|Procedure to adjust card reader speed in HP65/67/97|
Message #1 Posted by Bernhard on 11 July 2005, 2:13 p.m.
Dear fellow HP aficionados:
Some people owning old HP65/67/97 have reported difficulty
with card reader speed adjustment, especially after mounting
a new friction wheel. As my previous post on the topic did
not yield any useful information on how to do it, I worked
out a procedure that does not need a priori knowledge of
any factory adjustment parameters:
The proposition is that HP orginal cards were written with
in-spec bit density and did not change over the ~30 years.
Just as you would expect from a top-quality manufacturer.
(Hopefully so !)
So if we have a means to compare bit density on an original
card with its copy written on a HP65/67/97, then the motor
speed of the latter can be adjusted to yield the same bit
density as was on the original HP card.
As a means to compare bit density we use another HP65/67/97
(not the one to be adjusted) with good card drive (not worn)
but the speed of this one does not matter (may be a bit off)
as long as it does not change much between reads, which can be
verified by applying simple statistics on the measurements.
Bernies' procedure: put a digital sampling oscilloscope or
a 30ms retriggerable monoflop followed by a digital timer /
counter (I used a HP 5327A, has nice Nixie Tubes) on one of
the two read lines coming from the card reader amplifier chip
of the HP65/67/97 to be used as a measurement tool.
Set up to measure the total time interval from the first read
pulse to the last read pulse. Run original HP card through for
a few times (i.e. 20 times). Note the time intervals. Use a HP
calculator to get mean and standard deviation. In my case, I got
mean = 1110,6 milliseconds with a standard deviation sd = 4.4 ms.
Beware: other cards / calculators may give different results.
3 * sd / mean is a indication of the accuracy you can achieve
with this improvised measurement setup - in this case, about 1.2%
accuracy with nearly full confidence, enough for the purpose.
The mean is the value you want to get from a card written on
the HP to be adjusted. Then you have identical bit density on
the original HP card and its copy.
Now the repeat loop for speed check / adjustment:
Copy the original HP card on the HP to be adjusted. Run the
copy through the measurement HP. Compare the time interval
of the copy with the mean "reference" time interval taken
from the original HP card.
If the copy gives higher readings, bit density is too low,
speed must be increased, and vice versa. After each speed
adjustment, make a new copy to get the new bit density and
repeat the measurement.
Replacing the original speed setting resistor of the HP to
be adjusted with a trimpot / resistor combination is helpful,
but only needed if speed needs be corrected, so don't solder
before you have the first result. Some well - known tricks
as marking the previous settings on the pot or measuring
and noting the resistance before and during adjustment will
help to quickly zero in to the final setting, on which both
cards (original and copy) will yield the same time interval,
and hence, will have the same bit density.
The beauty of this scheme is that no reference other than the
original HP card (standard pac) is needed. The measurement HP
can be adjusted afterwards by interchangig roles. One final
hint: checking the measurement HP with the original card now
and then will give confidence during the adjustment procedure.
If the "reference" time interval starts to deviate, the motor
speed of the measurement HP is not very stable (weak battery?)
but the adjustment procedure can continue with targeting the
new "reference" time interval. Always do the statistics to be
sure to have reliable results.
I have tested the above on three HP67s, and all now work fine.
One of them has a trimpot already installed (original HP ?) and
someone had played with it. This unknown guy has forced me to
think out the above procedure.
Comments / success stories invited !