HP-65 goes on hold after reading a card

[albertofenini]

Hi everyone

this is a unit that we restored a while ago but never truly tested
Few days ago we replaced a coil following the advice of Teenix (thanks for help !) and fixed an issue that was causing the G segment to be always off (the related coil was open)

We discovered that when a card is read, a steady 0.00 appears, but when we try to execute the loaded program, the unit goes in a non responsive status.

With the diagnostic card, to name one, it happens just hitting the R/S key in order to start the test
With the Temperature Conversion, to name another, you can input a value and it locks immediately after that

When used for standard operations it works with no problem

I may be wrong but with a number on the screen just typed, pressing STO will make
a .00 appears after the digit, but then pressing RCL won't recall the number.

Thanks for any idea you may want to share in roder to possibly fix it
Let us know if we can upload anything that can be of help

Take care !

[teenix]

(11-15-2020 12:13 AM)albertofenini Wrote:  Hi everyone

this is a unit that we restored a while ago but never truly tested
Few days ago we replaced a coil following the advice of Teenix (thanks for help !) and fixed an issue that was causing the G segment to be always off (the related coil was open)

We discovered that when a card is read, a steady 0.00 appears, but when we try to execute the loaded program, the unit goes in a non responsive status.

With the diagnostic card, to name one, it happens just hitting the R/S key in order to start the test
With the Temperature Conversion, to name another, you can input a value and it locks immediately after that

When used for standard operations it works with no problem

I may be wrong but with a number on the screen just typed, pressing STO will make
a .00 appears after the digit, but then pressing RCL won't recall the number.

Thanks for any idea you may want to share in roder to possibly fix it
Let us know if we can upload anything that can be of help

Take care !

My emulator shows the entered number plus [00] on the display after STO is pressed if FIX 2 is selected.

Are you pressing (say) STO 5, then use RCL 5 to recall it.

Can you write a program and switch to RUN mode and execute it? Does it read with a known good CPU board?

If so, then problem could be the card reader, the CRC chip in the hybrid, the usual copper traces, connection from card reader to CPU board etc.

I guess invalid data could be read from the card which confuses the CPU board when the code tries to run.

cheers

Tony

[Kostas Kritsilas]

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.

[albertofenini]

Hi Tony
this is what we tested :

Storage Register : these all works well
Yesterday it was me using them wrongly
You can store in any of the 1 to 9 registers, you can recall the values, and you can do storage register arithmetic as well

Writing a program : we followed the example on the quick reference guide,
LBL E ENTER ENTER x x RTN, and it works, we even wrote it on a card,
turned off the unit, read it from the card and it worked, it was also read from
other units with no problems, I would say that the card reader works fine ...

Standard Pac programs : some programs like Day of the Week, Temp Conversion, Weight conversion are read with no problem and run, most like the Diagnostic don't
This behavior is consistent also using other Standard Pacs

Then, we have put into this assembly one known working hybrid board and everything worked perfectly, no problems at all
I would say the card reader and the led/keys boards are ok.

On the opposite, we have put this hybrid board in another unit and all the wrong behaviors were exactly repeated

I would say that the problem is within the logic board at this point ...

Could it be a ROM/RAM problem or one of the chips inside the Hybrid is failing ?
We have a set of the ROMs, would you suggest to try to replace them ?
All, one in particular ?

Thanks for help !!!

(11-15-2020 03:40 AM)teenix Wrote:  

(11-15-2020 12:13 AM)albertofenini Wrote:  Hi everyone

this is a unit that we restored a while ago but never truly tested
Few days ago we replaced a coil following the advice of Teenix (thanks for help !) and fixed an issue that was causing the G segment to be always off (the related coil was open)

We discovered that when a card is read, a steady 0.00 appears, but when we try to execute the loaded program, the unit goes in a non responsive status.

