HP-65 more on the card reader

[wallyt]

Yes, I would suspect the caps, too. Unless they check okay with a good quality ESR meter I would replace them with new ones, not just replacements from another 65. Also, at least in my 65, the motor speed resistor was really tricky. The original value varied by only 110 ohms from what I found to get the reader to work correctly.

[albertofenini]

Hi everyone,
we'll give a try with new capacitors, we still should have some spare new parts,
this is the original scheme : two capacitors are involved with the channels from the magnetic head,
the third however is more related to motor, which runs smoothly, which is also proven by the fact
that the unit reads 100% of the cards.

   

We'll post later ...

[albertofenini]

Hi everyone,
we changed all of the three capacitors, but nothing changed.
Reading works like a chime, while writing doesn't work at all (WA and WB traces are attached).
As said, we changed, or swapped EVERY possible item from other working units.
The only thing we couldn't change was the keyboard.
What chances there are that the fault is in the keyboard component ?
Thanks to everyone for the effort in helping us !

   
   

[teenix]

(02-03-2019 08:15 PM)albertofenini Wrote:  Hi everyone,
we changed all of the three capacitors, but nothing changed.
Reading works like a chime, while writing doesn't work at all (WA and WB traces are attached).
As said, we changed, or swapped EVERY possible item from other working units.
The only thing we couldn't change was the keyboard.
What chances there are that the fault is in the keyboard component ?
Thanks to everyone for the effort in helping us !

As your calculator recognises the PRGM/RUN switch, I don't think that would be the problem.

I imagine the bit density on the HP65 cards would be slightly less than the HP97 which stores them at roughly 1 KHz, ie. 1 bit per millisecond

Therefore on he HP65, I would expect to see bits changing at around 1mS.
Looking at the bottom trace, there appears to be some high frequency noise there, not a legible bit pattern. Assuming the horizontal trace is 2mS per division.

As I mentioned earlier, a full program containing all NOPs should give an alternating bit pattern on track 0 with a steady logic level on track 1.

A program memory full of value 63 would see an alternating bit pattern on track1 with a steady level on track 0. However [63] is a Top Of Memory marker and can't be done. The closest is to fill memory with ENTER which is [62].

Sorry I haven't been much help. I have to guess solutions as I don't have a 65 to help you compare results.

cheers

Tony

[albertofenini]

I agree, it looks like there is some high frequency noise there, kind of something would interfere ...
but what could possible generate that noise ?
The motor ? it was changed
The head ? it was changed
The capacitors or the resistance ? they were changed
Something on the logic board ? it was changed
I thought of a defect in the PCB ... it was changed as well
You are right, the switch program works ...
I don't think that the anode/cathode drivers can interfere that much
I could try to shield the card reader from the rest of the unit, but does it really make sense ?

On the other side, we also have a unit S/N 1333A20538 where both WA and WB looks good
but still it doesn't record anything on the card ...

We will keep trying ...

[teenix]

(02-03-2019 11:43 PM)albertofenini Wrote:  I agree, it looks like there is some high frequency noise there, kind of something would interfere ...
but what could possible generate that noise ?

On the other side, we also have a unit S/N 1333A20538 where both WA and WB looks good
but still it doesn't record anything on the card ...

We will keep trying ...

bad ground connection, on pcb, inside chip, on probe
probe not connected properly, scope not setup properly
open circuit from CRC, or inside CRC


If the WA WB signals appear at the sense amp and it reads ok, then I would imagine faulty write circuits in sense amp. The rest of the circuitry is common between read/write, except the outputs to the RW head.

Can you see any activity from the sense amp pins that drive the head?

cheers

Tony

[albertofenini]

Hi Tony,

scope setup is the same for all the other units, and we soldered same cables at the end of the card reader PCB to attach the probes,
hence I would exclude bad connections fo the scope or a bad setup.

