ACT available

[brouhaha]

(03-11-2015 12:18 AM)rprosperi Wrote:  Eric - I'm happy to lend you the machines if that's what held you up, and as long as they continue to work after dumping.

Thanks for the offer. I don't think it will be necessary, as Panamatik just dumped a 33E, and I think he has a 31E. Years ago I tried to dump the ROM from a 33E, but that particular calculator wouldn't operate with a 100 Kohm logic analyzer probe attached to the clock signal. All the others I dumped had no problem with that.

I do have another request, though. If you (or anyone else) is willing to open their 30-series calcs, is detailed records for each unit of the model (e.g., 33E), the markings of the CPU chip (40-pin) and memory chips (8-pin).
The positions (order) of the memory chips are less important; they are fully interchangeable on the E-series, while on the C-series the one closer to the display contains the continuous memory and has different power supply wiring.

The main chip is normally an 1820-2015, 1820-2122, or 1820-2162, optionally with a letter suffix, or 1MA4-nnnn. The base (numeric) part number chips are NOT interchangeable across all models. It is unknown whether the letter suffix (where present) affects interchangeability.

The 8-pin memory chips are normally have an 0031- or 00031- prefix, or 1MA1, 1MA3, or 1LB9, with two or four digits appended. The different part numbers may or may not denote different ROM code.

Unfortunately when I dumped a bunch of the ROMs from these years ago, I wasn't meticulous enough in my record keeping.

Thanks!

[PANAMATIK]

ACT Bug report:

Thanks to two attentive HP-25E users, that reported me both the same observation, I want to announce a bug in the extended ACT software Revision r1.00. The stopwatch calibration does not work. It will never calibrate the time to 05:00:00 minutes. I corrected this already in the next revision r1.01. Unfortunately there is no workaround, but at least the wrong calibration procedure doesn't change the initial accuracy. The uncalibrated accuracy of HP-25 calculators I tried was about 2 to 8 seconds off in 5 Minutes due to the lack of a quartz.

If the update release is available, this feature will work again.

Don't spend time for repeated calibrations. It will not work, it's just a bug.

I'm sorry for the inconvenience.

Bernhard

[PANAMATIK]

With the modified ACT hardware I was able to read the ROM codes of HP31E and HP33E for the first time. They were never published before.

Here some pictures of how the ROM from a spice calculator can be read:

The serial interface of the modified ACT scanner/recorder is connected to a PC, the relevant signals are connected to the ROM chip by 5 wires.



The calculator is switched on and the selftest is started by STO ENTER. This initiates reading every ROM address in ascending order for some seconds. The addresses and opcodes are recorded while the selftest is performed.

Reading of the HP-38C

With some post processing and stripping unnecessary data and de-interleaving bank switching instructions the received data is converted into a ROM code file of the specific calculator.

You can download the disassembled codes from my website.

It should be possible to run all of the HP31E HP32E HP33E HP33C HP34C HP38C firmware on an HP-25 hardware with the new ACT. But I'm still working on it.

Bernhard

[rprosperi]

(03-12-2015 01:22 AM)brouhaha Wrote:  I do have another request, though. If you (or anyone else) is willing to open their 30-series calcs, is detailed records for each unit of the model (e.g., 33E), the markings of the CPU chip (40-pin) and memory chips (8-pin).
The positions (order) of the memory chips are less important; they are fully interchangeable on the E-series, while on the C-series the one closer to the display contains the continuous memory and has different power supply wiring.

I have a 32E which I'm willing to open, but doing so with Spice machines is a pain. Asyou know well.

I also have an almost new 34C that I really don't want to open and a 38C that has been opened too many times already.

So if 32E is useful info, I can do that.

[PANAMATIK]

The recently extracted HP31E firmware is now running on my HP-25 calculator with help of the new ACT chip.

Why am I doing that? Obviously this is just a try and has no practical meaning, except that I can proof that the ROM firmware was extracted correctly from the ROM chips and I can learn something about the spice calculators.

Both calculators have the same number of keys but of course with different functions. The PRGM/RUN switch has no effect.

As it seems at first glance very easy to run the firmware, it was not. Keyboard rows were arranged in different order and produced different indexed jumps into the firmware. After I found the correct entry points in the firmware, I could setup an appropriate jump table. The spice calculators does have only 10 digits and a special digit with minus sign only. Thus data encoding for decimal point, minus and decimal comma is different. The HP-25 doesnt have yet a decimal comma. I had to find the differences and eliminate the decimal commas. The minus signs for mantissa and exponent are coded differently, I had to find out how they are coded and transform the data stream for the display drivers.