With the diagnostic card, to name one, it happens just hitting the R/S key in order to start the test
With the Temperature Conversion, to name another, you can input a value and it locks immediately after that

When used for standard operations it works with no problem

I may be wrong but with a number on the screen just typed, pressing STO will make
a .00 appears after the digit, but then pressing RCL won't recall the number.

Thanks for any idea you may want to share in roder to possibly fix it
Let us know if we can upload anything that can be of help

Take care !

My emulator shows the entered number plus [00] on the display after STO is pressed if FIX 2 is selected.

Are you pressing (say) STO 5, then use RCL 5 to recall it.

Can you write a program and switch to RUN mode and execute it? Does it read with a known good CPU board?

If so, then problem could be the card reader, the CRC chip in the hybrid, the usual copper traces, connection from card reader to CPU board etc.

I guess invalid data could be read from the card which confuses the CPU board when the code tries to run.

cheers

Tony

[teenix]

(11-15-2020 11:37 AM)albertofenini Wrote:  Hi Tony
this is what we tested :

Storage Register : these all works well
Yesterday it was me using them wrongly
You can store in any of the 1 to 9 registers, you can recall the values, and you can do storage register arithmetic as well

Writing a program : we followed the example on the quick reference guide,
LBL E ENTER ENTER x x RTN, and it works, we even wrote it on a card,
turned off the unit, read it from the card and it worked, it was also read from
other units with no problems, I would say that the card reader works fine ...

Standard Pac programs : some programs like Day of the Week, Temp Conversion, Weight conversion are read with no problem and run, most like the Diagnostic don't
This behavior is consistent also using other Standard Pacs

Then, we have put into this assembly one known working hybrid board and everything worked perfectly, no problems at all
I would say the card reader and the led/keys boards are ok.

On the opposite, we have put this hybrid board in another unit and all the wrong behaviors were exactly repeated

I would say that the problem is within the logic board at this point ...

Could it be a ROM/RAM problem or one of the chips inside the Hybrid is failing ?
We have a set of the ROMs, would you suggest to try to replace them ?
All, one in particular ?

Thanks for help !!!

Because the fault shifts with the CPU card, the only thing I can think of is that the 600 bit program storage unit or card reader control circuit (CRC) inside the hybrid may have some faults.

If this is the case, then it can only be determined by changing the hybrid chip module, or having a close look at the PCB tracks and pins associated with the card reader control signals.

I don't think it is a ROM/RAM problem because unlike the 67, the card read/write is a separate system, but you never know.

I'm guessing here, but from what I understand, the HP-65 is a bit sensitive to timing problems, (say) different card speeds compared to the CPU clock speeds or vice versa. I don't know if this is the issue, but reading long programs from the cards might have errors because of a timing imbalance, whereas short programs might run ok.

This might be determined by writing (say) a 30 step program to a card and reading it back. If it works, try a 50 step, 70 step, then a 90 step and finally a 100 step program. If things begin to fail with longer programs then it might be a timing issue, but as I say, just guessing.

You can check the card by monitoring the head switch and the RA/RB data. The data should clearly appear inside the points where the switch operates.


cheers

Tony

[albertofenini]

Thanks Tony,

do you mean that if we put a probe on the HDS pin we should see the flows of data
and see where it stops ?

Do we need to check also RA and RB ?

take care !

(11-15-2020 01:10 PM)teenix Wrote:  You can check the card by monitoring the head switch and the RA/RB data. The data should clearly appear inside the points where the switch operates.

cheers

Tony

[teenix]

I attached an image explaining the data flow while writing a card.

You can see the data on WA WB is between the head switch going low and then high as the card enters and exits the mechanism.

The RA RB data would be similar for a read.

cheers

Tony

(11-16-2020 09:07 AM)albertofenini Wrote:  Thanks Tony,

do you mean that if we put a probe on the HDS pin we should see the flows of data
and see where it stops ?

Do we need to check also RA and RB ?

take care !

