HP-65 motor always on - card reader

[Trevor Gale]

Hello again - this time I am being confused by my HP-65 which I am trying to repair - and from an electronics engineers' view it is a puzzle. Again! All the usual is completed, no gummy wheel, no loose soldered wires, etc.

The problem is that when the calculator is switched on, the reader motor immediately starts and never stops. Everything else in the calculator works (there was a keyboard error but that was due to a small break in one of the key strips, and that's repaired).

First-off, on opening the unit I temporarily replaced the motor with a 56-ohm resistor to (a) provide a load and (b) to avoid over-running th poor motor!).

Then I checked the contacts for the card finger-switches MTRS, WPS and HDS and they function as expected and in order (while moving a card in, first checked with a multimeter to ground, no supply, then with battery in and voltages measured).

Here's the puzzling part - (conditons: calculator in 'run' mode, no card inserted, no key pressed) with the CRC interface board plugged into the main CPU board as normal, upon power-up, the motor supply from the 'sense chip' pin 1 goes to +2.56V (which would drive the motor). Measuring the MTRS signal at pin 5 of the hybrid, pin 9 of the CPU, pin 2 of the CRC interface board, sits at +3.7V = HI, and WPS & HDS are also HI. VSS to the 'sense' chip is at +6V as expected, and the 'ground' pins are at 0V as expected.

Pushing a card in pulls the MTRS, WPS and HDS signals LO as they should. However, doing the same with the interface board unplugged from the CPU board does NOT result in the motor drive voltage being sent out. To be clear, I checked further with VSS and GND wired to the reader interface board, but I still find that MTRS is now at a LO level (0.49V) from the CPU board and is also LO on the interface board.

In the first case, pulling HDS low does not stop the motor drive voltage (I wondered if that transition would be latched by some other logic) yet pulling MTRS low on the interface card does NOT cause the motor drive voltage being set to the +/- 2.5V either!

Does anyone know if there is some kind of interaction with the data lines and the switching of the motor? One thing I have not yet done is to determine if the 'sense' chip is pulling the MTRS signal high or the CPU board is doing it - I have to cut a trace on the interface board to find that out, which I might do in the meantime.

Any ideas?

[teenix]

The MTRS switch (I think) does 2 things. It lets the CRC chip know if the motor is running or not and it also tells the sense chip to turn the motor on or off. OFF being a HI signal, ON being a LO signal on the MTRS line.

Both of the MTRS pins on the CRC and Sense chip are inputs, however it seems in your case there is some interaction between them causing the MTRS line to (perhaps) pull LO internal to the sense chip, even though you can read HI or LO externally. Maybe the isolation is affecting other pins like RA RB WA and WB which is where the fault may be coming from. It is easy to unsolder and resolder the connecting pins between the two boards.

I don't know which chip has the pullup function on the MTRS line. From isolating the sense board, and the resulting LO level at the sense and Hybrid, it seems it is the sense chip.

As far as I know there is no latching logic on the MTRS line, it is just a switched HI or LO signal.

cheers

Tony

[Trevor Gale]

I think there is a pull-up resistor internal to the chip for MTRS, and that's what might be at fault (or it simply relies upon the input already being pulled up by another resisitor). I have also "mimicked" the head input signals (i.e. two separate 180mV signals over two 47-ohm resistors with a 1.35V bias) connected in place of the real head connections) and found that the comparator outputs RA and RB do not toggle correctly either. So it looks as if the 'sense' chip may be fried.

I am wondering if the 'sense' chip is similar to the one used in the HP-41 card readers - I have a spare unassembled HP-41 reader that is well beyond repair but the small electronic boards are undamaged - and the connections seem to be the same (head connectons, VSS, GND, gb etc.) so if it is I could replace the chip on the HP-65 reader board. (I do not have any "spare" HP-67 boards or chips!).

I very much doubt that there is any fault with the CPU board on the HP-65 since all other operations and expected voltages are correct.

Any ideas? And thanks, yet again!

[teenix]

As the pin outs are the same, I guess it can't hurt to try the spare sense chip.