We checked all the joints on the original card reader and we built another one stripping
all the components, clearing the PCB and soldering again, from scratch, I would exclude bad joints on both.

The amplifier, unless two different chips have the same problem, which may happen but I think is unlikely, should be out of discussion

We have not measured activity other than at WA, WB, RA and RB pins.
We can check activity at the head cables and on the other Amplifier pins.
However, two complete different card readers giving the exact same waveform ??
It has to be something else ...

We'll check and post.

Anyone has a suggestion for the other serial number ???

[wallyt]

With reference to your waveforms for WB, as you noted there seems to be a lot of noise there. Have you used your scope to look at all of the power supply voltages to verify that they are clean DC? Scope all of the voltages from the power supply on the logic board to make sure they are clean.

Are you powering the calculator from a battery pack or external supply (is it clean)? Did you check or change the capacitor on the motor? Isn't there a filter cap on the keyboard for Vbsw?

[albertofenini]

Thank you very much for the tips
The unit is powered from a battery, so I take it is clean
We didn't change the capacitor on the motor, we'll do that
We'll scope all of the voltages on the unit and see what we find
What do you mean with filter cap(acitor ?) on Vbsw
The PSU provides Vcc, Vss and Vgg
What is Vbsw (which I have seen in some schematics) and how much should be ?
Is that the Battery Voltage ?
I have read that :
Vbsw : Now, with the logic board removed, apply the power to the
battery contacts. You should be able to measure the battery voltage
between Vbsw and Gnd on the 28 pin connector. Disconnect the power from
the battery contacts.
Should we do as described to measure Vbsw ?
Thanks again Alberto

[wallyt]

Vbsw is the switched battery voltage (pin 18 of the 0158 sense amp). If you are using the battery for power I would assume this to be clean (unless something is getting into this). The other voltages, however, are generated from the DC-DC converter on the logic board. I would definitely scope those.

The only other thing I could say about this is that in my experience I have run into a bad mag card now and then. I would clean these with an alcohol wipe (both sides) and let it dry. Lots of oxide would come off on the wipe. They would work fine after this. But you are showing bad waveforms so I doubt this would be your problem.

[albertofenini]

Last update on S/N 1511A

We scoped as suggested (thanks for the idea) all the Voltages, Vbsw, Vcc, Vgg, Vss in different points.
These were all clean.
Before replacing the capacitor on the motor, we decided to swap the only thing we hadn't swapped until now : The keyboard
And that was it.
We used the keyboard from S/N 1333A and combined with the 1511A assembly it was giving clean RA/RB and clean WA/WB.
A little adjustment of the eccentric cam and it was working perfectly, reading and recording.
In this moment it has the 2 x 22microF and 1x 15 microF capacitors new, we may as well leave them.

At this point we tested the viceversa : we installed the defective keyboard into the 1333A assembly
(logic board, card reader and battery plate) and as expected the problem we had before was now
affecting this unit.

We looked very carefully at the keyboard with a magnifying glass, we cleaned it and there are no visually noticeable defects.

   

We still have an additional keyboard, but it misses the anode and cathode drivers,
we should transplant the actual ones and we feel it's a bit too risky.

Now that we have isolated the problem, how can we check the integrity of the keyboard pcb ?

Thanks everyone for the support !!

[teenix]

(02-04-2019 10:22 PM)albertofenini Wrote:  A little adjustment of the eccentric cam and it was working perfectly, reading and recording.

Now that we have isolated the problem, how can we check the integrity of the keyboard pcb ?

Good to hear.

Check the PCB numbers probably etched into the board like the traces. They might be different types. The board to CPU connections might be different also.

It's interesting that this part affected the card write.

cheers

Tony

[albertofenini]

Here we are :
Looking at the picture, the two keyboards can be seen :

UPPER (Faulty) model 00065-60201 rev 51 803F D 61
LOWER (Spare without anode and cathode drivers) model 000605-60201 rev 3351 B-1341-61

The lower is much similar to the one we have in the working calculators,
however, the upper has been really carefully examined and appears to be perfect

   

Any idea ?