(11-15-2020 01:10 PM)teenix Wrote:  You can check the card by monitoring the head switch and the RA/RB data. The data should clearly appear inside the points where the switch operates.

cheers

Tony

[albertofenini]

Thanks Tony,

I need to figure out how to compare more than two channels since our scope has only two ...
How are related the length of the program to the length of this waves ?

I understand this : there are two channels on the head named A and B
from the head these go into the sense amplifier at pins HAC (connected to CA) and HA
and HBC (connected to the CB) and HB

The sense amplifier is responsible for handling the motor functioning and amplify those
signals before these get to the hybrid in the pins WA, EB, RA and RB

WA and WB should be for Write channel A and write Channel B
RA and RB should be for Read Channel A and read Channel B
but what is WE for ?

What documentation could we read (easy to understand) to better understand the functioning of this ?

For instance, if W or R signal is dirty, how can we track it back to the head and try to troubleshoot ?

Thanks a lot for your patience and your help, take care

(11-16-2020 09:22 AM)teenix Wrote:  I attached an image explaining the data flow while writing a card.

You can see the data on WA WB is between the head switch going low and then high as the card enters and exits the mechanism.

The RA RB data would be similar for a read.

cheers

Tony

(11-16-2020 09:07 AM)albertofenini Wrote:  Thanks Tony,

do you mean that if we put a probe on the HDS pin we should see the flows of data
and see where it stops ?

Do we need to check also RA and RB ?

take care !

[teenix]

(11-16-2020 09:46 AM)albertofenini Wrote:  but what is WE for ?

What documentation could we read (easy to understand) to better understand the functioning of this ?

WE is write enable. In the last image you can see that the signal went low to tell the sense chip that a write is required. That signal stays high for a read.

The low signal tells the sense chip that its internal buffers are going to be used to get data from the CRC chip and write it to the card.
The high signal tells the sense chip that its internal buffers are going to be used to get data from the card and send it to the CRC chip.

The image came from the Classic Notes PDF and there is some more info and oscilloscope traces for card read/writes.

cheers

Tony

[albertofenini]

Thanks Tony
I think I only have the Display Section of that document
Can I have it all ?
I understand there is a lot of effort put in that document and I would truly understand if it's not coming for free
take care !

(11-16-2020 10:12 AM)teenix Wrote:  

(11-16-2020 09:46 AM)albertofenini Wrote:  but what is WE for ?

What documentation could we read (easy to understand) to better understand the functioning of this ?

WE is write enable. In the last image you can see that the signal went low to tell the sense chip that a write is required. That signal stays high for a read.

The low signal tells the sense chip that its internal buffers are going to be used to get data from the CRC chip and write it to the card.
The high signal tells the sense chip that its internal buffers are going to be used to get data from the card and send it to the CRC chip.

The image came from the Classic Notes PDF and there is some more info and oscilloscope traces for card read/writes.

cheers

Tony

[teenix]

(11-16-2020 10:18 AM)albertofenini Wrote:  Thanks Tony
I think I only have the Display Section of that document
Can I have it all ?
I understand there is a lot of effort put in that document and I would truly understand if it's not coming for free
take care !

The PDF document is here...

http://teenix.org/ClassicNotes.pdf

The card info is towards the end.

All free and I'll add more as I can :-)

cheers

Tony

[albertofenini]

Thank you so much !!!
Let's go back to school now
Take care !

(11-16-2020 10:28 AM)teenix Wrote:  

(11-16-2020 10:18 AM)albertofenini Wrote:  Thanks Tony
I think I only have the Display Section of that document
Can I have it all ?
I understand there is a lot of effort put in that document and I would truly understand if it's not coming for free
take care !

The PDF document is here...

http://teenix.org/ClassicNotes.pdf

The card info is towards the end.

