TI-58/59 Modules Curiosity

[Matt Agajanian]

Hi all.

In the spirit of the NS-7100's recordable ROM potential, why were the calcs in the TI-58/59 series not designed to have a recordable Solid State Modules right from the calcs themselves?

[David Hayden]

I can think of two possible reasons. First, maybe it would have taken too much power to burn an EEPROM. Second, it may have been to ensure 3rd party ROM producers that their ROMs could not be easily copied. I know that my father paid a lot of money for one of those ROMs back in the day.

Dave

[Paul Berger (Canada)]

Why didn't the HP41? I would suspect that the designers did not think the R/W non-volatile memory technology available at that time was a viable solution so both opted for using magnetic cards for storage. Another factor that might have tipped the scale towards cards would be cost, the cost per byte of storage at the time would have been much less for the cards. They do not say what sort of memory was used in the 7100 modules, previously it was suggested it may have been EEPROM but at the time this calculator was scheduled to be released, EEPROM technology may not have been available. In my 1977 NS Memory data book they do list an electrically programmable ROM but it was not erasable and required a -50V programming pulse. The write up in the catalog that Katie linked does not say that the modules are rewritable, only that you can store program steps into it. I guess battery back RAM might have been a possibility but given memory and battery technology of the time that would have likely been problematic.

edit:
With the information just made available it seems the memory is indeed some early form of EEPROM even though in several places an Intel engineer is credited with developing EEPROM technology in 1978. It seems that the release was delayed a few times before it was cancelled, which would suggest there where significant problems with it, and that coupled with TI delivering the 58/59 probably doomed it.

BTW I read in one place that TI beat NS to market with the58/59 and that killed the 7100. The NS people probably saw what TI was releasing and threw in the towel it seems they exited from the business completely after this, and now National Semiconductor is owned by TI. The base 59 had more program steps in the base unit than the 7100 plus a "File cartridge" not to mention the more cost effective program storage and the extra routines that would be available in the plug in modules. Even the 58C had as many program steps available, but of course no way to save the other than in the calculator's continuous memory. It would probably be at least a couple years later before people began to find out how bad the keyboard of the 58/59 was, just in time for HP to release the 41.

[Matt Agajanian]

Thanks for this. Now that I see this sequence of events it makes more sense how and why the 7100 was held back.

[brouhaha]

The first handheld computer with non-battery-powered nonvolatile in-system-writable semiconductor memory for storage of user data may have been the Psion Organizer in 1984. It used EPROM capsules which were written in the device, but had to be removed to be erased by a separate UV lamp.

There were earlier handhelds that used EPROMs, e.g. The Matsushita/Panasonic/Quasar HHC, but the device did not write the EPROMs, and they were typically only used for canned programs in the same manner as HP-41 and TI-59 ROM modules.

The HP-71B was intended to support EEPROM modules, which are referenced in the IDS. Based on references in HP's bug tracker, apparently there were even prototypes, but they might not have been packaged to physically fit in a normal-sized 71B module.

[Didier Lachieze]

The Elektronika MK-52 around 1983 is a programmable calculator using an EEPROM memory to store the programs. It may have been the first one using this kind of memory after the 7100.

[brouhaha]

I have an MK-52. I'm fairly skeptical of the claim that they were introduced in 1983, but I don't have any hard evidence that they weren't.

[Didier Lachieze]

The earliest evidence I've found for the MK-52 is 1985, so 1983 may be too early.