Thanks again !!!

[albertofenini]

LAST UPDATE

We decided to be brave and we moved the Cathode and Anode Drivers from the faulty keyboard
to the spare keyboard.

The machine turned-on fine, the move didn't hurt the drivers,
reading from a card was good, but writing had exactly the same problem ...

Now ... given that last night we fixed the unit S/N 1511 by swapping the keyboard and it
worked out, the only thing that we can think of, is that the problem is in the active components
i.e. the Anode and the Cathode Driver.

Next time we will measure if the display signal is clean or not, as this looks to be the only possible
explanation

Take care !!!

[AndiGer]

I wonder what are the electronic parts above the two drivers? Inside the HP-35 it looks like it is 4 discret parts. Did you examine those? Another possibility beside the drivers?

[albertofenini]

In the scheme I have these looks like Inductor modules, which could be responsible for some noise.
Again, we have some spare parts, but I would find extremely unlikely that on both keyboards these are giving exactly the same fault.
It may be worth to shield the carder reader amplifier to see if in this way it is protected from interference.
What do you think ?




   

[teenix]

(02-06-2019 02:57 PM)albertofenini Wrote:  In the scheme I have these looks like Inductor modules, which could be responsible for some noise.
Again, we have some spare parts, but I would find extremely unlikely that on both keyboards these are giving exactly the same fault.
It may be worth to shield the carder reader amplifier to see if in this way it is protected from interference.
What do you think ?

They are there to drive the display LEDs. I guess nothing is impossible as the previous problem showed, but any radiated noise from these shouldn't drown out the WB signal.

cheers

Tony

[Harald]

(02-06-2019 07:00 PM)teenix Wrote:  

(02-06-2019 02:57 PM)albertofenini Wrote:  In the scheme I have these looks like Inductor modules, which could be responsible for some noise.
Again, we have some spare parts, but I would find extremely unlikely that on both keyboards these are giving exactly the same fault.
It may be worth to shield the carder reader amplifier to see if in this way it is protected from interference.
What do you think ?

They are there to drive the display LEDs. I guess nothing is impossible as the previous problem showed, but any radiated noise from these shouldn't drown out the WB signal.

cheers

Tony

I quess the most likely problem is the andode driver. This also contains the circuit that generates the clock signals. Maybe one of the clock signals phi1 or phi2 has noise on it. If you want to get to the bottom of why one of the keyboard PCBSs doesn't work, I'd suggest looking at those clock signal.

Here the anode driver is described in detail.

Cheers,
Harald

[albertofenini]

UPDATE 7/FEB/2019

This is getting intriguing ... so we decided to perform more tests.

We haven't had the time to measures signals at the anode and cathode driver yet, but we will do it in the future.

We did however another test.
Ignazio suggested to investigate, thanks for it, about a possible incompatibility about the two different boards
and the two possible card reader amplifiers : 1826-1058 and 1826-0332.

We performed the first, using keyboard model 000605-60201 rev 3351 B-1341-61 we tested a
card reader with 1826-0332 (originally this unit was completed by a card reader 1826-1058) but we
had exactly the same behavior.

For completeness we will perform soon the opposite test, we will use keyboard 00065-60201 rev 51 803F D 61
and we'll test the two card readers with the two different Amplifiers.

On the other side, what exactly should we measure at the anode driver to see it is noisy ?

Thanks for help to everyone

[Harald]

(02-07-2019 10:39 AM)albertofenini Wrote:  UPDATE 7/FEB/2019

On the other side, what exactly should we measure at the anode driver to see it is noisy ?

Thanks for help to everyone

Just measure the two clock signals on pin 7 and 8 of the anode driver. Both should be clean signals. Phase shifted but otherwise identical. If you measure them on both the good and the bad board, you will be able to tell if there is a difference.

Cheers,
Harald