HP-19C Repair Odysee

[teenix]

(07-06-2021 11:19 AM)Artur - Brasil Wrote:  Well, I checked again the voltages. All are there, but nothing on display.
I checked again and it seems to me that, as I tested display using + probe at left and - at right, the voltages of left should be more positive than right, but my measuring on real circuit shows the inverse!
I'll try again, but I'll give some time. I have other things to work in home...

Many thanks Tony, Teenix, Bernhard, Panamatik. I really like to have the opportunity to learn with you!

Sincelrely best wishes from Brazil !
Artur

Check that pin 10 for 1820-1620 (cathode driver) is definitely grounded. This is the return path for all the LED segments.

If you have a 1K5 resistor on hand perhaps for a temporary connection:

VCC -> Resistor -> single LED anode

Taking care not to short any pins, briefly touch the resistor end to any LED anode pins (one at a time) on the 1818-0432 (pin 1 or 2 or 3 or 4 etc). If you can reach the display better, then any of the anode pins (I think display pins 1, 2, 3, 4 etc)

This should show the same LED segment on for all digits (probably dimly). If so, then the LEDs work and the cathode driver is working. If not, it might suggest the cathode driver or connectors PCB traces etc might be faulty.

cheers

Tony

[Artur - Brasil]

I will check, thanks!

[jjohnson873]

(06-04-2021 09:01 AM)PANAMATIK Wrote:  I would really like to program a PIK with a FPGA, but the available FPGAs with the necessary amount of cells are rather big, with many pins and they have 3,3V or 1,8V signals. They would not tolerante the 6,4V power of the HP calculators. And even if the inputs are 5V tolerant, all output signals must be converted to 5V level. I cannot see a minimalistic solution which finds place on the small area of a 28-pin package.

Bernhard

Bernhard....have you looked at the Microchip/Microsemi 250MHz, Igloo nano series of flash based FPGAs? They come in as small as a 3x3mm BGA package. Unlikely you'd want to use the BGA package, but you could use the 6x6mm QFN package. You would have to add a voltage translation IC to move from the FPGA 3.3v I/O to 6.4v I/O. The Igloo series is also an ultra low power FPGA.
~ Jim J. ~

[PANAMATIK]

(07-21-2021 12:41 PM)jjohnson873 Wrote:  ....have you looked at the Microchip/Microsemi 250MHz, Igloo nano series of flash based FPGAs?

Igloo nano QN48 6x6mmm with 34 I/O would be the smallest to handle. This is bigger than an PIC16Fxxx 4x4mm 36 I/O, and needs a lot of additional ICs: 1.5V regulator 3.3V regulator, power sequencing and level shifters, some bidirectional which needs outputs for switching the direction. Very challenging!

Bernhard

[PANAMATIK]

After a long time has passed, I recently remembered my unfinished HP-19C repair project. Also I remember my words, that I never again will open an HP-19C calculator, a rule which I have broken now.

In the last days I convinced myself to start with a new approach. I dont want to build two separate chips, an HP-19C ACT chip and a PIK chip amymore as in my previous concept, but a combined solution with one PIC16LF19197 processor. This is much easier to handle. I already combined the two softwares from my ACT/PIK approch and now already have a source code for the new processor, which "only" need to be debugged. My concept differs from teenix HP-19C project, which replaces both PCBs (the complete electronics). Instead I will only replace the processor board, the power supply board will be left in place but mainly deactivated. It shall be as simple as possible to remove the old main board with the new PCB. Even for the display contacts I think I have a solution which doesn't need to transfer the original spring contacts to the new PCB.

The new HP-19E board will have following features.

1.) Low Power consumption <20mA when not printing, no power consumption when in Sleep mode
2.) running on 3.3V from batteries (also 4x1.5V Alkalines possible) with low quiescent current LDO regulator
3.) Alphanumeric LED display with program steps shown as plain text
4.) HP-19C 13-digit LED display or compatible Woodstock 12-digit LED Display if the original is not available
5.) printing with the built in thermo printer and to external Infrared printer HP82240B
6.) Real Time Clock with Time Date Alarm
7.) Beeper (also programmable as instruction)
8.) 512kB non volatile SPI Flash or pin compatible Fujitsu FRAM for many user programs
9.) optional HM-10 Bluetooth BLE module on board for PC communication (upload download programs, updates etc.)
10.) optional GPS module on board for satellite communication
11.) Printer intensity can be set by software key sequence

The features will follow the Woodstock LP (HP-29E LP) set of features.
Find attached a preliminary schematic for the new HP-19E main board. All components for the prototype board are ordered and I will start debugging in a few days. If the software works on the prototype board, I will design the PCB layout in a few weeks.

One of the reasons for my new start with the HP-19C repair project is that I had an HP-19C calculator from "Harald" for about two years and still couldn't repait it. Last week I made a new try and replaced the RAM chips by some spare RAM chips from an HP-27 calculator. After a first success I could enter program steps again, but it was not reliably working, then after soldering also the second RAM into the original board, nothing did work any more. Something must have happened with these sensible old chips or the second HP-27 RAM was defective. This calculator also has a power supply problem and needs >5,3V battery voltage to power up.

Now I think a new PCB will possibly give it a new future.

Bernhard

[AndiGer]

Possibly the cause is the HP-27 chips are NMOS?

E.g.: https://www.hpmuseum.org/forum/thread-19...#pid166218

Andi

[PANAMATIK]

(01-15-2024 03:25 PM)AndiGer Wrote:  Possibly the cause is the HP-27 chips are NMOS?

Andi

You are right Mr. Sherlock Holmes! The RAMs of HP-27 are incompatible. This cannot work. I had labeled the package wrong, where I have stored the chips. Then I double checked what I have done and recognized that I soldered the wrong chip HP5061-0459 (???) instead of HP5061-0469 (RAM). See the image attached. The label was only partially visible, but I should have seen it before taking action. I will unsolder the wrong chip and the calculator most likely will at least power up again. The first RAM chip which I replaced was correct.

Thanks
Bernhard