Help to identify National 14 pin DIP

I'm trying to reverse engineer a small circuit with a light detector
that's used to set end-of-stroke on a mechanical slide. The circuit has
a 7-pin header, three light emitter/detector pairs, two SIP resistor
packages, and one IC. I need help identifying the IC. It's a 14 pin DIP

It's labeled:
1st line: (National Semiconductor Logo) 158 9129
2nd line: SA 779197
3rd line: 8201 CPA

Anybody recognize this or know where I can get a data sheet for it?
Thanks.

 

You might want to look at the board, so as to see what might make sense.

When I had quad emitter/detector circuits to play with in the 80's we
were using an LM339 quad comparator. If the circuit looks like that
might be it (with one quarter out of use) then it's worth a shot. If the
pinout would not work for a 339, then that won't be what it is...

 

Thanks. I looked up the pinout for that, but the power pins don't seem
to be in the right spot for a LM339. Now I'm thinking it may just be a
hex schmitt trigger. I buzzed out the lines to the header, and if pins 7
and 14 on the chip are power, then pins 1 and 2 on the header are the
power connectors. Pins 3 and 7 are not connected to anything, and that
leaves pins 4, 5 and 6 as outputs, and they're on pins 2, 4 and 6 of the
DIP. I'll probably just put 5V on it and see what happens when I block
the light paths. I feel pretty sure it's just either totem-pole or open
collector outputs. Worst that can happen is I just have to build a new
one from scratch. What I don't understand is why they would use a house
part number on something like this. This is surplus equipment, I don't
know who the original manufacturer was.

 

It keeps the customer reliant on the manufacturer. At the very least it
means the parts come through the company so they get the profit, at worst
they can raise the price and make more profit.

It makes things a tad more difficult for someone to copy. It's no
different than some companies sanding off the IC numbers. Someone
tracing out the circuit has to work that much harder to figure out the
circuit, and if they don't find the IC through guess and such, then they
can't copy the circuit. It doesn't really stop those who have enough
knowledge, but it does keep out the simpler ones.

There was also a time when semiconductors weren't made as efficiently, so
there'd be lots of failures or semifailures in a batch. Sometimes the
manufacturer would put an official number on them, I'm thinking of the
time when 64K dynamic RAM were new and there was a high failure rate, but
it was limited to half the die. So they were perfectly normal 32K RAMs.
I suppose in some cases they might just sell the batch to one company,
with a house number, if the supply was such that a single company would
buy it all up. I suppose then things like transistors might be offered
ashouse numbered devices if they were out of spec, but a company that
could live with the lower specs could make use of them. You don't want
"bad" 2N706s to land in the hands of people who think they are full
spec'd.

Michael

 

IBM, too. Everything was house numbered and rarely did one group use
parts from another. Many had very unique specs. One engineer I
worked with spec'd an LF386 up to $30 each (put a tested limit on it
recovering after being driven into the rails).

 

Back when I worked for Fairchild/Schlumberger ATE we never used house
numbered parts. (But then again, maybe we did. Most of the chips we used
had Fairchild numbers and logos on them ;-)

 

We don't use many house numbered parts anymore. The things are too
damned small to print part numbers on them. ;-)

 

Well, there is that. Good thing someone invented the Mantis.

Laser? I thought it used electric fields.