After clearing out an old cupboard at work, I stumbled across a couple of old Motorola MC68705P3S microprocessors. What immediately caught my attention with these chips compared to others, was that they had a ceramic package which is not very common. Additionally, they have a glass window over the chip’s die (more details on the reason for this later) allowing you to peer deep inside the chip.

Top view of the chip clearly showing all of its markings.

Without diving into the device’s datasheet, interpreting the markings I know straight away that, “MC68705P3S” is the model of this chip. From the bottom line “47E8504”, I’m going to take a well educated guess that “8504” refers to the device being manufactured in week “04” of the year 1985 – this is very common method of date stamping the chip. The “47E” may refer to some sort of specialty or batch code, details may be within the datasheet.
However, before getting too technical and whilst I can’t find a good quality copy of the device’s datasheet yet, I thought I’d get a nice close up of the chip’s die. Doing so revealed some nice surprises (click on the images for the higher res images).

In the top right hand corner of the chip is Motorola’s “M” symbol etched into the die which is a cool feature – perhaps a way to hallmark the device and prevent cloning. Also at the top of the die in text that is upside down is the chip’s model number “MC 687005P” – its always a good idea to label your work. Also, there is a copyright mark “© 1982”, which supports the assumption that the chip was manufactured in 1985.

So why the glass window? Well, this device was developed before EEPROM (Electrically Erasable Programmable Read Only Memory) existed and at the time there was EPROM (Erasable Programmable Read Only Memory). To change this device’s program memory, was a bit more involved than a modern device with EEPROM memory where the erasure and re-writing can be done in a matter of seconds. With an EPROM device, erasure was completed by exposing the chip to a UV lamp for about 30 minutes (or direct sunlight for a few hours). After this time, the chip was able to be reprogrammed. (To prevent accidental erasure the windows can be covered with black electrical tape.)
Nowadays, waiting 30 minutes to erase a chip might seem slow and archaic. You need to remember though, at the time some programmable chips didn’t have a glass window and hence were not able to have their memory cleared. They could only be programmed once. If you got your code wrong and the chip didn’t operate how you wanted, then the chip was useless and you needed to start again with a new chip.
As a hobbyist in the 80s I had (and still have) several of these chips, along with the eraser and the programmer. The programmer I actually built from a kit. I believe it still works, and just ordered a USB -> Parallel adapter to try it out again.
What I would do is have 4 or 5 chips sitting under UV while I wrote code/programmed/tested the new version. By the time I was done with whatever version (even if it was just a couple of lines of code), one of the chips would be erased and ready to use. So you’d never really wait 20-30 minutes.
I designed and wrote some code so my 1986 corvette would hook to this thing. I hooked the transmission pulse up to it so I could count output shaft RPM. From that you can calculate vehicle speed, and this is exactly how the speedometer worked in the C4 vette. What I used it for, however, was to calculate 0-60 time (and ok, 0-100mph time). It was just a matter of calculating the time between pulses to determine current speed. Combine that with some internal clocks and you have a basic stopwatch.
The main problem, however, was getting that first pulse, since you’d already be moving, so it wasn’t perfect.
LikeLike