HP-65 goes on hold after reading a card

[albertofenini]

Hi Tony

Am I wrong or the HP-41 card reader uses the same Sense Amplifier 1826-0322 used by the HP-67 and thus compatible with the HP65 ?

Take care !

[teenix]

(11-17-2020 04:08 PM)albertofenini Wrote:  Hi Tony

Am I wrong or the HP-41 card reader uses the same Sense Amplifier 1826-0322 used by the HP-67 and thus compatible with the HP65 ?

Take care !

It looks like it but I haven't swapped the 65 and 67 sense chips. They have different part numbers but pinouts look the same/

cheers

Tony

[albertofenini]

Hi Tony
We did it in the past and it worked out so they are compatible

However, we did the following right now :
Replaced the HP65 sense amplifier with one from a working card reader for HP41
The test sorted out the same results, WA good WB dirty
We then excluded the card reader on the HP65 de-soldering WA WB pins where the card reader interconnect the logic board
Same as before WA good WB dirty
Looked like we found the culprit in the CRC right?

Well we put this logic board in another unit and while the display fault was consistent (all g segments on) the read and write operation was perfect

This would exclude the hybrid ....

We tried the opposite, we put a known working logic board in the faulty one and the display was acting crazy with random zero’s appearing

Since it’s also Tuesday 17th ... we immediately took the good logic board out

In summary if this board in an other Unit works, and in this even isolating WA WB from the card it doesn’t... where could it be the trick ?

I would not say traces on the logic because if this was the case it would have been consistent across different units, the sense amplifier has been changed twice and this one is known to be working ...

Card reader pcb has been tested and it is very clean ...

What are we missing ?

Thanks and take care !!!

[teenix]

The original hybrid obviously has a problem with the display.
The original keyboard also seems to have a problem causing display issues.

Maybe while operating together, they are causing too many issues for the CRC to operate properly.

Are the power supply voltages ok when both faulty boards are used together, no increase in supply noise?

Do any ICs or hybrid get warm to the touch after being operated for awhile.

Is the wiring between the boards frayed at the connection and maybe should be re-soldered?

If you are using a mains power supply to power the calculator, sometimes connecting a mains operated oscilloscope causes earth loops and upsets the operation.

cheers

Tony

(11-17-2020 10:15 PM)albertofenini Wrote:  Hi Tony
We did it in the past and it worked out so they are compatible

However, we did the following right now :
Replaced the HP65 sense amplifier with one from a working card reader for HP41
The test sorted out the same results, WA good WB dirty
We then excluded the card reader on the HP65 de-soldering WA WB pins where the card reader interconnect the logic board
Same as before WA good WB dirty
Looked like we found the culprit in the CRC right?

Well we put this logic board in another unit and while the display fault was consistent (all g segments on) the read and write operation was perfect

This would exclude the hybrid ....

We tried the opposite, we put a known working logic board in the faulty one and the display was acting crazy with random zero’s appearing

Since it’s also Tuesday 17th ... we immediately took the good logic board out

In summary if this board in an other Unit works, and in this even isolating WA WB from the card it doesn’t... where could it be the trick ?

I would not say traces on the logic because if this was the case it would have been consistent across different units, the sense amplifier has been changed twice and this one is known to be working ...

Card reader pcb has been tested and it is very clean ...

What are we missing ?

Thanks and take care !!!

[albertofenini]

Hi Tony,

it's a behavior that few other HP-65s have shown in the past as well
Normally we have fixed this by swapping and mixing parts until they were working
Like it was a matter of finding a good match between the three main components,
the keyboard/display, the card reader and the logic

Certainly the keyboard/led plays a role in here, and we need to find out what it is
On the other side, on the keyboard/led there are only two active components, anode and cathode driver,
the two inductances for the led and the LC components ... not much to test ...

We where thinking to buy one of those portable microscope to look for bad traces, will see

There are not ICs getting warm and we are powering the units both via a rebuilt battery pack or via an external battery pack with a switch,
however the behavior seems not related to the kind of power source

We'll do more checks later today and we will let you knew
Thanks for the huge support !

[albertofenini]

Hi Tony, Hi Everyone
we did a couple of tests, using the Biorithm Program which is 90 steps long

It reads ok from a card but if we examine the listing, after few steps a - minus appears at the right of the instruction code.
Usually this would mean that the end of the 100 steps memory has been reached, obviously this is not true.

The minus appears in multiple positions like at the 12th step, then at the 15th step, then in more positions but not in all the steps after the first appearance.

The unit goes on hold even just pressing RTN in the RUN mode after all the steps have been examined

We tried reading from different cards with the same result and if these cards
are used with a working unit everything runs ok.

