Restoring HP-65

[albertofenini]

Dear All, it's nice to be back again.
I had posted several years ago, then due to personal reasons I had to stop working on our collection.
I also have a new helper, now that my son has turned 11, he has joined me in this passion.
I will post later pictures of our collection.
We are now in the process of restoring three HP-65, trying to get at least one working unit out of the pieces.
We also have noted that although all the calcs had been stored away cleaned and sealed, some of them have developed oxidation when not corrosion, even if none of them was stored with batteries inserted.
I think it was mostly due to non perfect cleaning of the original corrosion.
Before posting I have searched in this and in the old forum, looking for a good recipe to obtain the removal of oxidation, of the corroded parts and to make shine again the contatcs or the soldering joints.
I remember using a dilution of vinegar with water and baking soda but I can't remember the exact composition.
Can anyone please help us ? Take a lok at the green thin lines close to the gold traces.

Thanks in advance to everyone for your support !
Alberto and the young collector Edoardo !

[albertofenini]

Just to add few things we tried meanwhile-
We have tried to clean some of the green spots from the motherboard with lemon juice
Apparently it worked out but on a closer look something is still there
Attached are two pictures of the CPU board
You'll notice the is not shiny as it should be.
Then since we are restoring this unit and we don't want to have too many unstable things like the keyboard or so, we have put this mother board in a working HP-65.
Turns out that it works, but it is not responsive to the last two bottom lines of keys
We know for sure that the keyboard circuit for that unit is fully working

[Maximilian Hohmann]

Hello!

(11-25-2018 01:13 PM)albertofenini Wrote:  Turns out that it works, but it is not responsive to the last two bottom lines of keys
We know for sure that the keyboard circuit for that unit is fully working

I wouldn't bother to remove the residual oxidation from that board. There was very little of it in the first place. I have seen many perfectly working calculators which looked much, much worse.
To remove the oxides both acidic (vinegar, lemon juice) and alkaline (baking soda) chemicals can be helpful, depending on the oxide. I usually use both, one after the other. Do not mix them as they will neutralise themselves. Afterwards thoroughly rinse the board. Remains of your acid and/or alkaline cleaners will quickly start corrosion of their own otherwise. And dry it completely - five miniutes with my wife's hairdryer usually do the job.

Here is an old thread which shows what could be wrong with your keyboard and where to start looking: http://www.hpmuseum.org/cgi-sys/cgiwrap/...ead=237065

Good luck with bringing the calculator back to life!
Max

[albertofenini]

Hello Max,
we wrote in the past, so it's nice to hear from you again, thanks for your reply.

Let's say that we have a working unit, and two units under restoration, these tow units are complete with keys, keyboard and power connection that we will call KBD001 and KBD002.
At the moment, other than the working CPU, I only have one CPU board under restoration , CPU S/N APD 1540 T.

This is what I experience so far :

PROBLEM 1 : If I put the CPU under restoration I didn't have the control of the last two rows of keys. This can't be because of a bad keyboard, since the unit housing the CPU is the working one.
So I re-passed the soldering points of the hybrid on the side that control the keyboard rows (I do have the schematics), and it has got worse, since now most of the keys are not responding

[attachment=6635]

Then I did the opposite : I have put the working CPU in the cases with KBD001 and KBD002. Both appear in very good conditions as you can see from the pictures, and in this test the CPU board is working since is coming from the working unit.
KBD001 as a weak metal plate, you don't hear the classic click when a key is pressed, kind of stressed metal. But if all the keys are pressed strong enough they work.


PROBLEM 2 : when the working CPU is put in the case with KBD002, on few led lights on ,as you will see in the picture.


[attachment=6636]
[attachment=6637]
[attachment=6638]

What can I do to solve PROBLEM1 and PROBLEM2 ?
Thanks in advance, Alberto & Edoardo

[teenix]

(11-25-2018 09:59 PM)albertofenini Wrote:  PROBLEM 1 : If I put the CPU under restoration I didn't have the control of the last two rows of keys. This can't be because of a bad keyboard, since the unit housing the CPU is the working one.
So I re-passed the soldering points of the hybrid on the side that control the keyboard rows (I do have the schematics), and it has got worse, since now most of the keys are not responding

Then I did the opposite : I have put the working CPU in the cases with KBD001 and KBD002. Both appear in very good conditions as you can see from the pictures, and in this test the CPU board is working since is coming from the working unit.
KBD001 as a weak metal plate, you don't hear the classic click when a key is pressed, kind of stressed metal. But if all the keys are pressed strong enough they work.