According to the HP Journal, there are latches in the sense IC for reading data. I don't know if these get confused if the input data is not alternating properly, or whether it is based just of the comparator switching.

cheers

Tony

[Trevor Gale]

Thanks Tony,

I shall double-check all of the pin connections on the old HP-41 reader boards to ensure that this chip has identical routing.

The HP-65 CRC chip is 1826-0158, while the HP-41 chip is 1826-0322 and I do not know enough about the HP chip numbering schemes to be sure yet.

If I manage to replace the chip this afternoon (here it is 13:43 hrs) I will come back and tell how I got on!

Kind regards - Trevor.

[teenix]

The chip pinouts are the same, maybe there are some internal changes.

Circuit-wise, these aren't on the 65, but there are some diodes to drop the battery voltage on the 41 for the VCC pin, and some pullup resistors on the RA RB pins, and series resistors on the WA and WB pins.

cheers

Tony

[Trevor Gale]

Hi Tony,

Yes, I also looked up the various chip info, and the connections, and functions, are the same.

Interestingly, from the HP-41 service manual from HP, the card reader circuit uses 1826-0322 for the 'sense' chip; however this same chip is also used in the HP-67. In both the '67 and the '65, pin 1 of this chip is the +ve supply to the motor, supplying the +/- 180mA for it - whereas in the HP-41 cardreader that pin drives a LED which couples to a phototransistor which (in parallel with the HDS switch-to-ground line) then activates a timer which actually drives the motor through an emitter-follower.

I do not see any problem with using this chip - and I will today attempt the replacement and get back to you.

[Trevor Gale]

"Bad news but good news" - the HP-41 cardreader 'sense' chip seems to be able to be used on the HP-65, HP-67 as well as the 41's. Anyway:-

Well, I did finally replace the 'sense' chip, the old one was easy to desolder but it had very clearly been 'fried' - as I gently extracted it the top of the chip came completely loose, revealing all of the internal gold links to the chip die (some loose)!!! I have kept it and I will try to make a macro microphotograph of it!!!

With the new one in, and the board inserted onto the CPU board, I checked voltages and the input bias (1.41V), installed it and ran it with a 56-ohm load instead of the motor to check, and it all worked as expected (with error due to no head input of course).

Before I connected the wiring to the motor, I applied my bench supply at +2.3V (a typical level) to the motor to check that it would pull a card through correctly. However it did not - it ran but very slowly, which was because it was trying to draw 235mA - which I had set as the current-limit on my bench supply... I took out the card and then set the current-limit to 275mA and the motor ran slightly faster but drew all that 275mA. I had a new tantalum capacitor across the motor so that was not the problem - the motor had obviously been damaged by the previous 'sense' chip driving the motor for far too long. Probably what 'fried' the chip in the first place.

I already set the eccentric pin to what felt like the 'softest' drive wheel pressure; I then extracted the motor and checked again on its own, and it still tried to draw excessive current. Thus I am now looking for a new HP-65 reader motor - probably better to try and find a 'broken' HP-65 to take the motor from (in my experience they do not normally seem to fail). So I put this HP-65 aside until I find such a solution - the motor in the HP-65 has differently-spaced mounting screwthread holes to the motors used in the HP-67 and HP-41 series readers..

All is NOT LOST however - I had a 3rd HP-67 with a bad sense chip in it (no data outputs on RA or RB) which I have on my desktop as a "no-card" calculator for occasional use. Since the card reader interface board is the same as the HP-65, I took my 'newly-chipped' interface board and wired that up to this HP-67, installed it, checked voltages etc., wired in the motor, and it seemed to run first-time.

I wrote several simple programs to cards and successfully read all of them back in, then checked that they could all be read by my other HP-67's (they could, no problem) - then read in a 'standard pac' card which also read in okay. Programs written to cards on the other HP-67's were also read in okay in this HP-67 too.

SO I have not solved my HP-65 successfully (yet) but I HAVE gained a fully working spare HP-67 with working card reader!

[teenix]

Thats great.

I have seen where a motor has been loaded too much and it stresses the sense chip which gets hot. I guess if this is repeated enough then eventually the chip and/or motor will fail.

Improper positioning of the worm gear on the motor shaft can cause excessive motor wear and possible increased loading. Someone might experience this and use the trimpot to speed up the motor to compensate and this could lead to premature failure.

cheers

Tony