interviews

I'm going by memory, so I could easily be wrong. I never heard about an instruments division at IBM until in the early 80's (82-83?) they bought several analytical chemistry product lines from Bruker, including the FTIR and NMR products. That's what I was thinking of, and why the Byte ad caught my eye, being an analytical chemist so of course the world revolves around what I care about :-). Then later in the 80's they sold everything that was originally Bruker and maybe some other stuff back to Bruker. If you were there working for IBM during that period I'm sure you have much more accurate memories than I do about when they acquired and sold various product lines.

IBM owned the Instruments Division well before '84. I designed the

I think it went away 1984-85ish. I didn't get there till late '87.

Cheers

Phil Hobbs

They were around in the mid-'70s, as well. They had a range of color (paint) matching tools, EKG machines (at least I remember them being in the instruments division at one time), chemical analysis tools, and lab instrumentation. The 68K based IBM9000 instrument controller came out about the same time as the PC ('82). That's when many first realized they were around. I left that group in early '80 and worked on bringing the Intel parts into IBM (and miscellaneous losers like bubble memory ;-).

nice calculator.

The old HP35 finally gave up. That old nmos logic wasn't all that reliable. I'm using an HP32SII these days.

My HP45 still works, sorta. Someone swiped my HP11C, so I had to settle for an HP35S (though have the 11C clone on my smart phone).

The 35S is pretty awful. Someone should do an accurate clone of the HP35.

Does anybody need a programmable calculator? I'd rather write a computer program.

The easiest programmable calculator is an analog computer, aka a simulator... just draw the blocks of your thought processes and it solves itself. ...Jim Thompson

I use my HP-35 and HP-67 daily (one on each desk). The 67 needs some care feeding magnetic cards, I think I need to disassemble and clean the reader. Don't use the mag cards much but still handy. It really wows visitors!

piglet

An analog computer consisting of blocks of summation, gain, integration, etc was used during the design/analyses of Lockheed's L1011 flight control system. Not a digital computer. How does that work? Ask any pilot and they'll tell you how much they prefer flying the L1011 rather than 747 etc.

Small tidbit, during simulated approach at low altitude the designers hit the L1011 with a stronger than required windshear [I think it was]. The plane promptly rolled over on its back and arched into the earth at nearly

Same for me while I was there, if ever one of my designs failed to meet spec and I could NOT point to the component out of spec; the design was scrapped and I had to start over, couldn't just fudge a number somewhere to make all work. In other words all designers were forced from day one to 'control' their designs, not herd them along to Production. Makes for some 'conservative' designing, though.

My general design attitude... one which often annoys clients, at least temporarily... if I find myself patching I wad up the design and start afresh (as you can see, the term "wad" dates to when I did strictly paper designs plus breadboard).

I still adhere to that principle today... that's why I have a 99.6% success rate on first-pass silicon.

And that 0.4%... that represents the single chip that happened to "latch-up" during power supply ramp-up... which is why my designs now include that procedure as a standard test. ...Jim Thompson

It's not *that* bad. I'd rather have a new HP45 than '35. It was much better for electronics work and there was nothing lost from the

I never used one.

...with appropriate scaling.

If you don't know how to scale, you're in the wrong business :-D ...Jim Thompson

Have that chip reverse engineering guy make you some drop in replacements.

I think airbus converted that feature into software that runs on real computers.

I seriously doubt that it was the chip that failed. There was nothing wrong with NMOS logic except that is was speed-limited on the rising edge... in essence a current pull-up that diminished in value as you got near the positive rail.

NMOS was certainly an economic process... one single diffusion step. ...Jim Thompson

The HP35 was just one generation past board-stuffed-with-diodes logic, it was probably bipolar (or maybe PMOS); the NMOS 8080 came two years later. I've tried to look it up, but all I can find for sure is that MOSTEK was the chip fab.

I used my circa 1991 HP48SX at work three years ago to figure out a problem that required a 48-bit word size; a 48-bit number was coming out of an FPGA and then being (incorrectly) converted to a 64-bit number in C. None of the "desktop" calculators I tried, Linux or Windows, 32 or

That's probably the last time I did something that required its unique abilities. When my cell phone died the other day, I used the alarm on the 48SX to make sure I got up on time, but that probably doesn't count.

Matt Roberds