The keyboard connectors on the restoration CPU look black with corrosion, maybe some of this black is in the pcb holes and is not making good connection to the keyboard pcb pins.

Sometimes the component pins have corrosion on them and new solder will not properly fuse to them causing open circuits or high resistance joints.

Just guessing, but it may be possible that resoldering the hybrid pins caused damage to the module, open circuited the pins as described above, or internally, too much soldering iron heat, static damage.

If the exact same problem exists between the different keyboards then the hybrid would be suspect and or the keyboard to CPU connectors.

(11-25-2018 09:59 PM)albertofenini Wrote:  PROBLEM 2 : when the working CPU is put in the case with KBD002, on few led lights on ,as you will see in the picture.

Does the display change when keys are pressed? If so, it might be the same problem. Corrosion on the keyboard to CPU connectors.

cheers

Tony

[albertofenini]

Hi Tony, nice to meet you,
thanks for your reply,
all the connectors have been cleaned, all the holes have been cleaned so the black you see is just a reflection of the flash.

PROBLEM1 I do agree that the Hybrid could have suffered damage because of the soldering, but I'm using a 30W soldering iron, which I have been successfully used for many other projects, so it would be difficult. I have a couple of untested Hybrid so I may as well swap them and see if it gets better

As for PROBLEM2, I don't know what to say, the keyboard PCB is all clean, and yes pressing keys cause more leds to light on. However in this case the cPU module is good, so I'm more thinking about the cathode and anode drivers on the keyboard PCB, do you know how I could check them ?

Thanks again Alberto

[teenix]

(11-25-2018 11:40 PM)albertofenini Wrote:  Hi Tony, nice to meet you,
thanks for your reply,

As for PROBLEM2, I don't know what to say, the keyboard PCB is all clean, and yes pressing keys cause more leds to light on. However in this case the cPU module is good, so I'm more thinking about the cathode and anode drivers on the keyboard PCB, do you know how I could check them ?

Thanks again Alberto

Same :-)

Look for common segments not lit when entering numbers, that might point to the anode driver. Only certain digits lit, might point to the cathode driver.

You have the schematic, maybe check the connections from hybrid to anode and cathode drivers and check the power supply voltages are correct.

Testing the CPU board, might be as simple as pressing [divide] which at power on would be a divide by zero causing an error. The display even if faulty should flash because of the error. That might verify (to a certain extent) that the actual hardware is operating.

cheers

Tony

[albertofenini]

Hi everyone,

firs of all let me introduce my partner in all these activities,
he is of great support, optimistic and of great help in this work,
let me introduce you Edoardo, my 11 years old son and probably one of the youngest calcs and
led watches collector.

Here we are working at our HP-65 restoring project, we will dea later with the card readers.
Following a old procedure found on the Forum we have better organized our work,
borrowed a working HP-65 working mother board from our collection and mounted the
keyboards pcb as shown.

So far we have managed to bring back to life a keyboard/led pcb from the broken units,
by cleaning with a strip of sandpaper (2000 grit) all the keys plates.

You will see in the pics is called Keyboard 001 and shows full array of 8.

Then we did the same with KBD002, but with no results.

Also, we had a third Keyboard PCB, we can't recall wher it was coming from, that was
with out LED modules. So we recovered the modules from an HP80, but also this had
some issues.

We had then some donors, like HP80 and HP45 and we recovered from those the
Anode Driver 1820-1029 (HP80 and HP45 use the same chipset the HP-65 does) and the
Cathode Driver (the HP45 uses the same that the HP-65 does) however, we also don't know
if those were working previously or not, and the results are not so good.

We are now thinking of soldering two 20 pins sockets on a Keyboard in order try different
combination with the hope to find at least a good couple of chipset so to bring back to life
a second keyboard PCB.

