Question about the technology itself... There are ...
# sky130
l
Question about the technology itself... There are two "resistor protect" masks, the rpm (for the high-resistivity poly) and urpm (ultra high-resistivity poly). What exactly do they protect against? Poly doping or silicidation?
h
Definitely, on a poly resistor, you have to suppress the silicide formation, otherwise, you would have about 7Ohm/sq. I guess these two separate masks are used to individually select the implant level so that they can realize the two resistor levels (usually around 300Ohm/sq and 1..2kOhm/sq).
l
I meant to ask that because I want to use the rpm/urpm mask inside the loop of an inductor to reduce substrate losses. If it blocks both silicide and doping, I think it solves all the problems :)
h
There is usually not much you can do about substrate losses associated with an inductor. Just think about dimensions: the inductor has a diameter of 100…200µm, and everything you can change in the substrate happens in the top 1..2µm of the silicon, the magnetic fields go deeper than that. Your best bet is to used patterned ground shield below the inductor to have a termination for the electric field lines and you can tie the shield to a suitable potential. However, this will likely increase the parasitic cap of the inductor, and thus lowers self resonance a bit.
👍 1
l
I agree. I've been reading some papers concerning ground shield and guard rings for on-chip inductors. The consensus appears to be that PGS helps mitigating substrate losses up until the lower mm-wave band, above that you're better off without it. Some technologies apply PWell and silicide blocking masks inside the inductor to minimize substrate losses, and I thought about applying this on the sky130 :)
h
You can do this, don't expect wonders 😀 You can skip this if you want to make your life easier. Watch out for filling, though, you should block the autofill in the inductor area as much as you can.
👍 1
l
Hehehehe Good advice, avoiding DRC-induced stress 🤣