All free and I'll add more as I can :-)

cheers

Tony

[albertofenini]

It is very rich of information and quite easy to understand
I have a few questions, if it's ok for you I'll post them here
may be this discussion can be of help for some other people
thanks again take care !
still I have tons of question

(11-16-2020 10:28 AM)teenix Wrote:  The PDF document is here...

http://teenix.org/ClassicNotes.pdf

The card info is towards the end.

All free and I'll add more as I can :-)

cheers

Tony

[teenix]

Absolutely, I think your continuing HP-65 problems are providing everyone with an interesting source of information.

cheers

Tony

(11-16-2020 12:04 PM)albertofenini Wrote:  It is very rich of information and quite easy to understand
I have a few questions, if it's ok for you I'll post them here
may be this discussion can be of help for some other people
thanks again take care !
still I have tons of question

[albertofenini]

Thank you so much Tony

From what I read this is what I understand :
When a card is inserted the MTR switch is closed and the CRC in the hybrid (?) gives power to the motor thru the Sense Amplifier.
Let's keep it simple with focus on the tracks, forget about HDS and WPS for now and suppose we are reading a card

The calculator can keep maximum 100 steps @ 6 bit each one, is this why a card can contain 600 bit at maximum?

Now the head: it has two channels, A and B but for each channel it sends back signals on three pins : HAC, HA, CA and HBC, HB, CB
Also HAC and CA are connected together and both benefit from a 22microF condenser and the same is for HBC and CB

What's the difference between HAC and CA or HBC and CB ?
What's the difference between HA and CA or HB and CB?
Is there a way to test a head ? Anyone knows what the inductance values should be ?

Than I have understood that the Sense Amplifier works these signals what is into RA and RB
These two combined have all the information including this auto clock feature
These are sent to the Hybrid to store the program steps, well the corresponding bits

Is it true that when reading : HAC, HA and CA are converted in RA and HBC, HB and CB are converted in RB ?

Now, when we record : WPS recognize that the card can be written, and suppose WE is coherent with this situation, so it's ok to write.
I have understood that the the Program Steps (bits) are coded into WA and WB and then the sense amplifier send these to the head
thru HAC, HA, CA and HBC, HB HC this time writing magnetizing the card

(BTW what all these acronyms stand for ?)

How far am I from understand the basics f the card reader ?

On some schemes there is also a CRT/P signal (on the card reader connector with the logic board is just below the WE pin)
Is this important ?

Having said this, if I'm not wrong this is the flow of data and the sequence :
when reading :from the head, to the sense, from the sense to the memory
when writing : from the memory to the sense and from the sense to the head

if the WA or the WB are dirty where the heck is this happening and why ?

Should that happen before the Sense ?
Or inside the Sense because it is responsible for taking the information from the Hybrid and coding into WA and WB ?

I think the flow is very important to find out what and where something can possibly go wrong

I may have written many non-sense, I'm sorry for that but it's fascinating how this works

Thanks for help and take care !!

(11-16-2020 12:36 PM)teenix Wrote:  Absolutely, I think your continuing HP-65 problems are providing everyone with an interesting source of information.

cheers

Tony

[albertofenini]

Hi Tony

after reading the document and asking so many questions in the previous post,
we have uploaded two more images, HB and CB during a READ and during a WRITE
but they are from the HP-65 that has a weird display and some card reader issues (WB dirty) and not from this unit.

About this unit we will make more tests with increasing length of programs and see if looking at RA and RB will show what goes wrong ...


Take care !

