Threaded Mode | Linear Mode

teenix · 10-21-2021, 01:50 AM

Hi all,

Well, the latest (last) board design has been submitted to the fabricator.

The board has single processor, printer and card interfaces, storage for 800+ cards, Continuous Memory, programmable print speed and intensity, battery backed alarm clock with 5ppm accuracy and piezo buzzer, Bluetooth, HP-97S style interface via electrically isolated serial port, direct plug in replacement - no soldering, re-flash capability and no need for the original printer interface card.

The PIC had one spare output remaining so I added an IR diode for an external printer. I have written the IR code and it seems to be in spec on the scope. I had to create a new print ROM for it as the HP82240 printer does not have the HP-97 characters like [sqrt] etc. You can also use an emulated 82240 style printer via Bluetooth on the PC, where you can store print lists and print hard copies on a standard printer. The font is similar to 97 printer. These interfaces were quite tricky as the 97 printer gets a line of data in spurts all terminated by end of line characters with no easy way to determine what is coming next.

Fingers crossed the new board has no problems.

cheers

Tony

hpux735 · (This post was last modified: 10-21-2021 03:12 AM by hpux735.)

That's amazing!

(apparently this forum software does not like emoji Wink

)

Massimo Gnerucci · 10-21-2021, 05:50 AM

(10-21-2021 01:50 AM)teenix Wrote: The board has single processor, printer and card interfaces, storage for 800+ cards, Continuous Memory, programmable print speed and intensity, battery backed alarm clock with 5ppm accuracy and piezo buzzer, Bluetooth, HP-97S style interface via electrically isolated serial port, direct plug in replacement - no soldering, re-flash capability and no need for the original printer interface card.

The PIC had one spare output remaining so I added an IR diode for an external printer. I have written the IR code and it seems to be in spec on the scope. I had to create a new print ROM for it as the HP82240 printer does not have the HP-97 characters like [sqrt] etc. You can also use an emulated 82240 style printer via Bluetooth on the PC, where you can store print lists and print hard copies on a standard printer. The font is similar to 97 printer. These interfaces were quite tricky as the 97 printer gets a line of data in spurts all terminated by end of line characters with no easy way to determine what is coming next.

Enough to revive the dead one I have! ;-)

Thank you once more, Tony.

teenix · 10-21-2021, 07:46 AM

Hi all,

I've been trying to dissect the sense chip for some time now, and have been making some progress.

The HP Journal mentions it as a Large Scale Integrated (LSI) circuit, which is a bit dauting as it probably means lots and lots of transistors were used to make it do its job. I'm pretty sure it is not a digital chip, it is most likely all analog with some level shifting to interface to the PMOS CRC chip.

I am about to try and put a circuit together to mimic the card read/write which will take a bit of time and fiddling but I can use one of my CPU boards and one of the new blank sense cards that I have as a programmable test bed. I just have to be very careful that I don't fry the head Undecided

cheers

Tony

TomC · 10-21-2021, 03:44 PM

Tony:

Don't be too 'daunted' - a 'Large Scale Integrated Circuit' of the 1970s may be quite minimal in the 21st century!

TomC

(10-21-2021 07:46 AM)teenix Wrote: Hi all,

I've been trying to dissect the sense chip for some time now, and have been making some progress.

The HP Journal mentions it as a Large Scale Integrated (LSI) circuit, which is a bit dauting as it probably means lots and lots of transistors were used to make it do its job. I'm pretty sure it is not a digital chip, it is most likely all analog with some level shifting to interface to the PMOS CRC chip.

I am about to try and put a circuit together to mimic the card read/write which will take a bit of time and fiddling but I can use one of my CPU boards and one of the new blank sense cards that I have as a programmable test bed. I just have to be very careful that I don't fry the head

cheers

Tony

teenix · 10-21-2021, 08:24 PM

(10-21-2021 03:44 PM)TomC Wrote: Tony:

Don't be too 'daunted' - a 'Large Scale Integrated Circuit' of the 1970s may be quite minimal in the 21st century!

