SR-56C? Interesting idea.

[Matt Agajanian]

Hi all.

By the time the HP-25C came out, Constant Memory was possible. So, hypothetically, if TI just released a 56C instead of the 56, how much would it’ve cost?

[Peet]

It might not have cost anything, the TI-57 had pseudo-constant memory via a software trick.

[Steve Simpkin]

(06-10-2022 11:21 PM)Peet Wrote:  It might not have cost anything, the TI-57 had pseudo-constant memory via a software trick.

https://www.pcjs.org/machines/ti/ti57/tips/

[pauln]

That's an interesting question. What was the price difference between the TI-58 and the TI-58C in 1979, when the TI-58C was introduced?

I believe that, in 1977, the TI-58 was around $130. By the way, to me, that was a great value compared to the $80 for the TI-57.

I'm looking at HP-25 vs HP-25c:
* https://www.rskey.org/hp25
* https://www.rskey.org/hp25c

According to rskey, the price, when introduced, was essentially the same. But of course, the continuous version came the following year. By extrapolating, it seems to me that TI could have made a 56C for not much more than a 56, but it would have made the 52 less desirable than it was.

[pauln]

(06-11-2022 12:30 AM)Steve Simpkin Wrote:  

(06-10-2022 11:21 PM)Peet Wrote:  It might not have cost anything, the TI-57 had pseudo-constant memory via a software trick.

https://www.pcjs.org/machines/ti/ti57/tips/

By using the page with diagnostics, one can see that the calculator doesn't actually go to sleep when using that trick.

https://www.pcjs.org/machines/ti/ti57/rev0/

Incidentally, before writing RCL-57, I didn't realize that the calculator is constantly updating the display even when the calculator is idle. I guess that for anybody into hardware this must be obvious but it was a big surprise for me.

[teenix]

(06-11-2022 01:05 AM)pauln Wrote:  Incidentally, before writing RCL-57, I didn't realize that the calculator is constantly updating the display even when the calculator is idle. I guess that for anybody into hardware this must be obvious but it was a big surprise for me.

I imagine the display is multiplexed similar to the HP LED models, so if the display is not constantly updated it would go blank, unless in the case of a fault, a single segment may be on very brightly and possibly damaging. Also the keyboard is scanned during the display updating, so the processor can know when a key is pressed. More modern machines with low power LCD displays probably have processor interrupts to detect key presses and wake a processor from sleep mode.

cheers

Tony

[pauln]

Thanks for the explanation.

In the TI-57, the display goes black when doing an "expensive" computation such as "88 sin" which takes a couple of seconds: in order not to be even slower, there is no refresh of the display.

Until I understood what was going on, I was puzzled why the display of the emulator would never go blank. At the end, to match the behavior of the actual device, I had to explicitly set the display to blank if it hadn't been updated in the last x ms (I think I ended up using 250 for x).

[teenix]

(06-11-2022 05:13 AM)pauln Wrote:  In the TI-57, the display goes black when doing an "expensive" computation such as "88 sin" which takes a couple of seconds: in order not to be even slower, there is no refresh of the display.

This could certainly be the case for the TI calculators, I'm not sure.

In the HP LED models the display can be blanked but the processor execution speed does not change. This is because each display digit is refreshed during one instruction execution (56 system clock cycles) . So even when the display "flickers" during calculations or running programs, the speed still stays the same.

cheers

Tony

[pauln]

We are probably saying the same thing but just to make sure: in the TI-57, the processor speed is also constant but sometimes the display doesn't get updated/refreshed for a while and therefore it goes blank.

This can be illustrated by looking at the ROM activity monitor in https://www.pcjs.org/machines/ti/ti57/rev0/: when idle, the program counter (brighter green dot) is around address 0x04a7 (DISP instruction) constantly refreshing the display.

If one presses "88 sin", the program counter goes to larger addresses, where there is no DISP instructions, for a couple of seconds and the display goes blank.

The point is that there is no instruction in the TI-57 to tell the display to go blank. The display goes blank because it hasn't been updated/refreshed for a few milliseconds and therefore the light has faded away.

[Peet]

(06-11-2022 05:13 AM)pauln Wrote:  In the TI-57, the display goes black when doing an "expensive" computation such as "88 sin" ...

Many calculators with LED displays from that time behaved like this. My first programmable calculator was a Commodore PR100 and I remembered this behave at e.g. 69!.

BTW, back then I always used the "." as a constant memory simulation.

[John Garza (3665)]

I wonder...

What about hard-wiring some power to the memory chip, bypassing the ON/OFF switch? And the CPU, if it contains a power-on clear memory routine. It would eliminate the drain from the LEDs. I wonder what current just the chips would pull?

I always preferred the SR-56 over the HP-25 as it had subroutines. And it had 100 steps memory to compensate for the unmerged codes - roughly on par with the 49 fully-merged steps in the 25.

However, I would always chose the 25C simply because of the continuous memory. So a 56C would certainly tip the balance for me. The 56 looked like a calculator. The 25C, though I still love it, looks like a bathtub.

-J