Does the 68LC040 use the same die as the full 68040?

[David Cook]

Does anyone know if the 68LC040 uses the same die as the 68040. I mean, is the LC just a binned version of the full 68040 where the FPU failed testing?

I suspect the 68LC040 is actually a different die, given that there seems to be completely different mask versions than the full 68040.

Here are some die images that suggest they are actually different chips, but I don't know how much a die varies from mask to mask revision:

Motorola - XC68040RC25B - CPU Graveyard - Die shots

happytrees.org

Motorola - XC68LC040RC33B - CPU Graveyard - Die shots

happytrees.org

https://micro.magnet.fsu.edu/chipshots/motorola/mc68040medium.html

In any case, I'd be interested if anyone knows the actual answer.

- David

 

[phipli]

Does anyone know if the 68LC040 uses the same die as the 68040. I mean, is the LC just a binned version of the full 68040 where the FPU failed testing?

I suspect the 68LC040 is actually a different die, given that there seems to be completely different mask versions than the full 68040.

Here are some die images that suggest they are actually different chips, but I don't know how much a die varies from mask to mask revision:

Motorola - XC68040RC25B - CPU Graveyard - Die shots

happytrees.org

Motorola - XC68LC040RC33B - CPU Graveyard - Die shots

happytrees.org

https://micro.magnet.fsu.edu/chipshots/motorola/mc68040medium.html

In any case, I'd be interested if anyone knows the actual answer.

- David

Afternoon David,

My understanding is that the answer is "Both".

Some LC040s are full fat dies that had a fault in the FPU portion of the circuit, so they blew some fuses or whatever to disable that portion of the chip.

On the other hand, some LC040 chips have a special mask that completely excludes the FPU. These can be spotted sometimes because they have a smaller cover on the underside of the chip because the die area is smaller... but be aware that the die size also changes depending on what process your chip was made to anyway.

I have some of each type of LC040 - some of mine have a mask that implies they were intended to be full 040s, others are a specific LC040 mask. There are lists on the internet that identify if a mask is one, the other or both.

 

[David Cook]

Very interesting.

because they have a smaller cover on the underside of the chip because the die area is smaller

I wonder whether the intentional LC040 mask with smaller die per process is better at overclocking, versus the faulty 040.

 

[phipli]

Very interesting.



I wonder whether the intentional LC040 mask with smaller die per process is better at overclocking, versus the faulty 040.

I believe that they did produce less heat than a full 040, but I'm not sure about the hobbled ones - it could be either case... should do some testing!

 

[phipli]

Very interesting.



I wonder whether the intentional LC040 mask with smaller die per process is better at overclocking, versus the faulty 040.

I'm afraid I've done some rummaging through photos in my phone (its been a bit tricky because it is refusing to identify photos containing the text "XC68LC040") after someone gave me a nudge and I'm thinking that I'm wrong on the re-used masks. I'm sure I saw it on one of the websites that listed what masks ended up in what products, but after about 15 minutes of searching online and my photos I can't find any evidence.

I'm going with "Phipli is talking nonsense unless evidence of it happening is found".

Sorry about that, I... try to not say things that I haven't checked, but... I thought I'd actually looked at stuff in the past that said that. Bleh. Sorry.