Does anyone know what kind of measures can we take to understand if those chipset are working ?

Then we also tried to recover the CPU board we have cleaned last time, but we no progress.
The only thing we noted is that some of numeric keys work (3,6 and 9) and that Enter works,
and the - and + keys work, so that we have been able to perform simple operations.
This makes me think that the Hybrid is at least partially working, am I right ?

We are thinking to by a USB oscilloscope and connected to a laptop, this could be a good improvement in our lab and could help us in make more efficient diagnosis.

Take care everyone, and if anyone wants to add any idea, please let us know

Alberto & Edoardo

[burkhard]

Alberto,
Thanks for recounting your adventure with Edoardo here... it looks like a good father-son project! I need to be doing the same with my own kids.

One question... lemon juice makes me a little nervous. Certainly it's acidic, but it also carries a lot of sugar that I don't think would be great to get all over the board. We tend not to think of lemon juice as being sweet but apparently on the same volume basic, it's got 85% the sugar of orange juice. White not white vinegar, which is otherwise pretty benign?

I had considered a USB or kit oscilloscope in the past, but the prices of real hardware digital scopes has gone way down and those are so much nicer. The Rigol 1054Z is particularly popular among electronics hobbyists right now: a 4 channel, 50 MHz for $US 350. Check some Youtube reviews; I don't have one myself, but am leaning. There are others as well; the market is getting more crowded.

Keep us posted! I've got a 65 I need to take apart as well, Besides the standard fun with the card reader, it's got some keys that I think need cleaning... feel fine, but hard presses required.

[Hans-Peter]

(12-02-2018 06:07 PM)albertofenini Wrote:  We are thinking to by a USB oscilloscope and connected to a laptop, this could be a good improvement in our lab and could help us in make more efficient diagnosis.

I have an old textronix since 30 years. But since I have a jye tech, I leave the tex behind the table.
https://jyetech.com/Products/LcdScope/e138.php
For most things its fine and with a 9 volt battery handy to use. I screwed them on a wooden plate, put a switch on it and that's all.
Today I would take this one
https://jyetech.com/Products/LcdScope/e150.php

   
There is a battery tester (resistance), multi tester (often in use) and the scope.

[albertofenini]

Dear Hans,
thank you very much, that could be an idea, although the nearest reseller is in Switzerland,
and I'm afraid that is not plug and play, I mean that could be some adjusting to be done, I have read somewhere there are variable capacitors that need to be adjusted. What was your experience with this kit ?

Also please note that in our documentation for repairing the HP-65 the oscilloscope must have a 1Mhz bandwith, while
the DIY kit arrives only to 200KhZ, it looks we shoould buy


https://jyetech.com/Products/LcdScope/e062C.php

At the moment we are looking at these models :

(For the moderator, if posting Amazon links is not appropriate please remove them, I'm in no way affiliated nor promoting and product/vendor)

1. https://www.amazon.it/dp/B07JZ69H94/?col..._lig_dp_it

2. https://www.amazon.it/dp/B01N404XK0/?col..._lig_dp_it

3. https://www.amazon.it/dp/B07GXSS9NC/?col..._lig_dp_it

4. https://www.amazon.it/dp/B015XTOOKY/?col..._lig_dp_it

and eventually this one, which is however out of our budget

5. https://www.amazon.it/dp/B00DRABSBA/?col..._lig_dp_it

Does anyone can suggest us a model ?
Thanks everyone for help, Alberto

[albertofenini]

Dear All,

we spent the last three days recovering the third Keyboard PCB.
We have now a working Keyboard (001), a second one (002) where we de-soldered and re-soldered
Anode Driver 1820-1029 and the Cathode Driver 1820-1226 as it appeared to be some oxidation below these two circuits, and a third one (003) where we replaced the battery terminals as they where missing and cleaned the bakelite part.

We also tested the LEDs components and they are all working

However, when the working CPU module is connected to boards 002 and 003 the LEDs behave as in the attached photo.

We decided then to take some measures, remember we are still lacking of an oscilloscope, so we only could measure voltage around the 1820-1029 and the 1820-1226 pins. Below you will find the readings.