[teenix]

(11-16-2020 04:08 PM)albertofenini Wrote:  From what I read this is what I understand :
When a card is inserted the MTR switch is closed and the CRC in the hybrid (?) gives power to the motor thru the Sense Amplifier.
No, the CRC tells the sense chip to power on the motor via the [Motor] pin which goes LO for Motor=ON, HI for motor=OFF
The sense chip gets motor power directly from the battery, but the sense chip has a constant current circuit that controls the motor speed.
The external resistor sets the current for this circuit

Let's keep it simple with focus on the tracks, forget about HDS and WPS for now and suppose we are reading a card

The calculator can keep maximum 100 steps @ 6 bit each one, is this why a card can contain 600 bit at maximum?
Yes, it only needs to store 600 bits

Now the head: it has two channels, A and B but for each channel it sends back signals on three pins : HAC, HA, CA and HBC, HB, CB
Also HAC and CA are connected together and both benefit from a 22microF condenser and the same is for HBC and CB

What's the difference between HAC and CA or HBC and CB ?
What's the difference between HA and CA or HB and CB?
See Below

Is there a way to test a head ? Anyone knows what the inductance values should be ?
Only that it should have a small resistance. I don't know inductance

Than I have understood that the Sense Amplifier works these signals what is into RA and RB
These two combined have all the information including this auto clock feature
These are sent to the Hybrid to store the program steps, well the corresponding bits
See HP Journal 1974 Page 4

Is it true that when reading : HAC, HA and CA are converted in RA and HBC, HB and CB are converted in RB ?
Yes

Now, when we record : WPS recognize that the card can be written, and suppose WE is coherent with this situation, so it's ok to write.
WE goes HI for a write, LO for a read. If WPS goes LO before HDS then the CRC decides the card is not write protected

I have understood that the the Program Steps (bits) are coded into WA and WB and then the sense amplifier send these to the head
thru HAC, HA, CA and HBC, HB HC this time writing magnetizing the card
Yes

(BTW what all these acronyms stand for ?)

Guessing, but seems logical
HA HB= Head A Head B
HAC HBC = Head A Capacitor, Head B Capacitor
CA CB = CapacitorA, Capacitor C


How far am I from understand the basics f the card reader ?
Pretty good :-)

On some schemes there is also a CRT/P signal (on the card reader connector with the logic board is just below the WE pin)
Is this important ?
Not sure, I haven't seen this

Having said this, if I'm not wrong this is the flow of data and the sequence :
when reading :from the head, to the sense, from the sense to the memory
when writing : from the memory to the sense and from the sense to the head
Correct

if the WA or the WB are dirty where the heck is this happening and why ?
WA WB come from the CRC chip in the hybrid. It (or the sense chip) could be faulty if the tracks between the CPU board and sense chip are ok

Should that happen before the Sense ?
Or inside the Sense because it is responsible for taking the information from the Hybrid and coding into WA and WB ?
You could disconnect the card reader connector and measure WA WB when writing a card

I think the flow is very important to find out what and where something can possibly go wrong

I may have written many non-sense, I'm sorry for that but it's fascinating how this works
All good, see the HP Journal 1974 Page 4 onwards

Thanks for help and take care !!

cheers

Tony

[albertofenini]

Thanks Tony !
Couple of questions more in-line below ...
We need you to realize a replacement for the old chips
Thanks again and take care !!

(11-17-2020 07:06 AM)teenix Wrote:  

(11-16-2020 04:08 PM)albertofenini Wrote:  From what I read this is what I understand :
When a card is inserted the MTR switch is closed and the CRC in the hybrid (?) gives power to the motor thru the Sense Amplifier.
No, the CRC tells the sense chip to power on the motor via the [Motor] pin which goes LO for Motor=ON, HI for motor=OFF
The sense chip gets motor power directly from the battery, but the sense chip has a constant current circuit that controls the motor speed.
The external resistor sets the current for this circuit

Let's keep it simple with focus on the tracks, forget about HDS and WPS for now and suppose we are reading a card

The calculator can keep maximum 100 steps @ 6 bit each one, is this why a card can contain 600 bit at maximum?
Yes, it only needs to store 600 bits

Now the head: it has two channels, A and B but for each channel it sends back signals on three pins : HAC, HA, CA and HBC, HB, CB
Also HAC and CA are connected together and both benefit from a 22microF condenser and the same is for HBC and CB