The flag bits for PRGM/RUN switch and Battery Low check state are reused for other purpose in the HP31E i.e. flag for decimal point entered. I had to remove this program part.

But then the result now is very convincing.

I did enter many keystrokes in parallel on both calculators and don't have any differences except one: If recalling a nonexistent register RCL > 3 the original shows Error 2, the clone shows Error 3. I have no idea why! Its exactly the same firmware, but I will find out.

Add: The problem was found, the first ever HP31E clone is working completely and shows correct error numbers now.



Bernhard

[BarryMead]

(03-13-2015 02:42 PM)PANAMATIK Wrote:  You can download the disassembled codes from my website.

What is your website's "URL"?

[Massimo Gnerucci]

(03-14-2015 08:41 PM)BarryMead Wrote:  What is your website's "URL"?

Here you go!

[PANAMATIK]

The new ACT made another step forward towards a possible replacement for the 30 series. I integrated the HP32E ROM firmware into the ACTs internal memory and it emulates the HP32E now on an HP-25 hardware, like it did with the HP31E firmware already. I had to add more RAM and now it is performing all mathematical functions of this mid-range scientific calculator. The most remarkable function IMO is Q^-1, which needs more than 10 seconds calculation time and shows a flickering display like a programmable while busy.

Unfortunately there is still one thing left, that does not work: the ENTER key. For some reason, it is not jumping to the right program location. I didn't yet find the reason, but I will work on it.

If ever your ENTER key is stuck, you can always workaround it by pressing the XY button twice. Is this true?

Bernhard

[N4TVC]

(03-15-2015 11:59 PM)PANAMATIK Wrote:  The new ACT made another step forward towards a possible replacement for the 30 series. I integrated the HP32E ROM firmware into the ACTs internal memory and it emulates the HP32E now on an HP-25 hardware, like it did with the HP31E firmware already. I had to add more RAM and now it is performing all mathematical functions of this mid-range scientific calculator. The most remarkable function IMO is Q^-1, which needs more than 10 seconds calculation time and shows a flickering display like a programmable while busy.

Unfortunately there is still one thing left, that does not work: the ENTER key. For some reason, it is not jumping to the right program location. I didn't yet find the reason, but I will work on it.

If ever your ENTER key is stuck, you can always workaround it by pressing the XY button twice. Is this true?

Bernhard

Yes, that does a stack push like enter. Amazing what you are able to do with this ACT platform, Bernhard. It is fun to watch the forum to see what you will do next.

[PANAMATIK]

HC-29E released!

I did already announce in a previous posting, that the new ACT chip is emulating also the HP-29C firmware correctly, which is the most complex firmware of the "Woodstock" models. But I had some more programming to do for teaching it all the extended features described for the HP-25E. This is accomplished now and there is an HP-29CE now available. The calculator, when switched on, however shows "HC-29E", because the HP-29C display driver unfortunately cannot form the letter P. The C replaces the P and E is for Extended version.

The first HC-29E. Thanks to W.B.

There are some differences between the HP-25E and HC-29E:

Due to the double size of the internal HP-29C ROM firmware, there is only flash program memory left for 30 times 98 steps instead of 110 times 49 steps.

The number of predefined and selfdefined constants is limited to 40 instead of 100.

Because nearly all the HP-29C keys are occupied with g and f prefix functions, two combinations g R/S and g GTO cannot be used for the extended functions, they are called now by double g g prefix.

Obviously there is no need for inserting and deleting program steps because this is already built in.

All other extended functions and features described in the ACT manual are now working with the HP-29C calculator. There are no hardware differences to HP-25E ACTs.

As a final point, I found a lot of program solutions for HP-29C on the Museum's DVD set beginning with "Electrical Engineering solutions", "Mathematics solutions" to "Navigation solutions" ..... It is impossible to select only 10 or 20 of them for a predefined program collection. So I decided not to include a program collection, but let you decide, which programs you want to store. All 30 program spaces are free for use.

Bernhard

[PANAMATIK]

Add: Oh, I forgot to mention, that the HC-29E also would run on any HP-25/HP-25C hardware.

Bernhard

[Les Bell]

(02-15-2015 10:45 PM)PANAMATIK Wrote:  The new ACT works completely with the HP-27 ROM code. I checked, that there are no unknown opcodes. I tried all scientific, statistical and financial functions of the manual and all seems fine.

I've always liked the look of the HP-27 - it looked like an ideal non-programmable machine for those back-of-an-envelope calculations in meeting, etc. Plus, I like the unique colour scheme.

