HP-82104A Card Reader Repair Questions

[matty_K]

I am trying to repair an HP-82104A card reader. I’ve successfully replaced the rubber on the feed roller, and the reader now pulls cards through “properly”, although it still isn’t actually working.

Write functions appear to execute, but any read operations (including “VER”) fail with a “CARD ERR” error. I’ve tried adjusting the feed roller axle to increase/decrease pressure on the card without any success. The feed switch is triggering (obviously) and the head switch seems to be triggering (based on continuity). Issuing a CAT 2 command gives me the function list, so the ROM is intact, and the CRC seems to be accepting commands and triggering the other circuits. The write-protect system is working properly (I get a “PROTECTED” error if I try to write to a clipped card).

I’ve tried using a ‘scope to see what’s going on the RA/RB and WA/WB lines, and I do see various pulses. I’ll admit, however, that I don’t really know what I’m doing here—I don’t know how I should set my time base or what I should be seeing during reads or writes to confirm that the sense amp is actually reading signals from the magnetic head.

Does anyone have any helpful suggestions for additional testing? I’ve read through all of the related threads I could find, but I’m still coming up short. Also, any suggestions on how to remove the IO connector cover? I’ve read there is a wedge-type lock mid-way in the housing, and the service manual says to insert a spatula to release the lock so the cover can be slid off, but I don’t see how to do this. Since the calculator connectors are sitting right there, I’m very wary of poking tools inside and bending contacts.

Thanks for any help anyone may be able to provide!

[teenix]

The top of the IO cover has a couple of lugs hanging down from the rear. These sit behind the rear wall of the connector housing. If you gently raise the rear of the top cover from inside where the terminals are, you should be able to raise the lugs high enough so that the cover slides off.

My Classic Notes pdf file has details on the wave forms for the 65 card reader, but they will be the same for the 82104A.

cheers

Tony

[matty_K]

Thank you for the reference to your notes. I had read the portion on repairing the pinch roller, but I hadn’t looked further to see the portions on waveforms.

Regarding the cover (and just to make sure I understand), how does one lift the rear of the cover to clear the lugs without having any surface to lever against? What is your preferred tool/method? One slip and you’ve bent the pins going up to the switch board or damaged the components just under the edge of the cover.

[teenix]

From memory I used a small flat bladed screwdriver (about 4mm wide) and slid it between the pins and levered some upward pressure on the top cover to raise the two lugs over the back of the contact housing which they sit behind.

The two small lugs holding it place overlap the contact housing only by about 0.4mm so you don't have to force it apart too much.

The lug centre points are about 1/4" in from each side of the cover.

see attached picture

cheers

Tont

[AndiGer]

I used some brass "blade" - 0.4mm thick, about 25 mm wide and tossed it inside from the calculator side pressing it upward to try to do no harm to the connectors. With that thickness there is some flexibility of the metal and strong enough to lift the lugs. Maybe a second person holding the reader in place and you try to lift the cover from the rear while push the "blade" inside will be of additional help.

Andi

[matty_K]

Thank you for the reference to your notes. I had read the portion on repairing the pinch roller, but I hadn’t looked further to see the portions on waveforms.

Regarding the cover (and just to make sure I understand), how does one lift the rear of the cover to clear the lugs without having any surface to lever against? What is your preferred tool/method? One slip and you’ve bent the pins going up to the switch board or damaged the components just under the edge of the cover.

[teenix]

(01-13-2023 04:16 AM)matty_K Wrote:  Thank you for the reference to your notes. I had read the portion on repairing the pinch roller, but I hadn’t looked further to see the portions on waveforms.

Regarding the cover (and just to make sure I understand), how does one lift the rear of the cover to clear the lugs without having any surface to lever against? What is your preferred tool/method? One slip and you’ve bent the pins going up to the switch board or damaged the components just under the edge of the cover.

A small enough flat bladed screwdriver sideways, fits between the pins and rests on the bottom of the casing.

Levering it gently downwards forces the end of the blade to push up against the top cover. Two screwdrivers would be better with someone holding the reader and pushing the cover forward.

cheers

Tony

[Hiwi]

Hi Tony,
my experience is, that the speed of the cardreader is not the main problem. It works in a great range.

If you record a card and read it again and it gives card errors there may be other issues, that could be
some tantalum caps or the CRC Chip.

If adjusting the pinch roller will not be successful check the waveforms or ripple of the supply voltage.

Ralf

[Papymouzot]

Hi,

I have tried in the past few days to bring three HP82104A card readers to life again with no luck so far for two of them. I have a similar issue, this is why I have not created a new post.

