even a hopefully working 8"-floppy. So data transfer 
would be (theoretically) possible as well. When I get one of those 
&#$§*-9845 working.
>> home-brew startup fixure? I guess the only signals which are really 
>> important are the NP and the NL signals. The first resets the address 
>> registers of the alpha circuit and grabs the fist word from the crt 
>> buffer in block 1, the other probably initiates another PPU bus 
>> arbitration for grabbing one complete line of text from the crt buffer 
>> in block 1. Maybe in absence of a crt the alpha circuit makes some kind 
>> of continuous bus capture and the whole system hangs. And maybe the 
>> mainframe can run even without crt when NP and NL are both pulled up to 
>> logical 1. Two resistors would then be enough.
>>
>> Another probably important signal is the HALT-signal, from which the PPU 
>> knows it's the PPU and not the LPU, and which is generated by the alpha 
>> circuit but probably triggered by the NP signal.
>>
>>     
>
> First problem (for you, at least). I drew all this out long before seeing 
> an HP manual, so I didn't use the HP names. And several of the signals 
> disappear into the PPU hybrid so it's almost impossible to work out what 
> they actually do there. I have now seen the HP boardwapper guide for the 
> monitor, but I couldn't easily work out the translation between my names 
> and HP's.
>   

In case I  get a copy of the schematics,  I will try to help. Actually, 
first of all, I'll try to build up a fixture replacement, the 
appropriate signal lines should not be too hard to identify. Then see 
what happens. By the way, if the HP guys were clever, they have placed 
checksums into the ROMs, thats another thing which I'm going to verify.

Today I had a look at HPs website, if you like, try 
http://partsurfer.hp.com/cgi-bin/spi/main, say "9845B" at "by model 
number or name", than hit the search button, click on "DESKTOP COMPUTER" 
and then select a part list for all parts...

-- Ansgar