HP97 The journey begins

[Massimo Gnerucci]

Well done Tony! I put it on my shopping list.

[teenix]

(04-25-2024 12:11 PM)rprosperi Wrote:  WOW!!

I've been following this thru its gestation, but honestly missed or forgotten some of the features, or some smaller details of features, but this is quite amazing, Tony, very well done!!

A couple questions:

A. Why does the 25C support fewer program files than the 29C? Given the same platform, one would think they could support the same. Unless simply because the 25C was less capable than a 29C and that's how things must be, which I could understand and support.

B. Can the original power adapter be used to charge the LiPo battery and operate the device?

C. Does the new battery fit in an original, or 3D printed equivalent, battery pac?

D. Purpose of new GSB 49 instruction?

E. It appears installation consists of simply opening the unit, replacing the old board with the new one, and reassembling, is that correct?

Looking forward to one of these...

A. Just the size of the memory device. I thought the 200 or so programs for the 25C would suffice, and as you mention the 29C is more capable, so I thought it may get used more often. I could get larger memory, but I was thinking about cost. The 29C can also load Program or load/save Data files while a program is running so the extra program storage may be of more use.

B. No, the original charger outputs AC and would damage the board. Dave (Dreato) from the Forum has made up suitable USB charge cables for this board, or you can use the original cable with a USB plug which is mentioned in the help file.

C. It should do, I removed the batteries from an original holder and inserted the LiPo easily. There is also a small circuit board to connect the battery to via a plug which also serves as the connection to the CPU board.

D. There was a discussion recently initiated by Namir [HP-25 Emulators suggestion] about adding GSB RTN capability. I added a new code GSB 49 as per the discussion. The instruction causes a jump to step 49 where the program has a GTOnn step allowing the subroutine to appear just about anywhere in a program. There are insufficient steps to have GSB's to a range of steps like the GTO instructions. It only supports one subroutine at a time.

All of the original functionality and code for the emulated models is intact so they will operate normally if desired.

E. I wish that were so, but due to the interconnect pins used by HP in the Woodstocks that is impossible without modifying the keyboard and display circuit boards. I have not been able to find replacements for these pins although I saw some that came close. Unfortunately either way, soldering is required. I know not everyone can solder, so I have offered, mentioned in the help file, to do the soldering if the pins or original maybe dead CPU boards are sent to me. If requested, what's left will be sent back with the new CPU board.

It's actually not that hard to do, maybe a bit fiddly, but to help make it a bit easier, I also supply a small circuit board template to help line the pins up while soldering.

The user manual is available in the CalCom download, (CPU_WS.pdf), if anyone wants to read through it.

cheers

Tony

[rprosperi]

Thanks for these helpful answers Tony!!

My own valuable contribution to the world of calculator electronics is to never again attempt soldering of this nature, so when ready to order I'll send you a machine to have it installed.

The idea of modifying a normal Woodstock power cable to replace the brick with a normal USB A connector is great. Simple, elegant and obvious. Well done Dave Eaton.

Looking forward to this Tony!

[MinkLib]

I read the CPU_WS.pdf with interest.

The FTDI serial link for reflashing now uses 5v. instead of 3.3v on the Spice board - why is that?
Also why (other than mechanical access) does the board have to be removed to use the serial link?
Can the calculator be powered on (again a la Spice) while serial reflashing?

An exciting new development for sure, for the trusty Woodstock family. Well done indeed, Tony.

[teenix]

(04-27-2024 10:35 AM)MinkLib Wrote:  I read the CPU_WS.pdf with interest.

The FTDI serial link for reflashing now uses 5v. instead of 3.3v on the Spice board - why is that?
Also why (other than mechanical access) does the board have to be removed to use the serial link?
Can the calculator be powered on (again a la Spice) while serial reflashing?

An exciting new development for sure, for the trusty Woodstock family. Well done indeed, Tony.

The power supply is a little bit different to the Spice. The FTDI powers the CPU board, whereas the Spice does not.

Edit: The battery is difficult to connect with the CPU/Display/Key boards removed from the case so the FTDI board supplies the power.

The 3 possible supplies all with different voltages are isolated from each other, except the battery could override. The display and Bluetooth are powered directly from the battery so while the FTDI is powering the board, or from the programmer which cannot supply enough current to drive a fully operating CPU board, they will not function. There shouldn't be issues with merging supplies, but it was a "just in case' thing.

The serial reflash is just meant as a temporary measure. For external power, the USB cable can be used via the charge connector. You can use the FTDI interface as you would the Bluetooth, to change the menu options or transfer programs via CalCom.

cheers

Tony

[teenix]

I opened up the Woodstock AC power supply and gave it a 5VDC output using a simple circuit consisting of a full bridge rectifier, 7805 regulator and a couple of caps. This allows the original charger to plug directly into the calculator to power the new Woodstock CPU board and charge the battery.

It's a bit bulky, so I'll see if I can miniaturize it to fit comfortably into the charger case.

cheers

Tony

[tangdfx]

Finally, I build the hp29C with Tony’s board.
   
From left to right ,there are HP 29C original 29C with Tony's board base on an HP25 and HP29C with Panamatik low power.

[teenix]

Hi all,

I just put together new charger internals for the new Woodstock CPU. It supplies a regulated 5V supply from the mains 120-240VAC 50/60Hz. The Woodstock has an internal LiPo charger and the charge indicator is showing on the right side of the display.

The total conversion was < $Aus5.00

cheers

Tony

[teenix]

Hi all,

Due to having to move into temporary accommodation, working on the 19C has become difficult as most of my stuff has been packed up.

