HP-65 more on the card reader

[teenix]

(01-31-2019 12:30 PM)albertofenini Wrote:  Hi Tony,
look at this picture,
the bottom (usually curved) should be common to all
the inner or shorter (the middle one) is the first activated since is the first encountered by the card and I believe turns on the motor
If this is true, then the upper one (A in my picture) should the one the is triggered by the upper left corner of the card.
What are B and C doing ?
And, how can I check if a recordable card is treated like a record protected ?
Thanks for help !

I'm pretty sure the four switches are, a motor inhibit switch, a motor on/off switch, read/write head activate and write inhibit.
Not sure what the motor inhibit switch does but probably for jammed cards or something.
The 4 switches should go to the CRC chip and WE comes from the ARC chip direct to the CRC chip.

It looks, as you say, the 5th contact is just a common connection for the 4 switches to (probably) ground.

Looking at the switch lengths should give the clue as to the sequence of operation.

My guess,

the switch marked "motor" is motor on/off, C is motor inhibit, or the other way around,

A and B would be the head activate and write inhibit. One of these will activate last if the card corner is cut. Probably A as it is near the card outer edge.

Assuming A is the write protect switch, then if it is not making contact at all, the CRC chip might assume that the card is write protected.

Assuming the common contact is ground, maybe you can connect an ohmmeter to ground and the control pins (one at a time) on the CRC, then you should be able to see what switch is operating if you carefully slide a card in with no power applied.

cheers

Tony

[albertofenini]

I think you are right, A could be the candidate,
However, what fails me is is the following.
Once the upper left corner has gone past the switch i.e. 3 o 4 mm max, there is no difference between a protected card or a non protected card, since the shape is same.
What I'm trying to say, is that the difference between a protected card and non protected card is what ? that switch closes (or open ?) a little before o a little after ?
How can I catch that ?
take care Alberto

[teenix]

(01-31-2019 01:20 PM)albertofenini Wrote:  I think you are right, A could be the candidate,
However, what fails me is is the following.
Once the upper left corner has gone past the switch i.e. 3 o 4 mm max, there is no difference between a protected card or a non protected card, since the shape is same.
What I'm trying to say, is that the difference between a protected card and non protected card is what ? that switch closes (or open ?) a little before o a little after ?
How can I catch that ?
take care Alberto

The switch will activate a little later write protected than it would have not write protected which is how the CRC detects the condition. It would be fiddly, but connecting an ohmmeter as mentioned earlier should detect the switch operation if a card is manually pushed across the activating balls. You should be able to test all 4 CRC pins one at a time while pushing the card in a little at a time.

cheers

Tony

[wallyt]

Regarding viewing the magnetic data on a card, if money were no object one could invest in a can of Magnetic Developer from http://www.magneticdeveloper.com/product/spray-2oz

I think this would render the card useless afterward, and heaven forbid any of it gets into the reader, but it should verify data recording. Didn't HP specify something like this when servicing the card reader on a 97 to verify that track length was within specs?

Many, many years ago the Ampex company produced a similar liquid called EdiVue for viewing magnetic domains on videotape so that it could be manually edited (with scissors and splicing tape!). I had a can of this but forgot where I put it; senior moment strikes again ;0)

[teenix]

(01-31-2019 05:09 PM)wallyt Wrote:  I think this would render the card useless afterward, and heaven forbid any of it gets into the reader, but it should verify data recording. Didn't HP specify something like this when servicing the card reader on a 97 to verify that track length was within specs?

Magnasee
It deposits minute iron particles on the magnetic surface so probably renders the card useless and abrasive.

Non invasive viewer...

https://store.arnoldmagnetics.com/sectio...tic-viewer

cheers

Tony

[AndiGer]

Once in HP Journal May 1974 you had the description of the tiny card reader's operation.
An extract is on our beloved site:
http://www.hpmuseum.org/journals/65crd.htm

The original HP Journal you can be accessed here:
http://www.hpl.hp.com/hpjournal/pdfs/Iss...974-05.pdf

[Dave Frederickson]

(01-31-2019 07:28 PM)teenix Wrote:  

(01-31-2019 05:09 PM)wallyt Wrote:  I think this would render the card useless afterward, and heaven forbid any of it gets into the reader, but it should verify data recording. Didn't HP specify something like this when servicing the card reader on a 97 to verify that track length was within specs?

Magnasee
It deposits minute iron particles on the magnetic surface so probably renders the card useless and abrasive.

Non invasive viewer...

https://store.arnoldmagnetics.com/sectio...tic-viewer

cheers

Tony

I used Magnasee back in the day with digital tape drives.

I don't know if this would work, but for $3.95 it may be worth a try.

http://unitednuclear.com/index.php?main_...cts_id=157

Dave

[teenix]

Hi Alberto,

If the CPU reads the card then the sense amp and heads must be working. It is unlikely that the actual write circuitry in all the controllers you tried are not working.

This seems to point to a mechanical problem. Dirty contacts, bent contacts, something shorted, something open. The motor inhibit, motor on/off and head on/off switches must be ok or it wouldn't read. That just leaves the write inhibit switch circuit.