What's the difference between HAC and CA or HBC and CB ?
What's the difference between HA and CA or HB and CB?
See Below

Is there a way to test a head ? Anyone knows what the inductance values should be ?
Only that it should have a small resistance. I don't know inductance

Than I have understood that the Sense Amplifier works these signals what is into RA and RB
These two combined have all the information including this auto clock feature
These are sent to the Hybrid to store the program steps, well the corresponding bits
See HP Journal 1974 Page 4

Is it true that when reading : HAC, HA and CA are converted in RA and HBC, HB and CB are converted in RB ?
Yes

Now, when we record : WPS recognize that the card can be written, and suppose WE is coherent with this situation, so it's ok to write.
WE goes HI for a write, LO for a read. If WPS goes LO before HDS then the CRC decides the card is not write protected

I have understood that the the Program Steps (bits) are coded into WA and WB and then the sense amplifier send these to the head
thru HAC, HA, CA and HBC, HB HC this time writing magnetizing the card
Yes

(BTW what all these acronyms stand for ?)

Guessing, but seems logical
HA HB= Head A Head B
HAC HBC = Head A Capacitor, Head B Capacitor
CA CB = CapacitorA, Capacitor C


How far am I from understand the basics f the card reader ?
Pretty good :-)

On some schemes there is also a CRT/P signal (on the card reader connector with the logic board is just below the WE pin)
Is this important ?
Not sure, I haven't seen this

Having said this, if I'm not wrong this is the flow of data and the sequence :
when reading :from the head, to the sense, from the sense to the memory
when writing : from the memory to the sense and from the sense to the head
Correct

if the WA or the WB are dirty where the heck is this happening and why ?
WA WB come from the CRC chip in the hybrid. It (or the sense chip) could be faulty if the tracks between the CPU board and sense chip are ok
if the same logic boards works well with another card reader this excludes the CRC inside the Hybrid and leave as faulty the Sense amplifier ?

Should that happen before the Sense ?
Or inside the Sense because it is responsible for taking the information from the Hybrid and coding into WA and WB ?
You could disconnect the card reader connector and measure WA WB when writing a card
We tried last night but when the card reader is totally disconnected the motor will not run and the card will not move, we'll try to disconnect everything but the motor

I think the flow is very important to find out what and where something can possibly go wrong

I may have written many non-sense, I'm sorry for that but it's fascinating how this works
All good, see the HP Journal 1974 Page 4 onwards

Thanks for help and take care !!

cheers

Tony

[teenix]

(11-17-2020 10:21 AM)albertofenini Wrote:  if the same logic boards works well with another card reader this excludes the CRC inside the Hybrid and leave as faulty the Sense amplifier ?

We tried last night but when the card reader is totally disconnected the motor will not run and the card will not move, we'll try to disconnect everything but the motor

I would think so, or the sense PCB and or connectors.

------

Yes that would be the case because the sense PCB is missing, but if you eased a non write protected card in just enough, the switches would operate and the CRC would think a card is going though and you should be able to capture the WA WB data on your scope.

Note that I haven't tried this, but as there is no feedback to the CRC except the switches it should work.

You will get an Error on display because the CRC wont see the card exit in time and think it is jammed.

Another thing you could do, is to remove the WA and WB pins on the card reader PCB so they are open circuit to the sense chip when the sense board is connected to the CPU. Then see if the WB signal improves.


cheers

Tony

[teenix]

(11-16-2020 11:51 PM)albertofenini Wrote:  Hi Tony

after reading the document and asking so many questions in the previous post,
we have uploaded two more images, HB and CB during a READ and during a WRITE
but they are from the HP-65 that has a weird display and some card reader issues (WB dirty) and not from this unit.

About this unit we will make more tests with increasing length of programs and see if looking at RA and RB will show what goes wrong ...


Take care !

Something definitely wrong with the WB signal.

cheers

Tony