"Weird Device" : a Calculator

[eccerr0r]

Well, normally calculators are disposable. Alas I found this Novus 821. Before this point I never realized that this was actually National Semiconductor's consumer electronics division that actually made ... calculators. Also the really old Nat Semi logo, I finally realized it's an N and an S, I was wondering what those two squiggly squares were.

Anyway, why this calculator was interesting? It has an LED display instead of the typical LCD!

Alas this thing isn't working that well. What was kind of odd when I looked at the guts, this thing has NO capacitors in it at all! Not one discernible!

Anyway I wouldn't have posted here if it didn't have problems: it, well, doesn't work 100%. First it doesn't reset on powerup, forcing me to press "C" every time. The other thing is that the LEDs go wonky. Sometimes I do wonder about the ICs in this thing, this should be a rudimentary processor with built in firmware I guess... and could the firmware actually go bad?

It has exactly two ICs in it: MM5738N and DM8864N - and both are made by Nat Semi of course.

Anyone have experience with these really old calculators? I guess this would be about 40 years old now... though I've had other ICs that have lasted longer than these.

[CapLeaker]

Re: "Weird Device" : a Calculator

Any pictures? Maybe have a look at the through holes if there are any???

[eccerr0r]

Re: "Weird Device" : a Calculator

Yeah those two ICs are through-hole ICs - only two.

Well I found out that the LED seems bad - I probed out the segments with my multimeter in diode check, and found a few bad segments. However I'm still not sure how the bad LED could cause the calculations to go all wonky though.

[TechGeek]

Re: "Weird Device" : a Calculator

readout not getting to the user...

[momaka]

Re: "Weird Device" : a Calculator

However I'm still not sure how the bad LED could cause the calculations to go all wonky though.

Are the bad segments shorted? If so, perhaps this is causing the driver IC to draw excessive current and possibly messing it or its signals up to or the other IC's signals.

[eccerr0r]

Re: "Weird Device" : a Calculator

Interesting... don't know, so far it looks like the bad segments are open. It's just weird that the open segment is causing another, completely wrong digit (but right segment) to turn on... it would imply a short but something strange is going on. LED is a standard multiplexed configuration.

One thing that was strange that I found: the LED display module is actually removable without solder...

[TechGeek]

Re: "Weird Device" : a Calculator

maybe make your own 7-seg display(get pinout and get a generic display then wire it up)?

[eccerr0r]

Re: "Weird Device" : a Calculator

If I could find a 9-digit common cathode (I think) display I may have to try it, though it's a lot of wires to keep track of... perhaps finding another display would have to be the only way to test...

I think this is pretty darn close but not sure if the HP display is the same as the NS display:

http://www.ebay.com/itm/2x-7-segment...-/201507164881

[fzabkar]

Re: "Weird Device" : a Calculator

There is an application block diagram for a calculator in the DM8864N datasheet:



Here is a datasheet for the MM5736 calculator chip:


... and a patent that uses the MM5738 chip:
https://cdn.badcaps-static.com/pdfs/...16b20c5e00.pdf

The MM5738 datasheet could be in the "1977_MOS_LSI_databook.pdf":

[eccerr0r]

Re: "Weird Device" : a Calculator

Yeah the behavior is very weird. The 1977 NS databook does not contain the MM5738, but does contain the MM5737 which is close (page 345 if anyone's curious)...

Still no clues why it behaves the way it does with the apparently bad LED display. I wonder how much current max I can drive into the LED to test the segments, I'm sure the diode check on my multimeter is pumping less than 1mA to make sure it doesn't fry delicate junctions.

[fzabkar]

Re: "Weird Device" : a Calculator

The MM5736 datasheet states that "two depressions [of the "C" key] are required at power on to clear the machine".

I don't know why, but each calculation example begins with a "C" to clear the display register.

[eccerr0r]

Re: "Weird Device" : a Calculator

interesting, perhaps the MM5738 also does not have power on reset?
Didn't look at the MM5737 too carefully...

[fzabkar]

Re: "Weird Device" : a Calculator

FWIW, the Novus 823 user manual also calls for two presses of the "C" key.
https://cdn.badcaps-static.com/pdfs/...3fe11bfee1.pdf

[eccerr0r]

Re: "Weird Device" : a Calculator

Interesting, so I guess it does not have power on reset. Ahh the conveniences of modern electronics :p I guess that explains the corrupt random display on startup.

Now what still remains is the weird behavior of digits when the LED segment is seemingly bad...

[fzabkar]

Re: "Weird Device" : a Calculator

Interesting... don't know, so far it looks like the bad segments are open. It's just weird that the open segment is causing another, completely wrong digit (but right segment) to turn on...

The calculator chip's LED digit select outputs also double as the scan outputs for the keyboard. Could two of these signals be shorted to each other???

[eccerr0r]

Re: "Weird Device" : a Calculator

The keyboard is basically garbage, they bounce like mad. However all the buttons work - implying that there probably aren't shorts...

[eccerr0r]

Re: "Weird Device" : a Calculator

WEIRD. I had a couple segments that didn't work but suddenly they started working again... No clue. Still not perfect but it's working MUCH better now, and can do some simple calculations with it when key bounce isn't an issue.

Well, pictures!

https://www.badcaps.net/forum/attach...1&d=1470881866
https://www.badcaps.net/forum/attach...1&d=1470881866
https://www.badcaps.net/forum/attach...1&d=1470881866
https://www.badcaps.net/forum/attach...1&d=1470881866

One decimal point is still not working on digit 3 from left though none of the pictures demonstrate this.

[momaka]

Re: "Weird Device" : a Calculator

Still not perfect but it's working MUCH better now, and can do some simple calculations with it when key bounce isn't an issue.

Lol, gotta love old technology.
About two decades ago, we used to have this landline phone in our house - one from the early digital era. That phone had terrible key bouncing, though. When dialing, you had to hit the keys very quickly and with just the right pressure. If we didn't, a key could register 2-5 times. As such, it typically would take us at least 2-3 tries to enter a 6 to 8 -digit phone correctly without errors . Probably the only time when I wish we had our rotary phone back.

[cyberhack]

Re: "Weird Device" : a Calculator

You where lucky we were poor so we had to make our own.
Lol

[eccerr0r]

Re: "Weird Device" : a Calculator

Wow, you could afford cans? I remember having to use paper cups and didn't even have enough string so we'd have to just yell...

I really wonder who came up with the rotary pulse system and get it so reliable (I don't recall having it get the wrong number unless you actually chose the wrong number or "user error" not hitting the endstop), as those switches themselves are prone to bouncing too.