The leaf switches look a bit tarnished in the picture are the contacts ok?

The HP-97 has switch contact adjustment screws. Does the '65? If so there will probably be test points to measure each switch between ground. The HP-97 service manual says turn the screws CW until the switch just shorts, then 1/4 turn CCW.

HP recommended not to fiddle with the leaf switches too much because if they bend they may not be able to straightened properly causing tension problems. Cleaning a contact was done with an ink eraser.

Have you been able to see any activity on the WA and WB pins with your oscilloscope when trying to write?

Does the calculator go into program mode with the WPGM/RUN switch?

Can you see the WE pin change when this switch position is changed? (Unless the WE logic only activates during a write)

cheers

Tony

[albertofenini]

Hi everyone,
last nigh we spoke with an old friend from the Forum, Ignazio, which added his contribute to all of yours,
and I'd like to thanks everyone and Ignazio for this precious support.
We did some tests on one of the four units, will do more tests on the other over the weekend.

This is a early unit, S/N 1333A20583 where we reinforced the worm gear and used o-rings.
The motor switch works, we tested it, but we kind of knew this since the card was pulled thru.

Tests have been done without the logic board, measuring the switches behavior at the bottom pins.
These tests were done with PRGM switch in PRGM position.
I can confirm that all of the units, including this, go in programming mode,
and we have been able to key in programs and let them run.

We used two multi meters and we observed that :
in the case of a Standard Pac card (corner clipped) HDS closes first, then WPS closes.
in the case of a recordable card (cirner unclipped) WPS closes first, then HDS closes.

Then, when the card is pulled a bit more inside, both switches close.

Nevertheless the unit is not writing
Even when it reads, the program is not working, trying to run it put the machine in endless loops.
Note : we have not checked the listing in the machine read from the card with the original listing, we'll do it very soon

However, only with this unit, something weird happens.
After the card has gone thru (in the picture in read mode, but it happens also in recording mode) the display shows a strange 000 with some missing segments.
Extracting the card from the unit will take the display back to a steady 0.00

I totally agree with Tony, it's unlikely that all of the four controllers are not working, as is unlikely that all of the head, or all of the motors are not working properly.

Tony suggested to use the oscilloscope to see the activity on WA and WB, we have not done it, we will do it soon.
It would be very helpful to know what kind of activity we should expect

If in the meantime anyone has a suggestion will be more than happy to follow it
Thanks everyone for the continuous support !

   
No card

   
Standard Pac card

   
Recordable card

   
Display problem

   
Backplate & Switches

[teenix]

(01-31-2019 07:33 PM)AndiGer Wrote:  The original HP Journal you can be accessed here:
http://www.hpl.hp.com/hpjournal/pdfs/Iss...974-05.pdf

The HP65 card data format is explained in the above article link on page 4.

If you try to write a bunch of NOPs to the card then Track 0 will show a square wave at the bit write frequency, while Track 1 should stay level.

I'd image Track 0 to be the WA pin.

cheers

Tony

[wallyt]

Regarding the "endless loop" behavior.

Like Alberto I, too, am experiencing this when I try to run a pgm from a card that I have just read into the calculator. I do believe, however, that I am able to record "something" on the card but that it gets confused on reading. For example, from a fresh start of the calculator I switch into pgm mode and clear memory with the usual f/PRGM sequence. This leaves pgm memory full of NOP's. I send a card thru and record this. Then I switch back to read mode and power cycle the calculator, which correctly loads the built-in functions (1/x, sqrtx, etc) into pgm memory. I run my card with the NOP's and take another look at pgm memory. The card replaced what was in pgm memory with NOP's. So far so good.

If I write a very short simple pgm (say, LBL A, ENTER 1 + RTN) and record this, upon reading and execution I get the endless looping behavior. Examining pgm memory I see that my short program's instructions are not what I had presumably recorded. They have been replaced with STO and RCL instructions. The RTN instruction that I included at the end of my short test pgm is missing, so the calc simply loops through memory without stopping. It is as though upon reading the calculator is confusing what has been recorded on the card. It does however seem to like only NOP's.

Could this be explained by motor speed or perhaps motor induced vibration? I have rebuilt the dampener clutch and this did not help. I am going to try adjusting motor speed again wih my 10k pot and probably also re-examine the dampener.

[wallyt]

Now I hope I don't jinx what I accomplished by saying this but I think I finally got my HP65 card reader to work. After adjusting the eccentric cam for minimum current drain I replaced the original 5.11k ohm motor speed resistor with a pot and found the sweet spot to be 5.0k ohms. I assume a change was needed here because of the very slight difference in the diameter of the newly replaced pinch wheel from the original gummy wheel. I can read-in and execute per-programmed cards as well as record to blank cards and read/execute those too. I need to find a 5.0k 1% precision resistor to replace the temporary pot. I'm keeping my fingers crossed that this operational functionality lasts.

