hp 82104A only reads 67/97 cards

[John Garza (3665)]

Anyone seen this problem?
I repaired (new drive rollers) 2 of these units today. One went fine, the other won't read 41 cards (MALFUNCTION or CARD ERROR), but reads 67 cards fine (7P via VER). I've tried various axle settings and get good reads on 67 cards on about half the range - but nothing for 41 cards. I've tried cards written on the other card unit, and this one. Although the card write gave no error - I don't recall if it checks that during a write.

Very strange.

[Raymond Del Tondo]

(08-28-2016 05:40 AM)John Garza (3665) Wrote:  Anyone seen this problem?
I repaired (new drive rollers) 2 of these units today. One went fine, the other won't read 41 cards (MALFUNCTION or CARD ERROR), but reads 67 cards fine (7P via VER). I've tried various axle settings and get good reads on 67 cards on about half the range - but nothing for 41 cards. I've tried cards written on the other card unit, and this one. Although the card write gave no error - I don't recall if it checks that during a write.

Did you also clean the magnetic head, and check all the wires?

[John Garza (3665)]

Yes, I cleaned the head and checked that all the wires were still connected.
UPDATE: I forgot to clean the leaf switches, so I disassembled the unit, and gently ran an alcohol qtip over them for a few passes, then reassembled. Now it doesn't work for any cards! The leaf switches looked perfectly clean and I visually checked their movement (while peering between the card reader frame and their circuit board) with a known good example. Also, I did notice some dry, crumbly, old-dead-drive-roller residue down in the lower circuit card area. Looks too difficult to remove; and since it's rubbery, I assume it's a non-conductor. But could I be wrong?

-John

[Raymond Del Tondo]

(08-29-2016 09:23 AM)John Garza (3665) Wrote:  Yes, I cleaned the head and checked that all the wires were still connected.
UPDATE: I forgot to clean the leaf switches, so I disassembled the unit, and gently ran an alcohol qtip over them for a few passes, then reassembled. Now it doesn't work for any cards! The leaf switches looked perfectly clean and I visually checked their movement (while peering between the card reader frame and their circuit board) with a known good example. Also, I did notice some dry, crumbly, old-dead-drive-roller residue down in the lower circuit card area. Looks too difficult to remove; and since it's rubbery, I assume it's a non-conductor. But could I be wrong?

-John

So it may be a bad contact related to the leaf switches. Check their electrical resistance. I had that phenomen on several card readers. In most cases cleaning them helped, but in some cases I had to replace them by others from another card reader. The failing ones all were in older card readers from 1980 to 1982. Sometimes the bad contact was caused from the PCB side. Similar to the above procedure, on some boards cleaning helped, and some had to be replaced. To be sure: With "leaf switches" you mean the two partly gold-plated metal "leafs" which bow up when a card passes one of the underlying plastic levers.

I also don't know if the roller residue is a non-conductor, however I'd try to remove that stuff, if possible at all.

[John Garza (3665)]

Good point Ray - since the leaf switches were touched, the response changed - so that's the likely culprit. And yes, by 'leaf switches', I mean the thin metal strips that 'bow' upwards to make contact with gold pads on the PCB. 'Leaf switches' is what we used to call the similar structures in the HP-67/97 card readers.

So I disassembled again; and this time scrubbed with deoxit. Now I'm reading my HP-67 cards again, and the 41C cards no longer show MALFUNCTION, but instead I get CHECKSUM ERROR. Which, I think is a 'lesser' error if there is such a thing. Cards written by this card reader also receive a CHECKSUM ERROR when loaded on a known good card reader. And cards written with the good reader still show CHECKSUM ERROR when loaded on the faulty unit. And again I tried 360 degrees on the eccentric axle in 45 degree increments with 2 reads each. The midpoint of my good reads on the 67 cards was with the little plastic 'dot' on the axle head pointed 'downwards' (towards the back of the calculator). This is also where I receive CHECKSUM ERROR on 41C cards. When I rotate the axle away from this point by 90 degrees or so either way, I start to get MALFUNCTION or CARD ERROR.

I also salvaged a little crumb of the dried goo and checked it with my HP-3476B meter. The stuff is definitely an insulator.

-John

[John Garza (3665)]

PROBLEM SOLVED

Apparently the replacement I used for the old gummy drive roller was 'out of round'. I replaced it and it worked!

Also apparently, the 67/97 data is likely of lower density on the card than 41C data and therefore more forgiving to variation in card speed. At least that's my best guess.

I use model aircraft fuel-line tubing for these repairs. I've found it superior to o-rings - more grip surface, no glue, looks like the original, etc. I've repaired about a dozen 65/67/97 card readers over the years, but this was my first 2 repairs of the HP82104A. I'm wondering now if this problem was a fluke or if the fuel line tubing has enough normal variation to make it problematic in 82104A repairs - yet 'good enough' for 65/67/97 repairs.

-John