On 24/01/12 20:40, Tony Duell wrote:
I've seen HP service manuals that give
signatures for various testpoint=
.=20
They generally have a 'faultfinding'
guide of the form 'If the=20
signatature at pin 3 of U6 is not 23H7 then change U1, U8, U12, U24 in=20
orde'.
I have a DIY spectrum analyser article that a friend photocopied for me.
Under "Aligning the IF1 filter", it reads:
First, obtain a calibrated spectrum analyser with matching calibrated
tracking generator.
I see....
The is rumoured to have ebeen a program for th C64 that was supplied on
floppy disk. The user guide included instructios for converting to
cassette, the list of equipment required to do this included a C64
computer, a C2N cassette unit and a 1541 disk drive.
However the service manual for my Tektronix 555 'scope does explain how
to debug one half of the unit usign the other half (this is a ture
double-beam ;'scome, apart from the CRT and low-votlage PSUs, it's almost
2 'sopes in one box, so you can use one to diagnost faults with the other).
I'd test with the scope (because it's
permanently set up on my bench
anyway) then dig in with a bus-sniffer and the logic analyser.
As would I....
I prefer looking at symptoms and working back to
probable cause. If I
can justify the time taken to set up the probes, I'll rig up the logic
analyser and sniff the CPU bus. The HP 16700's Inverse Assembler feature
really is the mutt's nuts.
I am otld there were inverse assemblers for my 1630, but I have no idea
where you'd get them now. I think they were suppied on little tape
cassetres to go in an HP82161 drive, I do have the drive...)
Total overkill for most things though.
Indeed, 0% of fautls are 'sillies' like bad swithc contacts ;-) But it's
very useful to have the capability to find nasty faults for the reamin
10%....
I guess that's why I'd not use singature analusis much. In a commercial
environment it may well find 90% of faults quickly and let you fix those
units and get them back in the filed. Thing is, I generally only have one
machine (fo a given type) to fix, and I have to fix that one, no matter
waht the fault is.
Some HP manauls are better htna others. In
particular many of the=20
computer nad desktop calculator manuals are pretty pointless. But since=
=20
I've figured out how to fix some of those
without any official technica=
l=20
info, I don't let the fact that they expect
me to use a signature=20
analyser to mean I have to use one :-)
Now why does this statement make me think of the "SX Engine Schematic
Set, A.R. Duell Special Edition" as a prime example? :)
Odd... I was thinking of the pile of diagrms of machines from the HP9100B,
via the 9800s and the 98x5s to the HP9000/200s
-tony