07-02-2014, 02:44 PM
I wanted to share some of my experiences repairing a HP 82104A card reader for the HP 41. I recently acquired this card reader and it definitely had a gummy wheel. I enjoy repairing HP calculators so I was anxious to get started on this project.
The two best "how-to" guides I found are listed below. I read both of them top to bottom because both of them cover some things the other does not. Very useful.
Link 1
http://www.hpmuseum.org/cgi-sys/cgiwrap/...gi?read=57
Link 2
http://www.pmgeiser.ch/calc/index.php?id=HP82104A
Regarding the white roller (Link 2, Pic 11): I found mine to be too snug in the bushings that hold it. The roller would not roll in place if I tried to move it with a pencil. I was able to expand out the plastic bushing area with some tools to allow the roller to spin freely in place. I think this improved card advancement operations. The tiniest amount of white lithium grease (applied with a needle) provides those with OCD some additional comfort.
Regarding the new rubber roller (Link 2, Pic 12): I opted to use the "o-ring" solution rather than the "fuel line" solution that is mentioned elsewhere in the site. However, the dual o-ring shown in this picture may not be the best option. As it is shown, the white roller will be situated directly between the peaks of the o-rings. Ideally, the peak of the white roller and the peak of the o-ring should be aligned. Also, the inner most o-ring is lined up at the very edge of the magnetic card (half grabbing the card, half grabbing the opposite side of the card tunnel). Therefore, I opted to go with a single o-ring that is directly lined up with the white roller and secured to the gear drive with super glue. I used a 1/4"OD,1/8"ID o-ring. However, another link I saw mentioned this is slightly too big OD. They mentioned the 15/64"OD is better. I agree, but I couldn't find those locally. So I carefully turned down the 1/4" o-ring until my calipers read around 0.24". This had the bonus effect of flattening the contact surface of the o-ring to increase the surface area, and it also exposed the softer (more grabby) rubber inside of the ring instead of the polished outer surface.
Regarding the "clutch" or "coupler": I found this link to be useful:
Link 3 (http://www.hpmuseum.org/cgi-sys/cgiwrap/...ead=163396).
My card reader's clutch was definitely deteriorated. (Basically, I'd recommend to anyone who has a gummy wheel to check the "clutch" coupler at the motor shaft. It's probably a 99% chance it's bad also.) So I cleaned out the aluminum cylinder, and also cleaned the motor shaft and brass worm gear. Then I used the "wire insulation" method used in the link above. However, I only used the insulation as a filler for the cylinder and to provide good material to bond with super glue. I agree with some posters that this connection isn't supposed to flex or slip; the motor can be stalled easily. So I went with a permanent solution.
After all that work I'm glad to say the card reader works like a champ. I've run the same card through the machine about 100 times and it didn't mis-feed or jam at all. I've yet to check the write protection feature or a HP67 card (the latter I don't have access to). Until then, I'm enjoying the whopping 200 bytes of data per card. Hope this post helps someone else out there.
The two best "how-to" guides I found are listed below. I read both of them top to bottom because both of them cover some things the other does not. Very useful.
Link 1
http://www.hpmuseum.org/cgi-sys/cgiwrap/...gi?read=57
Link 2
http://www.pmgeiser.ch/calc/index.php?id=HP82104A
Regarding the white roller (Link 2, Pic 11): I found mine to be too snug in the bushings that hold it. The roller would not roll in place if I tried to move it with a pencil. I was able to expand out the plastic bushing area with some tools to allow the roller to spin freely in place. I think this improved card advancement operations. The tiniest amount of white lithium grease (applied with a needle) provides those with OCD some additional comfort.
Regarding the new rubber roller (Link 2, Pic 12): I opted to use the "o-ring" solution rather than the "fuel line" solution that is mentioned elsewhere in the site. However, the dual o-ring shown in this picture may not be the best option. As it is shown, the white roller will be situated directly between the peaks of the o-rings. Ideally, the peak of the white roller and the peak of the o-ring should be aligned. Also, the inner most o-ring is lined up at the very edge of the magnetic card (half grabbing the card, half grabbing the opposite side of the card tunnel). Therefore, I opted to go with a single o-ring that is directly lined up with the white roller and secured to the gear drive with super glue. I used a 1/4"OD,1/8"ID o-ring. However, another link I saw mentioned this is slightly too big OD. They mentioned the 15/64"OD is better. I agree, but I couldn't find those locally. So I carefully turned down the 1/4" o-ring until my calipers read around 0.24". This had the bonus effect of flattening the contact surface of the o-ring to increase the surface area, and it also exposed the softer (more grabby) rubber inside of the ring instead of the polished outer surface.
Regarding the "clutch" or "coupler": I found this link to be useful:
Link 3 (http://www.hpmuseum.org/cgi-sys/cgiwrap/...ead=163396).
My card reader's clutch was definitely deteriorated. (Basically, I'd recommend to anyone who has a gummy wheel to check the "clutch" coupler at the motor shaft. It's probably a 99% chance it's bad also.) So I cleaned out the aluminum cylinder, and also cleaned the motor shaft and brass worm gear. Then I used the "wire insulation" method used in the link above. However, I only used the insulation as a filler for the cylinder and to provide good material to bond with super glue. I agree with some posters that this connection isn't supposed to flex or slip; the motor can be stalled easily. So I went with a permanent solution.
After all that work I'm glad to say the card reader works like a champ. I've run the same card through the machine about 100 times and it didn't mis-feed or jam at all. I've yet to check the write protection feature or a HP67 card (the latter I don't have access to). Until then, I'm enjoying the whopping 200 bytes of data per card. Hope this post helps someone else out there.