We were surprised to find values very similar to the working unit althoung these last two 002 and 003 were not working.

Keyboard 002 lits a faint, not even complete 0) in position 10 out of the 15 leds
Keyboard 003 lits segments e and h of all the last five leds
Both blinks if a divede by 0 or R/S key is pressed

Anyone can suggest a better way to diagnose what the problem could be ?
Anyone can suggest a better way to test the Anode and Cathod drivers ?

Thanks in advance from Edoardo & Alberto

[teenix]

(12-09-2018 08:07 PM)albertofenini Wrote:  Keyboard 002 lits a faint, not even complete 0) in position 10 out of the 15 leds
Keyboard 003 lits segments e and h of all the last five leds
Both blinks if a divede by 0 or R/S key is pressed

Anyone can suggest a better way to diagnose what the problem could be ?
Anyone can suggest a better way to test the Anode and Cathod drivers ?

Thanks in advance from Edoardo & Alberto

Have you looked at the inductors. They are a passive component and I would imagine quite robust, but they may fail over time I guess.

I believe that improper timing of these parts can cause damage to the anode driver. Harold, a list member, might have more info on this.

There is not really much you can do without better test equipment like an oscilloscope, without actually exchanging the anode and cathode drivers for known good ones.

cheers

Tony

[_dev-null]

Hello community, ciao Alberto!

I have a similar problem with my 65. The CPU is working - Phi_1 and Phi_2 Signals appear to be correct. Sometimes if i touch the Phi or Sync testpad with the Scope probe, the Display works until I switch off the calculator.

Next week i receive two more Hp-65, then i can exchange parts and maybe i get one unit working again.

I will post my results ...

Greetings
Stefan from Austria

PS: I have a UTD - 1025CL Scope, which is pretty nice for testing.

[albertofenini]

Santa Claus brought us the Oscilloscope !!!
Just give us the time to install the software and we will start post again
Meanwhile, Merry Christmas to Everyone (sorry for being late with our wishes)
Edo & Alberto

[albertofenini]

Dear All here we are !
Happy New Year everyone !!

We installed the software and started to use the Scope.
It is absolutely the tool that makes possible repair this stuff.

We worked on both the keyboards and the CPU boards and we have got so far two units working.
On one CPU board we discovered that the CR CK signal were not handled and you will see why in
the next pictures. We then de-soldered the Hybrid and replaced with a spare one. This was from a
old project where the board had some PSU issues but the CPU was eventually working.

We also fixed another CPU board that had probably a PSU issue replacing the diodes and a
capacitor that we took out from a donor. We also stripped all the components and re-soldered on a different PCB.

With the scope we have been able to investigate about anode and cathode drivers on the two
keyboards that were not working properly. Both had the ICs non working and we soldered a
DIP socket, so that we could test all the spares that we had from two old HP45 that share the
Anode and Cathode drivers. We have found out only a working couple so that now one of the
keyboards is also working.

Next step will be reassembling these two units and start repairing their card readers.

What's left :
A bare HP65 CPU PCB
Enough components to rebuild it, although one of the ram modules could be faulty has it had a note saying "faulty"
What's missing :
a good couple of anode and cathode drivers
a good hybrid
If you look at the pictures you will notice that we have two hybrids left.
These are "first generation" and the pins are physically connected to the ceramic and
then a gold strip connects the pin to the internal modules.
Later versions are more robust and pins come out from underneath.

One of these two is in unknown working conditions, since it was given to me from a collector
many years ago, but I can't recall if it was working or not, also has some broken pins.
The second one, is logically working since it was able to perform math, but has a lot of missing
pins, and what's worst the gold strips coming out of the hybrid are too short.
This is the one that was not addressing many rows and columns exactly because of this problem,
We have collected all of the missing pins, so we have those, but has anyone ever succeeded in
re-connecting these ? How could we get those short strips of gold soldered with the pins ?
Any idea or suggestion is welcome.

Also, if some one has spares to share or willing to trade them for something else please let us know.

We will keep you posted on the card reader in the next days

Take care and thanks to all of you that has provided suggestions or hints

Edo & Alberto

[Kees Bouw]

(01-02-2019 12:11 AM)albertofenini Wrote:  With the scope we have been able to investigate about anode and cathode drivers on the two
keyboards that were not working properly.

