wafer.space Community - Ideas for different size silicon area.

8:02 p.m.

To be a little clearer;

(a) Eventually I hope that I can support any die size and use something like laser dicing. For now that is an added complexity / risk it doesn't make sense to take on (kinda like the gf180mcu process optimization idea that I abandoned).
- (b) In the shorter term, given that you get back bare die by default, I believe there could be a few options for post processing to split them into even smaller units.

8:04 p.m.

At the moment it is $7 USD per die -- so if we could divide the die into 8 parts, that would make it under $1 per die.

8:05 p.m.

Given that 3 * BOM cost;

$7 USD per die really means a >$100 USD product.

$1 USD per die means that <$30 USD or even $10 USD type products become viable.

Ok, I wasn't really sure that dicing "a second time" was something doable at all at scale ...

@tnt - "scale" is a relative term, ~1,000 units it pretty small for PCBA and non-existant scale for semiconductor people.

8:13 p.m.

Actually, if it's $7 USD per die and you want $1 USD per sub-die, then you can spend ~$1 USD per die to do the dicing.

By scale I meant, not 10 units done by hand. Because if you get the die in a gel pack like we dit from the skywater runs, I'm not sure how you put them in a laser dicing machine "in bulk" since the ones I've seen take full wafers ...

8:18 p.m.

But it's not like I have tons of experience with dicing so it's very possible there is a way