▼
Posts: 275
Threads: 41
Joined: Mar 2010
I'm finishing a TI57 emulator for android, based on the TI patents. When reading them, a question arise as I don't have a real TI57 (anymore, got one in 1980).
When running a program or doing a calculus, did the dots in the led display 'flash' ?
The patent explain how this flashing can be done but say that it can be disabled depending of the chip hardware.
Thanks in advance.
Olivier
▼
Posts: 536
Threads: 56
Joined: Jul 2005
I think the display would run a single point dot along the digit positions as it ran.
i cant exactly remember the behavior, but i used to refer to it as a screen saver :-)
-- hugh.
Posts: 24
Threads: 4
Joined: Feb 2010
Hello,
It's been some years since the last time I used my TI-57 but I very clearly remember its behaviour.
When doing a computation, the display blanks (duration ranges from a few milliseconds to almost 1 second for trigonometrics).
When running a program, the display acts as if you were keying in manually as fast as the calculator could allow you, pressing a key as soon as the previous result is displayed.
Visually, you got a briefly flashing display of each step result : numbers appear during not more than 80 mS (estimated).
I even remember a "program backup" apparatus described in the French magazine "L'ordinateur Individuel" built around an EPROM, a crude clock and counters, hard wired to the TI-57 keyboard ! In some way, it replaced the little built-in user-program memory and sequencer to extend capacity. Those were the days...
It also makes me think that a "fast" mode (à la ZX 81) could have been implemented by TI engineers : blank display during program execution, and a flashing dot to tell the user that something's going on.
Hugh : the screen saver you talk about was TI-30's, its goal was to save battery life since that model used non rechargeable 9V batteries (unless you bought the optional battery pack).
Hoping it well help,
Regards,
Marc
Edited: 22 May 2012, 4:33 p.m. after one or more responses were posted
▼
Posts: 275
Threads: 41
Joined: Mar 2010
From the patent, when not displaying a result, the dot of each digit is activated (with a lower duty cycle) by the second bit of nibble 0 to 11 of B cpu register (16 nibbles) so it can appear as random dots ...
The digits are blanked during this time.
Olivier
Posts: 430
Threads: 86
Joined: Jul 2007
WATCH OUT!
I am still waiting for feedback from TI's management about the TI-59 emulator running the microcode and eventually emulating it on an ARM CPU to provide a faster replacement board (sorry, no red case planned).
You don't want to deal with their legal department....
Joerg
▼
Posts: 275
Threads: 41
Joined: Mar 2010
The patents contains a rom listing (57 or 59), certainly not a final one, but this one is now in public domain ? (nearly 33 years old)
P.S. the go57c emulator is now on google play (with a rom taken from patents, even if a tmc1501 has a special test mode to dump its rom content...)
▼
Posts: 32
Threads: 2
Joined: Dec 2006
Yay! PI * 1 INV lnx - 1 INV lnx * PI = 2E-10, as it should be. Thanks! But somehow it also emulates the flaky keyboard, giving both visual and tactile feedback but not registering the key. Is that intentional?
▼
Posts: 275
Threads: 41
Joined: Mar 2010
Not really, but it's due to the debouncing behavior of the rom, you have to 'press' the key for a certain duration otherwise it is taken as a 'bounce'.
So, like the real hardware, 'press' firmly (i.e more than 0.3 sec) to have a real key press :)
Perhaps I could patch it ?
As other 'bug', the RST opcode stop the program at step 0 with this rom.
▼
Posts: 32
Threads: 2
Joined: Dec 2006
I don't think you should patch it, it makes it more authentic. It would be nice if it was the real ROM from the actual calculator. I think the real one showed the calculations as they happened when running programs, not moving dots. I used it in a game with a graphic explosion when the gun had the right angle to hit the target.
▼
Posts: 275
Threads: 41
Joined: Mar 2010
For the real rom, someone with a 57 and some time can use the 57 patents where it is explained how to 'dump' the rom. For the display behavior, I have done as described in the patents as I don't have a 57 anymore (I got one in 81).
Posts: 381
Threads: 32
Joined: Mar 2006
Quote:
As other 'bug', the RST opcode stop the program at step 0 with this rom.
Some 6 years passed since I was into a TI-57 emulator, but IIRC a NOP at 20E solves this problem:
...
20D 0E04 STAX
20E 0EFF NOP ; RST breaks running program
20F 140D CALL 40D
...
There are some other bugs in this ROM. For example, I have one fix labeled "SBR NONEXISTENT BUG" at 279:
...
278 0C74 SF B15 1
279 1B83 BRNC 383 ; SBR NONEXISTENT BUG
27A 1601 CALL 601
...
I don't remember what exactly was the problem there ...
▼
Posts: 275
Threads: 41
Joined: Mar 2010
Yes, the 'nop' at 20E fixed the RST behavior, but for nonexistent SBR label, actually the display just blink and if in a program, it stops and blinks, so I don't see here a problem.
(btw, EFF opcode is your creation ? I use E0F as 'nop' myself)
For the rom at 0279, the patents 4125867 already has the 1B83 opcode ...
▼
Posts: 381
Threads: 32
Joined: Mar 2006
Quote:
(btw, EFF opcode is your creation ? I use E0F as 'nop' myself)
Possibly. I defined some other opcodes to improve my TI-57 emulator for HP-48/49, so I used the last code in Exx range for NOP, don't remember exact details.
Quote:
For the rom at 0279, the patents 4125867 already has the 1B83 opcode ...
Could be. I see that Google patents PDFs are of a higher quality than those I had at that time - should do another compare when I'll have some time ...
Posts: 1,830
Threads: 113
Joined: Aug 2005
Quote:
Not really, but it's due to the debouncing behavior of the rom, you have to 'press' the key for a certain duration otherwise it is taken as a 'bounce'.
For a really faithful emulation, you could simulate the keys having inconsistent responses. For extra points, you could start failing them randomly after they have been pressed a dozen times or more. I especially like the behavior where the key gets stuck and the calculator has to be rapped against a table to unstick it. It could be worthwhile to simulate what effect this sort of shock has on the electronics, too.
This is all based in my experiences with the TI-30. It may not apply so much to the TI-58/59. My collection includes both, but I never use them.
Posts: 2,309
Threads: 116
Joined: Jun 2005
If there was a US copyright on it, the copyright lasts 95 years. Only a little over 62 more to go.
▼
Posts: 275
Threads: 41
Joined: Mar 2010
There is no copyright on those patents, so how long will they last ?
See with 'google patent 4107781' with google search. Other patents are : 4125901, 4125867 and 4079459
▼
Posts: 2,309
Threads: 116
Joined: Jun 2005
Patents aren't copyrighted, but sometimes portions of the material contained therein is, especially in the case of software. (I personally think that to be a crock, but my opinions are not worth the paper they're not printed on.)
Anyhow, the only thing I'm asserting with regard to TI-57 firmware is that *if* it is copyrighted, the copyright still has many years to go.
Posts: 430
Threads: 86
Joined: Jul 2007
I had a few weeks ago with TI a discussion about this question.
I WOULDN'T USE THE CODE!
Regards,
Joerg
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
I thought the code had to contain a copyright message or it wasn't copyrighted. Hence the trouble with the HP Voyager series...?
|