4 ms·

Author here if anyone has questions...

by kens 6y ago

jhfdbkofdcho 6y ago
Saw this listing on eBay with what looks like other chips from the same family. https://www.ebay.com/itm/OQ-1xx-Vintage-GOLD-ceramic-IC-for-collection-or-scrap-like-Intel-C4004-/254833220887?_trksid=p2349624.m46890.l49286 https://www.ebay.com/itm/OQ-1xx-Vintage-GOLD-ceramic-IC-for-...
- kens 6y ago
  That's where the original chip came from!
mysterydip 6y ago
Love this kind of stuff, thanks! What's the upper limit to analyzing a chip like this? How complex or small can the components get before it's not feasible?
- kens 6y ago
  The limit for me is the early 1980s. After that, the feature sizes get too small for my microscope. (Red light has a wavelength of 700nm, so things start to get difficult below about 1 micrometer.) And when chips have two or more layers of metal, it is much harder to follow the wiring. Companies that do chip analysis commercially, and have tools such as electron microscopes, can process much more complex chips. The limit for them is probably how much you're willing to pay.
  - Koshkin 6y ago
    There was an interesting post here on HN recently about a home-built STM. https://news.ycombinator.com/item?id=26740968 https://news.ycombinator.com/item?id=26740968
  - SAI_Peregrinus 6y ago
    Christopher Tarnovsky has given a few DEF CON and BlackHat talks (eg [1]) where he describes reverse-engineering Trusted Platform Module(TPM) and smartcard chips. It's a good demo on the capabilities of the sort of thing where companies are limited only by the time they can afford to take. [1] https://www.youtube.com/watch?v=Ed_9p7E4jIE https://www.youtube.com/watch?v=Ed_9p7E4jIE
userbinator 6y ago
Have you tried looking up "ASP104" or "ASP104N1" in any of your databooks/catalogs? That's the marking on the die and might yield some closer clues. Unfortunately this is the type of thing that Google is absolutely horrible for finding information about, although it wasn't that bad a decade or two ago. Asking on some electronics forums might also get some better suggestions or even an exact ID. I doubt the "P" is for "Philips" but just indicates Pin 1.
- kens 6y ago
  I've tried slicing and dicing the part numbers, but couldn't come up with anything. I've found that Google is not bad; the biggest problem is IC suppliers who list basically every possible part number with no information. I think their business model is that if you want to buy the part for a high price, then they will see if they can find it somewhere.
PhantomGremlin 6y ago
See my comment at Yet Another Dupe. I don't believe that most of the die is "unused". I believe it's being used as a capacitor. https://news.ycombinator.com/item?id=26847762 https://news.ycombinator.com/item?id=26847762

Author here if anyone has questions...

by kens 6y ago

Author here if anyone has questions...

Saw this listing on eBay with what looks like other chips from the same family. https://www.ebay.com/itm/OQ-1xx-Vintage-GOLD-ceramic-IC-for-collection-or-scrap-like-Intel-C4004-/254833220887?_trksid=p2349624.m46890.l49286 https://www.ebay.com/itm/OQ-1xx-Vintage-GOLD-ceramic-IC-for-...

kens 6y ago
That's where the original chip came from!

mysterydip 6y ago

Love this kind of stuff, thanks! What's the upper limit to analyzing a chip like this? How complex or small can the components get before it's not feasible?

kens 6y ago
The limit for me is the early 1980s. After that, the feature sizes get too small for my microscope. (Red light has a wavelength of 700nm, so things start to get difficult below about 1 micrometer.) And when chips have two or more layers of metal, it is much harder to follow the wiring. Companies that do chip analysis commercially, and have tools such as electron microscopes, can process much more complex chips. The limit for them is probably how much you're willing to pay.
- Koshkin 6y ago
  There was an interesting post here on HN recently about a home-built STM. https://news.ycombinator.com/item?id=26740968 https://news.ycombinator.com/item?id=26740968
- SAI_Peregrinus 6y ago
  Christopher Tarnovsky has given a few DEF CON and BlackHat talks (eg [1]) where he describes reverse-engineering Trusted Platform Module(TPM) and smartcard chips. It's a good demo on the capabilities of the sort of thing where companies are limited only by the time they can afford to take. [1] https://www.youtube.com/watch?v=Ed_9p7E4jIE https://www.youtube.com/watch?v=Ed_9p7E4jIE

userbinator 6y ago

Have you tried looking up "ASP104" or "ASP104N1" in any of your databooks/catalogs? That's the marking on the die and might yield some closer clues. Unfortunately this is the type of thing that Google is absolutely horrible for finding information about, although it wasn't that bad a decade or two ago. Asking on some electronics forums might also get some better suggestions or even an exact ID. I doubt the "P" is for "Philips" but just indicates Pin 1.

kens 6y ago
I've tried slicing and dicing the part numbers, but couldn't come up with anything. I've found that Google is not bad; the biggest problem is IC suppliers who list basically every possible part number with no information. I think their business model is that if you want to buy the part for a high price, then they will see if they can find it somewhere.

PhantomGremlin 6y ago

See my comment at Yet Another Dupe. I don't believe that most of the die is "unused". I believe it's being used as a capacitor. https://news.ycombinator.com/item?id=26847762 https://news.ycombinator.com/item?id=26847762