There was a Stanford project on one of the early ChipIgnite runs that was using a 2 GHz transfer, although I don't remember if that was input, output, or both. 2 GHz is impossible on the GPIOs. It should be able to be achieved on the caravan bare pads, since they are not capacitively loaded.

Is this pad part ok to use if I add it to caraval? @Tim Edwards . Can u share me the pad part name to be added in my gds which don’t have diodes etc so I can hook up high speed signals

That pad would work. Are you planning to do your own wirebonding or use probes?

I am doing flip chip . Thanks I will continue to use these pads

The empty space underneath on metal 4 is mainly for avoiding problems with wirebonding. For a flip-chip solution, you can probably get by with something simpler with just metal 5 and passivation opening. You shouldn't need the metal 4 ring around the edge, or have to keep out of the metal 4 region underneath the pad.

@Tim Edwards is there a part as u describe which I can use from pdk? no ring in metal 4 and just metal 5 with oxide open. or i just delete metal 4 ring and make it via to metal 5?

Just copy the cell and delete the metal 4 ring.

Thanks, will do

@samarth jain pad.dg is the glass opening. Layer 76/20. Put that where you want bare metal.

@David Lindley I have placed the Met5 and PAD drawing overlapping with 0.27 enclosure as shown in the snap shot below. But after the drc run i see some violations on these COuld you please suggest on solution or can it be waived

We don't check either of those pad rules. You should be good.