Reverse engineered Created the schematics from the netlist for the gpio

bi_t

block here. Note that it’s slightly different than the block diagram here. The output buffer connects directly to the pad, bypassing the esd protection. #ieee-sscs-dc-23

IMHO, OUTPUT MOS Trs have to behave as primary ESD diode in this case. Since the ESD diodes size between PAD and VDD/VSS are small, then it would protect for input MOS circuit only.

@Mitch Bailey: I have definitely been guilty of the same thing, but please be careful not to use the phrase "reverse engineering" which has implications of illegality. All of this is open source, so you're just creating schematic views of the netlists we have, not reverse engineering anything.

I do like the schematic though!

@Mitch Bailey: Any chance we can get schematics for all of the I/O pads? It would really be nice to have those in the repository!

@Tim Edwards I can probably do that. sky130

gpiov2

and gf180mcu

bi_t

are the most complicated, so the others should be relatively easy. I’ll focus on the pads actually used in caravel/caravan.

Mitch, It says it has programmable drive (4-16mA) in the docs but doesnt list the pins that control the drive current. Are we sure the drive current is programmable? (Basically asking what are PDRV0 and PDRV1 pins hooked to in the caravel implementation?)

Right, the output buffer has 4 drivers, 1 is always on, PDRV0 turns on 1, and PDRV1 turns on 2, giving the option of 4mA, 8mA, 12mA, or 16mA.