TomC

OK, Just a little bit then ;-)

I've cobbled together a circuit for the read/write and if it does what I hope then it should be a single chip solution. This will be good as there is one taxing problem, there is not much room where the original sense chip resides and this limitation may kill off any chance of a plug in replacement. There are also a couple of sense chip types and I'm not sure of the differences.

I need some odd value parts to experiment with the circuit, so I will try to put something together when I can.

The worlds longest lockdown here has begun to lift today, so I can start to get some freedom :-)

cheers

Tony

hpux735 · 10-21-2021, 08:32 PM

On the topic of card reading, Tony, I've been looking at the contents of the "hpp" files you have in the emulator. What's the format of those files? It doesn't quite look as I'd expect. Are they 6-bit words concatenated together?

I've collected the trace from each of the cards in the stat pac-1 using my saleae.

They're in the Saleae save file format and CSV export here: https://gitlab.com/hpux735/hp97-cards/-/.../resources

I'm working on a program now to parse the data and make the bitstream. I was hoping to validate the scanning, and my logic, using the card files you already have.

teenix · 10-21-2021, 11:51 PM

(10-21-2021 08:32 PM)hpux735 Wrote: On the topic of card reading, Tony, I've been looking at the contents of the "hpp" files you have in the emulator. What's the format of those files? It doesn't quite look as I'd expect. Are they 6-bit words concatenated together?

I've collected the trace from each of the cards in the stat pac-1 using my saleae.

They're in the Saleae save file format and CSV export here: https://gitlab.com/hpux735/hp97-cards/-/.../resources

I'm working on a program now to parse the data and make the bitstream. I was hoping to validate the scanning, and my logic, using the card files you already have.

I have put the card file format in the Classic Notes PDF file.

This may change at some stage to simplify them a bit although the emulator will still recognize both types.

cheers

Tony

hpux735 · 10-22-2021, 12:55 AM

Oh, cool! That's a newer version of the notes file than I had before Smile

teenix · (This post was last modified: 10-23-2021 01:03 AM by teenix.)

Hi all,

The 97 boards have been manufactured, should be here early next week.

Meanwhile, I have been playing with circuitry for the sense head writes.

The attached image shows the normal signals at the head at top, and the lower is what I created from a mock up. The first pulse has a smaller amplitude than the others for some reason. This will probably be a problem for the first data bit not recording properly on the card. The rest of them are a bit higher in amplitude, but it looks the part so far.

Unfortunately, I am about out of my allocation of storage space on the Forum again, so I am going to have to start deleting more stuff to make room. I could be cheeky and reinvent myself and keep going in a different guise I suppose ;-)

cheers

Tony

hpux735 · 10-23-2021, 01:58 AM

Those waveforms look excellent! It'll be interesting to see how well those cards are read.

The forum space allocation is kinda crazy, to be honest. Most hosting these days lets you have unlimited space. Even then it's a few hundred dollars a year.

hpux735 · 10-25-2021, 08:56 PM

Hi Teenix!

I've made a lot of process reading cards, and comparing them to your hpp format. Thanks for adding that to the document. I'm running into an issue now with the checksum computation. Would you mind taking a look at this post and proving some feedback?

https://housedillon.com/posts/Reading-cards-hp-97/

Thanks!

(10-21-2021 11:51 PM)teenix Wrote:
(10-21-2021 08:32 PM)hpux735 Wrote: On the topic of card reading, Tony, I've been looking at the contents of the "hpp" files you have in the emulator. What's the format of those files? It doesn't quite look as I'd expect. Are they 6-bit words concatenated together?

I've collected the trace from each of the cards in the stat pac-1 using my saleae.

They're in the Saleae save file format and CSV export here: https://gitlab.com/hpux735/hp97-cards/-/.../resources

I'm working on a program now to parse the data and make the bitstream. I was hoping to validate the scanning, and my logic, using the card files you already have.