So when I stumbled across one for sale this morning - probably with a fried ACT chip - the existence of this thread was the source of great joy to me!

I'll wait until the machine arrives, just to make sure, but I expect to be ordering an ACT for the -27, and I might as well upgrade my -25 to a -25E while I'm at it. My wife is probably going to kill me, but what the heck . . .

[PANAMATIK]

(03-24-2015 12:51 AM)Les Bell Wrote:  I'll wait until the machine arrives, just to make sure, but I expect to be ordering an ACT for the -27, and I might as well upgrade my -25 to a -25E while I'm at it. My wife is probably going to kill me, but what the heck . . .

The HP-27 ACT is an NMOS chip and the new ACT is slightly different to the HP-25E. If you order both you have to be careful not to mix them. Of course they will be labeled.

This would be the first new ACTs in .au

But first try whether the HP-27 ACT is fried. As mentioned earlier I bought two defective HP-29C and one of them had just a contact problem.

Good luck
Bernhard

[Les Bell]

(03-24-2015 06:08 AM)PANAMATIK Wrote:  This would be the first new ACTs in .au

But first try whether the HP-27 ACT is fried.

Cool! Give it a week or three to arrive, and I'll be in touch. Thanks, Bernhard!

[pito]

@PANAMATIK: would not be better to wire at least a zener from D2 anode to GND in order to protect the pic16? Otherwise it will fry for the same reasons as the old ACTs did

[PANAMATIK]

(04-04-2015 10:30 PM)pito Wrote:  @PANAMATIK: would not be better to wire at least a zener from D2 anode to GND in order to protect the pic16? Otherwise it will fry for the same reasons as the old ACTs did

Hello pito

Thanks for your suggestion.

A protection with Zener Diode or a three pin voltage regulator for additional protection is a good idea. I used a simple resistor at VCC instead. The PIC should have a build in diode at VCC as all modern chips normally do. The resistor limits the current if voltage is too high at pin 1. I admit, a zener diode at the right place could protect the other chips on board as well and would have been the better choice. Perhaps there will be a hardware update in the future, but still the best protection is the knowledge not to charge without batteries (which I didn't take serious enough some month ago).

[pito]

(04-05-2015 06:51 AM)PANAMATIK Wrote:  

(04-04-2015 10:30 PM)pito Wrote:  @PANAMATIK: would not be better to wire at least a zener from D2 anode to GND in order to protect the pic16? Otherwise it will fry for the same reasons as the old ACTs did

..The PIC should have a build in diode at VCC as all modern chips normally do. The resistor limits the current if voltage is too high at pin 1. ..

The PIC and all other chips do not have a build-in diode at VCC. They have got clamp diodes at signal pins, which may limit the over-voltage at signal pins to something like VCC+0.7V. However, the pic16's VCC is not protected (the highside clamp diodes have they cathodes at VCC, so no VCC against GND protection at all..). When the voltage at Vss(ACT pin 1) reaches let say 8V, the voltage at pic16's Vcc will be 7.3V (or less by V_R11). A 5.6-6.2V smd zener will do the job..

Of course a better protection would be a 5.6V zener at the battery contacts - you'll protect other chips too - but that is not so critical today as your superchip can replace all the old silicon stuff there easily
Great work!

[Les Bell]

(03-24-2015 06:08 AM)PANAMATIK Wrote:  But first try whether the HP-27 ACT is fried. As mentioned earlier I bought two defective HP-29C and one of them had just a contact problem.

The postman just dropped off my "new" HP-27 so I dropped what I was doing and unwrapped it. With a new battery pack, it's dead - nothing on the display at all. The seller had said "Powers up but flickers, it will not work properly until batteries are replaced.", but I expect he was trying to run it on the charger - something I'm not game to do, for fear of causing further damage. So this is pretty much what I expected - it might well be a fried ACT, or there might be something else dead as well.

However, the calculator is in excellent, in fact pristine, physical condition, so it's well worth while trying to bring it back from the dead. I just need a day or two to gather my notes on Woodstock, and particularly, HP-27 repair, and I'll be putting in an order for a couple of ACT replacements.

I must say, the 27 is a pretty little calculator, though, with its khaki colour scheme, so I'll be well pleased if I can resurrect this one.

[bshoring]

I saw the photo of the HP-27 on Facebook. Nice looking calculator. I'll be interested to know if the ACT will get it working.

Good luck!
Bob

[Oulan]

Just for curiosity did you plan to read the HP19C rom ?
(I just put some woodstock emulators on google play)

P.S. on go32c emulator I use the ENTER key as 0x73 keycode (in hexa)