Thanks to a person I met through eBay I got some silicon tube to solve the gummy wheel issue.

One reader works fine now. I can read and write cards. When I changed the "gummy wheel" I didn't pay attention to the excentric pin.

For the other two readers the cards go through them but one reader gives me a CARD ERR message when I try to read a card, and the second gives me a MALFUNCTION message. I don't really wish to swap parts of the working reader with the other ones as I am afraid to make a mistake on the working reader.

I have not tried yet to adjust the eccentric pin to have a motor consumption of 200mA. I actually didn't find out what voltage I had to apply to the motor to measure the current if I wished to test it out of the reader main body.

I have read in an old archive that it could be a tantalum cap issue, but both readers have a vertical board assy of 82104-60007 and the post mentioned a board with vertical PCB being 82104-60025.

Can somebody help me ?
Thanks in advance

[Hiwi]

Hi,
i always dismantle the reader and apply 3 V to the motor. I disconnect the cable to the PCB.
Then I measure the current. It should be about 200 mA. All from 180 to 240 should be fine.
When a card is inserted.

Before I adjust the idle current to a minimum, the worm gear should rotate smooth, the motor mount has influence on the current. Try to get it to a minimum. I suppose you have all cleaned thoroughly (the head, the roller etc.) and put a very small amount of lube (Mos2) on the gears and bronze bearings. The worm gear has a clutch, which has to be firm.

Then put a card in the slot and measure the current.
The card should be drawn evenly through the slot. If the silicon tube is too soft or not adjusted exactly (centered) the card does not go evenly through the slot. You can hear this.

Adjust the axle to the correct current. It is not a rocket science, the range is wide.

Try some more cards, because they have not all the same thickness.

This is all from the mechanical side. Tony has made a nice guide called the classic notes, you will find some very useful hints there.

I have replaced only very seldom tanatalums in the HP41 card reader. There is a 6.8uF tantalum on the sensor board, which may be faulty. But it is not a general solution for card reader problems.

On the bottom PCB, there is a small potentiometer, this would be my last advice to adjust the speed a little bit, try only with small rotations.

Double check all connections, measure the resistor of the maghead coils.

That is all for today :-)

Ralf

[Hiwi]

CARD ERR and MALFUNCTION message are in most cases mechanical problems with the contacts, may be there is some gum or dirt in the reader. Try to clean the contacts very cautious. They are located under the top pcb over the head mount.

And there is a small hole for the IR diode, clean this as well, since there could be some gum or dirt located.

The other message is CHECKSUM ERR, this may occur when teh speed is deadjusted to much.

Ralf

[Papymouzot]

Thank you Ralf for your replies.

I will try to follow your instructions to adjust the eccentric to get 200mA +/- as soon as I will be back from the slopes . and will post the results.

I will be very happy if I can make these readers work again.

I already cleaned the best I could the plastic parts where the card goes through as well as worm gears and the heads. They were quite clean already actually, I was warned to open the card readers and replace the gummy wheel before inserting a card so the disintegrated wheels gum didn't go all over the place.

I tested with a multimeter the contacts to detect if a card is inserted/out. These work well. In fact we can hear the "click" they make easily. The cards appear to run through the reader smoothly in both ones.

However, in order to test both faulty readers I only had one motor/worm gear assembly working so I moved the motor/gear assembly from one reader to the other to run some tests. One of the readers had its clutch disintegrated and I don't know how to repair this properly. The bronze gear, when inserted without the clutch on the assembly with faulty clutch is off centered and is inserted very tight, is it normal that the bronze gear inserts firmly onto the motor axis or it should be completely free ? I would guess that it should be free

I have tried a few cards and it is the same punition with four cards I tested with. I have no error messages when I write cards on the faulty readers, but they can't be read on the working one.

Regards,
Charles

[Hiwi]

The worm gear:

The wormgear has a small bore into which the motor axle immerses. This should run free.
There is a small piece of plastic or silicon tube that works as a clutch. This plastic is decomposed over the years. A small appropriate cable isolation or the inner part of a RG178 koax cable fits best. Put over this plastic clutch the small aluminium tube and squeeze it a bit together.
The silicon or plastic works as adhesion or grip and the bore centers the gear.

There are other options:

- You can try to squeeze the aluminium tube a bit, may be this makes the clutch stiffer.
This was also done by HP when the card reader was in maintenance.

- Some people have success with a drop of CA, to fix the worm gear to the motor axle.
Use medium ca, so you have enough time to mount the motor in the correct position.
I made this sometimes and it is a good alternative.