[teenix]

(02-02-2019 02:50 AM)wallyt Wrote:  Now I hope I don't jinx what I accomplished by saying this but I think I finally got my HP65 card reader to work. After adjusting the eccentric cam for minimum current drain I replaced the original 5.11k ohm motor speed resistor with a pot and found the sweet spot to be 5.0k ohms. I assume a change was needed here because of the very slight difference in the diameter of the newly replaced pinch wheel from the original gummy wheel. I can read-in and execute per-programmed cards as well as record to blank cards and read/execute those too. I need to find a 5.0k 1% precision resistor to replace the temporary pot. I'm keeping my fingers crossed that this operational functionality lasts.

Could be what is wrong with Alberto's too.
I don't know if they make 5K, but 2 x 10K 1% 1/2W metal film resistors in parallel would do.

cheers

Tony

[grsbanks]

4990Ω is a standard value with 1% tolerance if that will do:

http://www.rfcafe.com/references/electri...values.htm

[teenix]

(02-02-2019 09:31 AM)grsbanks Wrote:  4990Ω is a standard value with 1% tolerance if that will do:

http://www.rfcafe.com/references/electri...values.htm

Maybe, it will be +/- 50 ohms. If the reader is that finicky, then match the resistance to 5K with a small resistance in series.

cheers

Tony

[albertofenini]

Here we go with some results !!
Of the four units we were able to fix two of them.
Here is what we found.

S/N 1503S09967
This unit had a broken trace from pin 14 of the Amplifier, that's the motor pin.
I did the repair myself years ago, and while doing this some soldering material had short circuited
with the adjacent trace.
We found out this with the oscilloscope.
RA and RB where good, WA was good while WB was always at +4 volts, since it was in short circuit
with the motor trace.
Once cleaned the short it worked straight away.
Amplifier on this unit is : 1826-0158

S/N 1333A06154
This unit had the HDS switch always closed to ground.
In this case we simply used the multimeter.
The brass switch was too lifted up, we did a small adjustment with the screw and it started to work.
Amplifier on this unit is : 1826-0158

Then the failures ....

S/N 1511A01736
This unit had a pinch roller done with silicone tubing, we replaced this with two orings
It had a 1826-0322 controller
The HDS WPS switchs sequence is good, i.e it knows when a card is protected or not
However, RA & RB are good, it read all the cards at the first try
We have a good WA and a bad WB signal (see attached photo)
We tried then the following :
replaced the Amplifier with 1826-0158 : same result
replaced both 22K capacitors : same result
replaced the magnetic head : same result
replaced the resistor : same result
swapped the logic board with another one : same result ...
What can we test now ???

RA & RB 1511A01736 :    

WA & WB 1511A01736:    

S/N 1333A20538
This unit had the worm gear rebuilt
The HDS WPS switchs sequence is good, i.e it knows when a card is protected or not
It reads the cards 80% of the times
It never writes
However :
RA & RB are good
WA and WB are good
What could it possible be ?

RA & RB 1333A20538 :    

WA & WB 1333A20538 :    

For all of them we have tuned the eccentric cam
For all of them we cleaned the head

Any suggestion will be more than welcome !!!

[AndiGer]

Did you observe on 1511A01736 that as WA changes from low to high or from high to low WB drops? WB seems to be somehow influenced / connected to WA ...

[albertofenini]

Thanks Andy,

that's a good point that we missed, however, we replaced everything on that Card Reader.
Other than the PCB itself and the capacitor close to the Amplifier what could possibly be ?
Remember that we even changed the Amplifier itself, the Head, the resistor, the other capacitors and swapped the logic board.
Thanks for help, take care !!

[EDIT] we obviously checked for all possible bad traces, or short circuits and there are none, it's a very clean PCB

[albertofenini]

Hi everyone,

few more tests that we did on the S/N 1511A01736 :
1. we rebuilt completely from scratch another card reader, everything, a new reinforced motor,
"new" components, "new" motor all coming from some spare material we have.
The result was, however, exactly the same.
2. We swapped the plastic chassis, including battery terminals and brass switches from another HP-65.
The result was, however, exactly the same
3. We mounted a 10K Ohm trimmer instead of the usual resistor (which measure 5,58 K Ohm in this unit).
Playing with this trimmer we were able to change the speed from almost nothing to maximum speed.
In same cases Channel WA (which is normally working) didn't report anything at all, while channel WB was
either "mute" or showed the usual erratic behavior

The only thing we cannot change is the keyboard plate, we have a spare one from our initial restoration project,
but it's missing Anode and Cathode drivers and we don't want to transplant these since it is too risky.

Remember we already swapped back and forth different logic boards without any improvement.

Thanks for any thoughts you may want to share with us !
Take care !

[AndiGer]

Do you know this thread on a weird HP-67 card reader? Writes but doesn't read in any cases - only battery voltage going down.
http://www.hpmuseum.org/cgi-sys/cgiwrap/...?read=9683
Tony Duell suggested to check the capacitor over the VBatt line im post #9. Victor did also not suspicious. And it solved the problem.
Is there something comparable inside the HP-65?
Just some guess ...
Andi