i mean c'mon folks, it's got the part number right on the top! and the date code, and the whatever code
I have a tube full of these things. it's a very old style of tube.
scanned the pins. these represent all the detectable pn junctions between pins.
resistances
two reasonable candidates for the power connections
this is starting to feel familiar
it's a 555 timer!
so presumably the 7304 marking is the date code. 4th week of 1973. fairly early, the 555 came out in 1972! but why would they call it the 52B? strange.
and what is the 01003 all about?
this die image from wikipedia, purported to be the first 555 timer, has "1000" marked in the lower left corner. so maybe my device is mask revision 3?
fwiw, there is a paper label over the end of the tube, marked "52B"
how about i prove what this chip is, once and for all?
let's try and decap this little guy!
how? let's start by sanding it down.
it's starting to get thinner.
if you look closely, you can see the first spidery hints of the gold bond wires.
up until this point, the chip would still work as expected. but here the sandpaper has just kissed one or two of the bond wires and severed them.
all the bond wires have been cut.
something new appears: our first glimpse of the die--the fleck of silicon containing the circuit--flanked by the shadowy forms of the metal lead frame.
shall i keep going?
one false move and the sandpaper could shatter the fragile silicon
a little closer. you can see that i'm putting a little bit too much pressure on the area in the lower right, so i'll have to adjust my sanding technique.
the lead frame has gold plating so that the gold bond wires adhere better. i've also sanded halfway through the ball bonds on top of the die, thus the fat gold dots.
😬oh no
i'm going to try to remove the material from around the die and on top of it using a fresh exacto knife. i wonder if this will work, or if the die will just get all scratched up.
yeah, no dangerous chemistry here.
ok, i uhh have some news...
it's definitely a 555 timer! and an original, early production Signetics one!
compare my pic to the one from Wikipedia
dang, and i have a whole tube of them! I'M RICH haha
maybe i should give this one to @kenshirriff since he has a much better microscope and can take good photos.
oh yeah and this whole process took more like 3 hours. i just wasn't sure it would work, so i didn't live tweet it. mechanical decapping like this is always touch and go
the one question i still don't have an answer to is why these parts are marked "52B" when the part is clearly an NE555. maybe a customer wanted a house marking?
i'd love to be able to ask the IC's designer, Hans Camenzind, but he passed away in 2012. one time i had a chance to meet him, and he told me that he never expected that his chip would become iconic.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
i'm working on a Micro Channel POST (power on self test) card. the PS/2 BIOS writes checkpoint values to an IO port, and this card will be able to read and display those values. i'll keep this thread up to date as i design it.
yesterday on the live stream i designed the address decoder. i started to design the LED display but hit a snag when i realized that the 4511 chips only do BCD to decimal and can't handle A-F.
there aren't any readily-available chips that do binary to 7-segment hexadecimal. alternatives include
❓PALs/GALs
❓a microcontroller
❓a CPLD
❓a pile of TTL
❓a ROM chip
this 1920s clock technology could be accurate to within 1 second per year, and measurements in the 1980s confirmed that at least one such clock was accurate to within a tenth of a second per year.