If the program is inserted with the keyboard, everything goes well until about the step 70th,
after which more steps can still be keyed in but all will have the minus at the right of the instruction code.

Looks like not all the memory is working and can be used

Is there any possible fix ?
Is the problem in the Hybrid or somewhere else ?

Thanks for any possible hint !
Take care !

(11-18-2020 09:34 AM)albertofenini Wrote:  Hi Tony,

it's a behavior that few other HP-65s have shown in the past as well
Normally we have fixed this by swapping and mixing parts until they were working
Like it was a matter of finding a good match between the three main components,
the keyboard/display, the card reader and the logic

Certainly the keyboard/led plays a role in here, and we need to find out what it is
On the other side, on the keyboard/led there are only two active components, anode and cathode driver,
the two inductances for the led and the LC components ... not much to test ...

We where thinking to buy one of those portable microscope to look for bad traces, will see

There are not ICs getting warm and we are powering the units both via a rebuilt battery pack or via an external battery pack with a switch,
however the behavior seems not related to the kind of power source

We'll do more checks later today and we will let you knew
Thanks for the huge support !

[teenix]

I can only go by the block circuit diagram for the HP-65 which has a separate storage circuit for the programs, which is the 600 bit (plus program and return pointers) carousal buffer.

This is connected to the CRC chip via 5 lines for control, timing and data.

It is also connected to the CTC chip, which probably controls access to the program steps and pointers. This may also have a problem.

It would seem that there is a problem within this arrangement, either the storage/retrieval logic or the buffer itself. If this is the case, and as there is no easy way of changing these chips or checking the internal connections, then I would guess that changing the hybrid may fix the problem.

cheers

Tony

(11-28-2020 08:34 PM)albertofenini Wrote:  Hi Tony, Hi Everyone
we did a couple of tests, using the Biorithm Program which is 90 steps long

It reads ok from a card but if we examine the listing, after few steps a - minus appears at the right of the instruction code.
Usually this would mean that the end of the 100 steps memory has been reached, obviously this is not true.

The minus appears in multiple positions like at the 12th step, then at the 15th step, then in more positions but not in all the steps after the first appearance.

The unit goes on hold even just pressing RTN in the RUN mode after all the steps have been examined

We tried reading from different cards with the same result and if these cards
are used with a working unit everything runs ok.

If the program is inserted with the keyboard, everything goes well until about the step 70th,
after which more steps can still be keyed in but all will have the minus at the right of the instruction code.

Looks like not all the memory is working and can be used

Is there any possible fix ?
Is the problem in the Hybrid or somewhere else ?

Thanks for any possible hint !
Take care !

[albertofenini]

Hi Tony
Too bad we don’t have one ...
I guess only way would be sourcing it from a broken unit ...
Could it be a problem of bad connections between the hybrid pins and the hybrid internals ?
Take care !!

[teenix]

(11-28-2020 09:59 PM)albertofenini Wrote:  Hi Tony
Too bad we don’t have one ...
I guess only way would be sourcing it from a broken unit ...
Could it be a problem of bad connections between the hybrid pins and the hybrid internals ?
Take care !!

Hi Alberto,

I'm not sure, but I suspect it is more a chip problem internal to the hybrid, especially since normal key programming has issues as well as the card read/write. All other functions seem to work so the internal timing etc and associated connections would seem ok.

An internal connection issue maybe the cause of the broken WB signal but this maybe a symptom rather than the root of the problem. Any sort of surgery to the hybrid internals would be highly problematic and maybe impossible even with specialist equipment.

cheers

Tony

[[kby]]

Were you told where the wires were broken? They usually just break where they’re soldered in which case that is not too difficult to fix. I probably have some heads that could be sacrificed; they aren’t thst uncommon as they are shared with both the -67 and -97(s).

(11-15-2020 05:38 AM)Kostas Kritsilas Wrote:  I have a HP-65 in which the card reader cycles cards through, but doesn't read them. The card goes through, and the display shows 0.00, even though the calculator is otherwise fine. I sent it off for repair, and it appears that the hair thin wires going to the read/write head are broken. I'm just going to use it as a keystroke programmable, since the supply of HP-65 card reader parts is not exactly abundant.

[Kostas Kritsilas]

I was told it was probably the read/write head or the wiring to it was at fault when I sent it out to have it repaired. The motor comes on, and pulls the card through well, it just doesn't read anything. All the keyboard functions work, and I can type in and run programs. Even though the card reader is pulling in the card, the display does not show any data has been read.

If you want to sell a spare read/write head, please let me know what you want for one. If you are willing to attempt a repair, please let me know what you would like to charge for that. I would very much like to have the calculator fully working, but cannot, due to vision issues, do the work myself.