What is the correct position?

One end should fit into the bronze bearing without any force.
There should be a bit play or axial movement of the axle.

And the motor mount should be optimized to the lowest idle current, somewhat under 50mA if I remenber correct

Ralf

[Papymouzot]

Thanks Ralf !!! ?

With some luck I have some rg178 coax cable somewhere, if I didn’t trash it ?

I’ll see what can be done if I manage to have at least one reader working.

Is the clutch supposed to be tight ?

Regards,
Charles

[Hiwi]

Every ham radio amateur has some coax on hand :-)

And yes the clutch should be tight.

Tight means that it should not slip under normal conditions, only when blocked. The current must not exceed the specs. This stresses the sense controller.

Ralf

[Hiwi]

Once again:
download the classic-notes.pdf from tony!!!!!!!!!
Find them on google with "classic notes teenix" first search result. This is essential.

Ralf

[Papymouzot]

Hi everybody,

I finally had some time today to run some tests and adjustments as you advised.

Note: In the meantime I received another card reader, a younger one, serial 2548S while the two others were 2052S and 2040S (both non working). The one I managed to make work once the gummy wheel was changed and a bit of cleaning is 2640S.

I opened the latest one I received, removed the gummy wheel which was very sticky, cleaned everything, inserted a bit of silicone tubing I got from an Internet contact (thanks to him) on the black wheel (same tubing in the 4 readers) and tested it... same punition as the two others : CARD ERR. I did apply a very small quantity of teflon grease on the worm gear and on the axis of the little white plastic wheel on both units plus at the end to the worm gear. I did not open the part where the card goes through just in case the pressure applied by the screws would change something.

So... the tests results. The latest one I received: Card reader Serial 2548S

Using regulated lab power supply set to 3V

Running motor with no card inserted : current equal around 80-90mA
When I insert a card, it goes up to 170-180mA
Turning the eccentric pin makes little change unfortunately, I can go up to 100mA with no card and 180-190 with a card, I never reached 200mA
CAT 2 tells me CARD RDR 1G Motor is German
Inserting a card gives CARD ERR.

With the Card reader serial 2052S

Running motor with no cards: current equal 120-140mA, this is much higher than the other one!
When I insert a card it goes up to 240-290mA
Turning the eccentric pin makes very little change
CAT 2 tells me CARD READER
The motor makes at times much more noise than the newer one. On this old one, motor is Swiss, when I touch with my finger the brass gear without applying pressure (motor outside the unit), the noise reduces drastically and current goes to 80-90mA
The motor, removed from the unit consumes 30-40mA when I don't touch the worm gear. I therefore checked the piece (hole ?) where the end of the gear goes, I tried to clean it and reapplied teflon grease. There is a little play along the motor axis so the brass work gear isn't blocked in the "hole/bearing". I don't really understand why the current doubles when the motor is attached to the unit. Once I put the black wheel inside the unit, current goes up again... Although the wheel seems to turn freely, I also applied teflon grease to the inside of the black wheel with no improvement.
Inserting a card gives CARD ERR.


The electronics and motors are different in the two units.

[Papymouzot]

Photo of both readers inside

Left the 2548S, right the 2052S

Sorry, I could not find how to insert a photo



Edit: I ran other tests on the motor units, if I hold the card when it goes through the reader, both motors go to 350mA when stalled, it seems that the clutches are not doing their job...

Edit 2: I "repaired" the "clutch" in between the motor axis and the worm gear on my very first Card reader serial # 2040S. (The fourth one, which was actually the very first I purchased in 1980) The motor turns fine, no strange noise, quite silent actually, the card goes through ok, but this one gives me a Malfunction message.... A shame, I really would like this one to be working, I am sentimentally attached to it.

Your advice on what could be attempted to repair these three card readers would be greatly welcomed !

Charles

[AndiGer]

Charles,
Did you try with several cards?
My first thought was the card itself is the problem when you get CARDERR on (nearly) all readers.
I have some defective card - clearly visible as some magnetic coating is missing
Andi

[Papymouzot]

Hi Andi,

Yes, I have tried different cards. Four to be precise. I tested the cards on the card reader which works and on the three which don’t work no way to write nor read any of them.

Of course I have tried with each reader to read a card from the one which works but also read a card written by the CR I am testing with no luck.

I have taken the electronic out of the case for one of them, looking at how difficult it would be to change the caps, but it is going to be difficult. Components are packed tight on the circuits.

I found diagrams with caps values but two readers have different circuits, the more recent one has a different layout and it also seems that there are less components.