I have put the card file format in the Classic Notes PDF file.

This may change at some stage to simplify them a bit although the emulator will still recognize both types.

cheers

Tony

teenix · 10-26-2021, 01:12 AM

(10-25-2021 08:56 PM)hpux735 Wrote: Hi Teenix!

I've made a lot of process reading cards, and comparing them to your hpp format. Thanks for adding that to the document. I'm running into an issue now with the checksum computation. Would you mind taking a look at this post and proving some feedback?

https://housedillon.com/posts/Reading-cards-hp-97/

Thanks!

I've updated the Classic Notes with this info in it.

cheers

Tony

hpux735 · 10-26-2021, 01:43 AM

Thanks!

This is enough to keep me going, for sure.

(10-26-2021 01:12 AM)teenix Wrote:
(10-25-2021 08:56 PM)hpux735 Wrote: Hi Teenix!

I've made a lot of process reading cards, and comparing them to your hpp format. Thanks for adding that to the document. I'm running into an issue now with the checksum computation. Would you mind taking a look at this post and proving some feedback?

https://housedillon.com/posts/Reading-cards-hp-97/

Thanks!

I've updated the Classic Notes with this info in it.

cheers

Tony

hpux735 · 11-01-2021, 02:43 PM

I've got all the card data (other than the two cards that had missing bits, that's a project for another day) loading into my program, and I've been able to "disassemble" the card opcodes into instructions. Also, I can read and write hpp files (including from Saleae capture files).

I've documented that in my third post...

https://housedillon.com/posts/disassembl...ram-cards/

teenix · 11-01-2021, 05:21 PM

(11-01-2021 02:43 PM)hpux735 Wrote: I've got all the card data (other than the two cards that had missing bits, that's a project for another day) loading into my program, and I've been able to "disassemble" the card opcodes into instructions. Also, I can read and write hpp files (including from Saleae capture files).

I've documented that in my third post...

https://housedillon.com/posts/disassembl...ram-cards/

Nice read.

I didn't know you wanted the program codes for the 67/97 or I could have listed them for you. More fun finding out though :-)

cheers

Tony

hpux735 · 11-01-2021, 05:24 PM

Oh, absolutely. I had a great time.

I also found this appnote from Microchip about card reading.

https://ww1.microchip.com/downloads/en/D...0Guide.pdf

teenix · 11-01-2021, 05:35 PM

Hi all,

There appears to be an overseas mail delay problem so I still haven't seen the new 97 board yet.

While I'm waiting I have been fiddling with the card read/write. I managed circuitry to mimic the write signals and an adjustable speed motor controller, so lately I've been looking into the reading.

This is more complex as the signals are quite small from the magnetic head and are easy to get lost in noise. As I mentioned, I'm no analog wizz but I've managed to get the read signals direct from a card.

The time bases are different but the attached image shows the original AC signal (upper) from a 67 and the one I captured in the mock up circuit (lower). Mine is looking correct and has reasonable amplitude but it has more noise superimposed which is understandable as it is just a breadboard circuit.

Still some more head scratching to go as post processing this signal might give spurious data points. Still, it is encouraging.

cheers

Tony

teenix · 11-04-2021, 06:13 AM

Hi all,

The new 97 CPU boards have finally arrived. It will take some time to assemble it in sections while enabling and testing the single processor code in functional blocks.

Hopefully I did my sums right and Murphy has gone on holidays for awhile.

cheers

Tony

Massimo Gnerucci · 11-04-2021, 06:37 AM

(11-04-2021 06:13 AM)teenix Wrote: Hi all,

The new 97 CPU boards have finally arrived. It will take some time to assemble it in sections while enabling and testing the single processor code in functional blocks.

Hopefully I did my sums right and Murphy has gone on holidays for awhile.

cheers

Tony

Excellent! I just bought another problematic 97... ;-)