I found this talk by Enrico Di Lorenzo of https://www.fastfieldsolvers.com/ quite interesting: https://peertube.f-si.org/videos/watch/81951ef3-5914-4d21-99cf-b8019e698711

A heads up to anyone trying to use the RF PMOS devices, there's an issue with the pin ordering. It can easily be fixed by altering the sub-circuit but incredibly frustrating if you're not aware of the issue - I know from experience https://github.com/google/skywater-pdk/issues/212

That must have been hell to debug.

@User has come cross some important information regarding the RDL and future shuttle runs. They tagged the channel but I didn't receive a notification. Sharing in case people haven't seen it.

@User: You can try to persuade Tim Ansell otherwise. I was in favor of removing the four bump bonds from the middle of the die and using that space for inductor tests. However, keeping costs down seems to be the overriding factor here. Personally, I don't see it as a huge overhead because the masks for the bump bonds are the eqivalent of a 2um process or something like that and are much cheaper than the SkyWater masks. You can try to lobby Tim Ansell with all the great radio projects you could potentially do. Personally, I would prefer to set aside one of the reticle areas with an experimental inductor. Ultimately, even if we decided to do it in the future, I think there is just not enough time for this shuttle. From a thread in #sky130

@User If on subsequent tape-outs people will not have the automatic option of putting more inductors on the RDL, then I'm not going to push to do it this time to characterise/test them. For next time it would be interesting to quantify the mask cost though - individual users might think it worth paying. Anyway, if RDL inductors are not going to happen then it would be good to announce that in #radio so that people have time to adjust their designs accordingly. Provided there is a >=500um diameter region somewhere with no RDL or bumps above it, the Skywater layers are thick enough to make usable inductors, albeit probably a bit worse than the RDL. From a thread in #sky130

@here any recommendations for EM simulators

?

Does the PDK provide the eps of all layers?

@User The stackup diagram provides all the dielectric constants and the resistivity is available in the PDK docs. Alternatively have a look at this file It contains all the required data to reconstruct the geometric structure for meshing in a convenient place.

@User For EM simulation, the open source ones that are most accessible are openEMS, and FastHenry2 + FaserCap. Geometry description in FH can be a pain if the design is complicated. openEMS needs processing of the GDS files to extract the polygons. KLayout does a good job here. It has a CSG engine that is easy to work with. Here's a sample setup file that should get you started. https://github.com/diadatp/sky130_rf_tools/blob/main/openems/main.m Be prepared to wait a long time for simulations to run. Especially if you set the correct boundary conditions.

Hi @here Posted in #analog-design

We are happy to announce that we have created a script that installs all the tools necessary for starting with Skywaters 130 Posted in #analog-design

mabrains/sky130_ubuntu_setup Ubuntu scripts that are used for setting up your machine for Sky130 designs. mabrains/sky130_ubuntu_setup

@User has left the channel

@User has left the channel

@Nathan Olson has left the channel

@vks has left the channel

@Michael Langford has left the channel

Hello, does anubody know how to install sky pdk in order to use rf devices?

@SimonM Use volare

I do use it. Can u take a picure havin rf inductor in your xschem design please?

I don't believe there is a symbol for inductors

In raw sky pdk there is folder and gds for rf components but they are not implemented to xschem

Yes, that's true

But I wouldn't use the one that comes with PDK. I don't think it's reliable.

How can i implement it into xschem?

It's not just that the inductor in the PDK is unreliable, it's that it does not correspond to the process. The original Cypress S8 process had some process variants with thick metal for inductors, and the ones in the PDK come from (and are characterized for) that process variant, which is not equivalent to sky130 (the thick metal in that variant was something like 3um or more, whereas the sky130 process has a top metal thickness of about 1 1/2um). The process variant used the same GDS layer:purpose pair for top metal, so the device will read in from the library GDS into the sky130 technology, but the device model is completely wrong for that layout.

@Mohammed Fayiz Ferosh has left the channel

@Abhinav Uppal has left the channel