On Wed, 1 Jun 2005, Tony Duell wrote:
You would
advocate physically marring a board as "reversible" since you
could "solder a piece of wire over the break", but someone swapping screws
is permanently altering a machine's make-up?
Please tell me you're joking.
No I am not joking...
A PCB cut and repair is (a) generally a useful modification (either to
trace a fault, or to improve the machine in some way, and then to go back
to the original configuration). And (b) it's bleeding obvious what's been
done.
You're telling me that permanently altering a computer in a museum setting
by cutting traces and modifying components is less intrusive than swapping
screws? Please explain how this is the case.
It's not. But then I don't claim to run a museum. I do claim to use
classic computers. A museum should change _nothing_ unless absolutely
necessary to keep the machine operations (I feel that non-working
computers have no interest whatsoever :-)). _I_ should change only that
which is necessary to keep the machines useful.
It is _not_ obvious that the original screws
were, say, Bristol Spline
head if they've been replaced with Pozidriv or whatever.
It is equally non-obvious that the modification was done contemporarily or
prior to the museum taking possession.
True, and I would therefore agree that if a museum got a 5155 where the
screws had already been replaced, they should leave the 'wrong' ones in,
but if possible make a note of the fact that that screws don't appear to
be original (so that some future historian doesn't think this is a
variant of the machine -- yes, that has happened before).
My moan from the start is that there is no good
reason to replace the
original Bristol Spline screws with anything else. Period. And if there's
no good reason to change something, you don't change it.
In this case, there was what I would consider a good reason: Jules didn't
have a Bristol Spline tool on him and wanted to swap out the screws to
something more serviceable for future work. As far as I'm concerned (with
the caveats I have expressed) that's reasonable.
To be honest, if you don't think you'll sometimes have to stop and buy a
special tool, or stop and make a special tool, then you have little
business working on any mechanical or electronic device! Maybe it's a
standard tool (as here, or like that 11/32" spanner I mentioned in
another message) that you don't hapeen to have. Maybe there is a
manufacturers tool -- an alignement jig or something -- that is still
available (and at a sane price), so you buy it. Maybe it's a special tool
that's no longer available, or ridiculously overpriced (I was once
quoted several _thousand pounds_ for a set of feeler gauges, mandrels,
etc that I could make in very little time to an adequate degree of
accuracy). Maybe the tool never was availabe because that section was not
considered to be field-repairable. In that case, you make the tool, or
find somebody who will make the tool.
At the end of this message, I'll stick my notes on repairing the HP9810
card reader. I've done this to my machine. It is no longer original in
that the drive roller hub has been replaced with a home-made one which is
not, therefore, the HP part. But I have no problem with this
modification because the HP part, and the tool to remove/fit it was never
available. Yoy were supposed to replace the complete reader mechanism,
and that is no longer available. It's also a useful modification in that
it means I can now read HP9810 cards. And note that I had to make a
special tool to remove the old roller.
When there is a good reason to change something,
then I see no problem
with changing it, provided you document the changes and attempt to make
sure said documentation is preserved.
I agree. So how do you rationalize the cutting of traces and
modifications of parts to make something work to be acceptable but the
swapping of screws to make something more practical to work on is
unacceptable, given in both cases you fully document the alteration?
It's whether or not the modification is useful. Cutting traces for the
fun of it is not acceptable. Cutting traces to section-isolate a PSU or
bus short, or to make a useful change to the machine is acceptable to me
(as I don't claim to run a museum). Changing screws because you can't be
bothered to buy the right tool is not acceptable. Redrilling and
re-tapping the holes one size larger, and then fitting new screws is
acceptable if the original threads have stripped.
HP9810 reader repairs
----------------------
Take off the drive belt from the LHS.
Then undo the 2 screws and remove the head. This is not essential, but
you'd feel a right idiot if you damaged the head in subsequent work!.
Remove the top left and bottom left card sensor bulbholders from the
plastic frame, leave them hanging on the wires.
Remove the screws from the pillars connecting the 2 sideplates. The front
bottom one on the right hand side, all the others on the left (you have to
frob one of the top ones to get it clear of the pulley).
Separate the plates, and slide the 2 left hand card sensor photodetectors
out of the plastic frame also. Contine separaing the frames and remove the
pinch roller/pressure pad assembly.
Remove the rear cover from the motor (2 tiny screws, you need a nutdriver
a few sizes smaller than most people have!).
Undo the 2 screws holding the wires to the brushgear.
At this point the 2 plates can be completely separated, set the right hand
one (with the cableform, etc) aside.
Remove the motor brushes (one more screw each) to protect them.
Remove the screw holding the lower front pillar to the left plate, remove
it with the card guide.
Rmeove the roller from the shaft using the puller tool. If
necessary, file off any burrs from the end of the shaft. Carefully
slide the roller shaft out of the side plate and recover the metal strip.
Make a new roller as follows :
1" (or so) length of 7/8" (actually 0.87") brass rod. Face off both ends.
Centre-drill one end, then drill 1/8" (or 3.1mm). Check it will fit onto
the roller shaft (but not over the knurled part)
Turn a 0.06" deep groove (so 0.75" dianmeter at base) 0.1" from the end
and 0.25" from the end. Part off a total of 0.35" of the rod.
--------------- <--------- ) 0.1"
) ( <--------- )
| | 0.15"
) ( <--------- ) 0.1"
--------------- <--------- )
Fit 2 O-rings (1" OD, 0.75" ID, 0.125" think) into the grooves. Check
overall diameter of roller (should be 1")
Refit roller shaft through metal strip (DON'T forget this!) and left side
plate. Press new roller onto shaft using a suitable drift and bench vice.
Reassemble the other parts of the reader.
Puller tool to remove HP9810 card reader roller
-----------------------------------------------
2 pieces of steel, 2.5" by 2"
Top plate, 1/4" thick :
4mm holes at (0.25,0.25), (1.25,0.25), (2.25, 0.25), (0.25,1.5), (2.25,1.5)
10.6mm (M12 tapping size) at (1.25,1.5), Tap this hole M12 (1.75mm pitch)
Lower plate, 1/8" thick
4mm holes at (0.25,0.25), (1.25,0.25), (2.25, 0.25), (0.25,1.5), (2.25,1.5)
1/8" hole at (1.25, 1.5). Mill 1/8" slot to nearest long edge (about 0.5"
long).
5 pillars, 2" long, 3/8" diameter steel. Drill 3.5mm and tap M4 in centre
of each end.
Assemble with M4 bolts through 4mm holes in plates to ends of pillars. Fit
M12 bolt to tapped hole in top plate.
Drift to refit roller
---------------------
1.5" length (or thereabouts) of 7/8" brass rod. Face off both ends,
centre-drill, and drill 3.2mm hole. This should be a sliding fit on the
roller spindle.
-tony