I was however able to put together the HP-91/92 CPU board. It only required my laptop and a soldering iron. Still early days, but it is up and running.

The image shows the board installed in a HP-91. It will fit into the HP-92 as well and the CPU can be switched between the two models on the one calculator.

cheers

Tony

[tangdfx]

Good news Tony.

We are close to complete the final block of the HP classic LED calculator.

[Paul Dale]

Tony,

Is there anything I can do to assist you?


Pauli

(07-09-2024 01:49 PM)teenix Wrote:  Due to having to move into temporary accommodation, working on the 19C has become difficult as most of my stuff has been packed up.

[teenix]

Hi Paul,

The offer is most appreciated, thank you :-)

I can't really do anything with the 19C at the moment though, I just don't have the resources where I am. I was fiddling with the motor driver hardware, but I have no idea where a new test board ended up, or even if the idea will work. Oh well, more time to think about it.

I hope it won't last long, but I can fiddle with the new Topcat board as it will only be software, assuming it doesn't show a hidden hardware fault, then I'm probably stuck.

cheers

Tony

(07-10-2024 08:59 AM)Paul Dale Wrote:  Tony,
Is there anything I can do to assist you?
Pauli

[knife31]

Hi
HP97, Printer, Tape Drive, 45, 55, 67
Made some battery connectors and stripped out the connectors on my old batteries. Took them down to "Batteries Plus" and had them soldered and shrink wrapped. They supplied the batteries. Reasonably priced. Did the same thing for my 45/67 etc.. Only problem is they are pretty loose in the compartment, but I can live with it.
RH

[teenix]

A little bit of soft foam on top of the battery pack might do

cheers

Tony

[BobVA]

(07-12-2024 02:48 AM)knife31 Wrote:  ...Only problem is they are pretty loose in the compartment, but I can live with it.

+1 on Tony's recommendation. For shrink-wrapped Classic 3xAA packs I usually wind up gluing a strip of neoprene, cut from an old mouse pad, to the left edge to make up for the length. (By "left" I'm referring to the side of the battery pack closest to the contacts when installed - i.e. the left edge if you look at the installed pack with the back of the calculator facing you.)

[teenix]

Well, it looks like the 91/92 CPU board is behaving itself.

Figuring out the problems with the 92 which I posted earlier on a different thread manifested in this board as well, so that is sorted, and I learned a few other bits about the ACT in the process.

This board was a bit tricky because even though they have the same keyboard layout, the actual key matrix is different, although not sure why. That in itself was not much of a problem, but when you want to use the 91 Topcat to emulate the 92, or vice versa, the keys need to be decoded differently.

There are also differences in the switches and how they are interpreted by the ACT. The actual switches are different as well, so some switch functions are set from the menu when emulating a different model to that of the actual hardware.

I have the 2 PIK printer ROMs installed, and the printer is responding as it should, but I haven't connected the print head up yet. I want to make sure everything is working properly, and the print head driver shuts down if an error occurs. If the software fails, as it does sometimes during development, I need to make sure the print head doesn't stay energized, which would be a disaster as it will probably burn out quickly.

This has the same printer driver software as the 97 board, so it should(?) be ok.

That will be the next task.

cheers

Tony

[teenix]

Hi all,

I have the HP-91 and HP-92 board pretty much done. It is happy running in 91 or 92 hardware and will also emulate 92 in 91 hardware and the 91 in 92 hardware. I thought this would be easy, but there are differences to account for.

I added a clock which also has the funky text display and this all works ok.

The printer is being a bit cranky when paper is put in, so it looks like it needs a service, hopefully just lubrication. It is interesting that the idler gear for this printer is made out of black plastic and not the crumbly waxy looking ones I've seen in the 97's.

I also added Continuous Memory for both models and doesn't require battery backup.

cheers

Tony

[tangdfx]

Hello, Tony

So the question is when could we place the order?

[teenix]

(08-29-2024 01:21 AM)tangdfx Wrote:  So the question is when could we place the order?

It should be ready soon(?) :-)

I just have to sort the printer mechanism. It is hard to do this in the temporary living conditions I find myself in, most of my stuff is packed away.

cheers

Tony

[teenix]

Hi all,

While I can't do too much where I am currently staying, I took a plunge into the abyss.

The attached image shows my HP-95C emulator turned on with a new battery, ie Continuous Memory reset. I have previously done a simulator for this model, but this one is running HP microcode.

I cannot get the original microcode as it is very rare, so I am writing it myself. I gotta say, it's a big job and getting used to writing this style of code is a bit of a test of the old grey matter, especially in a tiny code development environment. I did the HP-10 this way, but this one is orders of magnitude larger.

I have so far managed to create all the key entries, display formatting, and about 30% of the functions on the right side of the key board, + - mul div, fix sci eng, deg rad grd, and some error codes.

There is no way I can exactly copy the original code as it is quite rare and I doubt I will ever see it, but I'm pretty confident I can bring it to life. I plan to include it in my replacement HP91/92 CPU board (about 99% working) which shares the same layout, so it will work in the real world.

Going by the available user manual, if I add up all the programable functions, the total is greater than 256. That means they all cannot fit into the standard way of coding program steps which are only 8 bits wide. I am therefore going to reduce some of the functionality which means things like steps (say) LBL 0 - LBL 9, I can only implement LBL 0 - LBL 7, which I don't think is too big a deal.

I'll keep progress posted, but it will take a while, there is going to be a lot of 2 steps forward and 1 step back - programming pun :-)

cheers

Tony