Can you tell a bit more about exactly what and where you measured with the scope, and how the display was helpful in determining the problem?

[albertofenini]

(01-02-2019 12:49 AM)Kees Bouw Wrote:  

(01-02-2019 12:11 AM)albertofenini Wrote:  With the scope we have been able to investigate about anode and cathode drivers on the two
keyboards that were not working properly.

Can you tell a bit more about exactly what and where you measured with the scope, and how the display was helpful in determining the problem?

We followed the procedure outlined in this document

http://home.comcen.com.au/therobinsons/h...repair.pdf

It has been very helpful and I strongly suggest anyone trying to troubleshooting to do so.

We had one CPU that was doing the math but only the number 6 and 9 and the enter and
+ and - keys were working, so that only column 1 and column 5 and two rows were working
only using the scope we have checked this out, because we have compared the waves forms from a working unit with this.

We have then discovered that signlas like Phi1 and Phi2 were good, while some of the CR and CK were not active, once we have find out this we have decided to change the hybrid.

Also with the cathode and anode driver, you read the document, you will see exactly which
kind of waves you have to expect and where you can find them, and this has been of great
help in finding which driver was not working.

Also, the HP65 schematics from the CD have been used, since they give you a good overview
of how things are connected and what the ICs pinout are.

Hope this helps, Alberto

[albertofenini]

As promised, we used some of the weekend time to restore one of the card readers.
It had the gummy wheel problem and the head wires were all disconnected.
Some of the traces of the PCB were also heavily oxydated.
We cleaned them with lemon, washed them with isopropylic alcool and then we paper sanded them
with 2000 grit just to be sure they were clean.
It would be a good idea to paint them with transparent paint, which we don't have, in order
to protect them from future oxidation.
Then we cleaned the little spheres and mounted the o-ring.
We then soldered the head wires, if someone is interested we can post a document describing
the correct color order.
Removed all the cables used for the test phase and put everything together.
It was reading the cards with no problems since the beginning, while it was not able to write cards.
By adjusting the white axle we were able to record cards as well in a good percentage (9/10).

Belowe are some photos and a link to a small youtube video

What's been done :
1 removing all the keys and sof washiing them
2. washing the upper half of the case and polishing with carnauba wax
3. restoring the silver trim line
3. cleaning with small strips of san paper the keyboard under the metallic plates
4. removing all the components from the old logic PCB and mounting them on a spare one
5. this keyboard already had working anode and cathode drivers
6. greased the card reader motor
7. re-soldered the card reader magnetic head wires
8. mounted new o-rings

What's left to do :
1. clean the lower half of the case
2. replace the instruction label on the back
3. clean and replace the original rubber feets
4. clean the leather pouchette
5. find a complete Standard Pac and Standard Pac Manual

Note : if the Calc is connected to the charger with the battery inserted, the card reade operates
weird, anyone has an idea why this is happening ?

What we learned
1. Never give up !
2. Don't be afraid to break anything
other than breaking something what could possibly happen ?


Thanks everyone for the support provided during this restoration, if we can be of help in any way during this activity.

We will restore the second one, but more over, our collection needs more than 20 repairs on
totally different models from Woodstock, to Spice to Topcat, to Coconut, so is going to be fun !

If we can help for anyone, drop us a line and thanks again for your contribution !

HP65 Final Test

[albertofenini]

Hi everyone,

after restoring the HP65 and its card reader, we have been able to read most of the pre-recorded cards, and record some new cards.
However, what surprises us, is that we are not able to successfully write the type of cards that should really belongs to the HP65.

Look at the attached picture, we have even been able to record on both sizes of a card, and I believe that some of the cards successfully recorded are from a HP67,
while we can't succeed in writing on the first type that was coming with the HP65 itself.

Has it happened to someone before ?
Any idea ?

One more question.
If we record a program on a card, says about 65 steps, and then we override this with a new recording for shorter program, says about 50 steps,
will the old program overridden or the 15 more steps will remain and is this goin to be a problem ?

take care Alberto