[teenix]

(11-29-2020 06:50 AM)Kostas Kritsilas Wrote:  I was told it was probably the read/write head or the wiring to it was at fault when I sent it out to have it repaired. The motor comes on, and pulls the card through well, it just doesn't read anything. All the keyboard functions work, and I can type in and run programs. Even though the card reader is pulling in the card, the display does not show any data has been read.

If you want to sell a spare read/write head, please let me know what you want for one. If you are willing to attempt a repair, please let me know what you would like to charge for that. I would very much like to have the calculator fully working, but cannot, due to vision issues, do the work myself.

You didn't mention card writing, but if it is the coils inside the head, (which would not be repairable), then I would assume the card cannot be written too as well, as these perform both read and write functions. The same for the external wiring to the sense board. If one is broken, then a read or a write will not succeed.

If it is the external wires, I'd be pretty sure like [kby] says, that the wires will most likely be broken where soldered into the circuit board, which is due to moving the board around too much putting stress at these points on the wires. Should be easy to spot and easy to fix. This is also a problem with the power wiring at the top of the key/display board.

For the same reason, it might be possible that the wires break at the point where they enter the epoxy in the R/W head. Repairs of this nature might still be possible, but a bit more fiddly as you would have to carefully dig the epoxy out around the wire end, solder wires back on and put more epoxy back on to secure them.

Hopefully all data checks were done with an oscilloscope to prove that the data is not actually appearing at the sense chip inputs and outputs.

cheers

Tony

[albertofenini]

Hi Tony

Do you think that rework the hybrid would help ?
Could it possibly be a cold join ?
Take care !

(11-29-2020 12:57 AM)teenix Wrote:  

(11-28-2020 09:59 PM)albertofenini Wrote:  Hi Tony
Too bad we don’t have one ...
I guess only way would be sourcing it from a broken unit ...
Could it be a problem of bad connections between the hybrid pins and the hybrid internals ?
Take care !!

Hi Alberto,

I'm not sure, but I suspect it is more a chip problem internal to the hybrid, especially since normal key programming has issues as well as the card read/write. All other functions seem to work so the internal timing etc and associated connections would seem ok.

An internal connection issue maybe the cause of the broken WB signal but this maybe a symptom rather than the root of the problem. Any sort of surgery to the hybrid internals would be highly problematic and maybe impossible even with specialist equipment.

cheers

Tony

[teenix]

(11-29-2020 08:13 AM)albertofenini Wrote:  Hi Tony

Do you think that rework the hybrid would help ?
Could it possibly be a cold join ?
Take care !

It wouldn't hurt to try :-)

cheers

Tony

[albertofenini]

Hi Tony, hi everyone

It has hurt a little since pin 4 and pin 11 broke ...
Anyway, we fixed with two ultra thin cables
Inspection one removed the hybrid didn’t reveal anything
Once the hybrid has been put in place again the behavior was exactly as before

When loading or executing a program it goes in a hold status

Tony How much is worth to replace the ROMs set ?

Updated pictures here : HP-65 1603S03504

(11-29-2020 09:00 AM)teenix Wrote:  

(11-29-2020 08:13 AM)albertofenini Wrote:  Hi Tony

Do you think that rework the hybrid would help ?
Could it possibly be a cold join ?
Take care !

It wouldn't hurt to try :-)

cheers

Tony

[teenix]

(11-29-2020 05:03 PM)albertofenini Wrote:  Hi Tony, hi everyone

It has hurt a little since pin 4 and pin 11 broke ...
Anyway, we fixed with two ultra thin cables
Inspection one removed the hybrid didn’t reveal anything
Once the hybrid has been put in place again the behavior was exactly as before

When loading or executing a program it goes in a hold status

Tony How much is worth to replace the ROMs set ?

Hi Alberto,

Sorry, I thought you would just re-solder each pin one at a time while the hybrid was still in place.

I don't know about the ROMs as I have never replaced any.

cheers

Tony

[Kostas Kritsilas]

(11-29-2020 07:49 AM)teenix Wrote:  

(11-29-2020 06:50 AM)Kostas Kritsilas Wrote:  I was told it was probably the read/write head or the wiring to it was at fault when I sent it out to have it repaired. The motor comes on, and pulls the card through well, it just doesn't read anything. All the keyboard functions work, and I can type in and run programs. Even though the card reader is pulling in the card, the display does not show any data has been read.

If you want to sell a spare read/write head, please let me know what you want for one. If you are willing to attempt a repair, please let me know what you would like to charge for that. I would very much like to have the calculator fully working, but cannot, due to vision issues, do the work myself.

