HP-19C Repair Odysee

[PANAMATIK]

The "newPIK" schematic is ready!

As always there was a tradeoff between possibilities and reality.

The disadvantages:

1.) The newPIK will not be working standalone with the original ACT chip. For now it is only usable for repair of an HP-19C together with the newACT. The newACT repair kit hardware is unchanged when programmed as HP-19E. Using the newPIK in an HP-97 requires to use also a newACT for the HP-97, which I still have to develop.
2.) The newPIK relies on the original display driver chips, therefore there is no alphanumeric LED display possible.

But there are some advantages:

1.) Very easy to install. Just remove the original ACT and PIK chip and insert the new circuits.
2.) There will be a Real Time Clock which can display time, date, day, alarm time, stopwatch.
3.) There is a Low Power mode. When you switch OFF the calculator the state of all registers and programs will be retained.
4.) You can store many user programs in flash memory. The original HP-19C/29C application programs are preprogrammed.
5.) There is a beeper for alarm time and stopwatch countdown alarm.
6.) Display of battery voltage and temperature



Why did I not develop a PIK chip, which can be used in any HP-19C and HP-9X calculator with the original ACT?

In my opinion this is impossible with any of the 5V PIC or ATMEL devices available. The reason is the timing of the synchrounous ISA and DATA signals of the ACT. These signals are synchrounous bidirectional communication lines between ACT and PIK, where the ACT is master and PIK is slave. If a PIK instruction is output on the ISA line, the PIK has to decode the instruction and react before the next clock arrives. This is only 5,5us. No way with 4-8MIPS interrupt driven software. This can be done by a FPGA but not with a small processor.

But why then did I not go for a FPGA solution?

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