Posts: 1,841
Threads: 54
Joined: Jul 2005
Hi,
great work!
For the 42S I/O : Yes, via an indirect proof. Christoph Giesselink demonstrated on the famous Allschwil meetings how to output and input data over a serial connection between the HP-28S and a normal PC terminal program, Hyperterminal in this case.
This should also work with an HP-42S .
Ray
Posts: 774
Threads: 93
Joined: Aug 2005
Hi Jose,
I don't know of any (successful) project allowing a PC<>42s interconnection.
A few years ago, I built a very rudimentary MLDL-like interface by connecting the spare ROM pads to a PIC18F452 microcontroller via a couple of Flex PCB's salveged from LCD display units.
Eventually, this should have allowed full control/access to the RAM as well as an extra space for saturn code expansion.
The HW part was (more or lees) easy; but the required SW interfaces went far beyond my limited programming skills. Also the fact that an HP-42s must be "destroyed" to built such device, represents a serious drawback in order to find beta testers; not to say "user's target market" for the project... :-)
It was very funny nonetheless.
Best wishes.
Diego.
Edited: 26 Mar 2013, 9:45 p.m.
Posts: 355
Threads: 67
Joined: Nov 2005
Some years ago I worked on an RS232C interface for the Lewis chip based calculators like the HP28S and HP42S. This project never leaved the proof of concept state on a HP28S.
It's quite unknown that the Lewis chip output port for the IR-LED is bidirectional. So I made a primitive hardware interface to connect the IR-LED port of a HP28S to the RXD and TXD pins of the RS232C interface. The power supply of this interface is generated over the RTS/TXD pins and in idle state < 4 uA from the calculator battery.
The RS232C protocol 2400 baud, no parity, 8 bit, 2 stop bits (2400,N,8,2) was implemented in software. Because of polling the input pin, standard baudrates over 2400 aren't possible with a Lewis chip at a nominal strobe clock frequency of 1MHz.
I mentioned this project here: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv016.cgi?read=101692
Quote:
7) HP28S serial
What's new on the HP28S serial project. I recognized many problems on running when I tried to port this method for ROM uploading over the Ir port. The main problem was the time delay for creating the RS232C signal. In the version I demonstrated two years ago I hadn't thought about different timing on Saturn commands executed on an even or add address. Because the program is lying in calculator RAM the transmitter routine can begin on an even or odd address depending on the program version and the base address of the complete program. I fixed this in spring this year by executing the time critical part in display RAM which is directly accessed by the internal Saturn bus having equal timing on even and odd address.
Next step is a protocol (XMODEM) layer.