You didn't mention card writing, but if it is the coils inside the head, (which would not be repairable), then I would assume the card cannot be written too as well, as these perform both read and write functions. The same for the external wiring to the sense board. If one is broken, then a read or a write will not succeed.

If it is the external wires, I'd be pretty sure like [kby] says, that the wires will most likely be broken where soldered into the circuit board, which is due to moving the board around too much putting stress at these points on the wires. Should be easy to spot and easy to fix. This is also a problem with the power wiring at the top of the key/display board.

For the same reason, it might be possible that the wires break at the point where they enter the epoxy in the R/W head. Repairs of this nature might still be possible, but a bit more fiddly as you would have to carefully dig the epoxy out around the wire end, solder wires back on and put more epoxy back on to secure them.

Hopefully all data checks were done with an oscilloscope to prove that the data is not actually appearing at the sense chip inputs and outputs.

cheers

Tony

I don't believe that the person attempting the repairs used an oscilloscope, as his background is in chemical engineering. He said he did inspect all the traces going into and out of the card reader ICs under magnification, and said they were fine, and verified them with a multi-meter.

He had some spare HP-65 card reader parts, and said that he tried to repair the unit with them, but then found out that the spare units themselves were bad. he also said that he would be putting his own spare HP-65s up on eBay as non-working units. he also said that the HP-67 heads are better in this regard, as they use a thicker, better insulated wiring between the board and PCB.

This is all I have on the repair process. The statements about the read/write head being bad, whether it was the leads going to the head, or the head itself, came from the person attempting to repair the unit. I can forward the relevant emails in direct messages if anybody would like to see them.

[teenix]

(11-30-2020 02:25 AM)Kostas Kritsilas Wrote:  I don't believe that the person attempting the repairs used an oscilloscope, as his background is in chemical engineering. He said he did inspect all the traces going into and out of the card reader ICs under magnification, and said they were fine, and verified them with a multi-meter.

He had some spare HP-65 card reader parts, and said that he tried to repair the unit with them, but then found out that the spare units themselves were bad. he also said that he would be putting his own spare HP-65s up on eBay as non-working units. he also said that the HP-67 heads are better in this regard, as they use a thicker, better insulated wiring between the board and PCB.

This is all I have on the repair process. The statements about the read/write head being bad, whether it was the leads going to the head, or the head itself, came from the person attempting to repair the unit. I can forward the relevant emails in direct messages if anybody would like to see them.

What a shame the parts were unusable.

Without an oscilloscope check, it would be hard to define where the problem might be.

On reflection, I guess it is hard to test a card write because you cannot read it back again unless you have access to another 65.

cheers

Tony

[Kostas Kritsilas]

(11-30-2020 03:34 AM)teenix Wrote:  

(11-30-2020 02:25 AM)Kostas Kritsilas Wrote:  I don't believe that the person attempting the repairs used an oscilloscope, as his background is in chemical engineering. He said he did inspect all the traces going into and out of the card reader ICs under magnification, and said they were fine, and verified them with a multi-meter.

He had some spare HP-65 card reader parts, and said that he tried to repair the unit with them, but then found out that the spare units themselves were bad. he also said that he would be putting his own spare HP-65s up on eBay as non-working units. he also said that the HP-67 heads are better in this regard, as they use a thicker, better insulated wiring between the board and PCB.

This is all I have on the repair process. The statements about the read/write head being bad, whether it was the leads going to the head, or the head itself, came from the person attempting to repair the unit. I can forward the relevant emails in direct messages if anybody would like to see them.

What a shame the parts were unusable.

Without an oscilloscope check, it would be hard to define where the problem might be.

On reflection, I guess it is hard to test a card write because you cannot read it back again unless you have access to another 65.

cheers

Tony

He did have a working HP-65, with which he used to verify that my card reader was bad, and his own HP-65 spare card reader parts were bad as well. I don't think he was aware that the HP-67 read/write heads are interchangeable with the HP-65 read/write heads. He did say that he could not repair the HP-65 read/write heads as their wires were, to quote him "hair thin" and he did not have the soldering skill to attempt that. He did say that he would not be working on HP-65 card readers any more, and that the read/write heads on the HP-67, because they were larger in diameter, were a lot more reliable, and easier to work on.

I think I am hijacking Alberto and Edouardo's thread, so I think I should stop posting about my HP-65 issues. If anybody would like to discuss this in further detail, please direct message me, and I will give you any information that I can.

[[kby]]

I will say that one other thing I recently was reminded of (think I read it somewherelong ago, but just encountered it) is that a missing (or presumably open)
Capacitor across the card reader motor can cause read failures.

[additional information for completeness: this problem turns out to have been caused by something on the CPU board, and is therefore likely the card reader controller part of the hybrid chip]