open-source-silicon.dev

I'm seeing an error while writing GDS: `CIF error in cell twin_tee, layer RPM: parent and child disagree on CIF`

image.png

gds is here: &lt;<https://github.com/mattvenn/tt07-twin-tee-opamp-osc/blob/main/gds/twin_tee.gds>&gt;

and the mag is here: <https://github.com/mattvenn/tt07-twin-tee-opamp-osc/blob/main/mag/twin_tee.mag>

I think this happened because I was cleaning up the directory and accidentally removed a resistor before commiting it

then I had a hard time re-creating it for some reason, and in the process I also lost the top 2 resistors.

So I had to make them again with pcell generator

I don't know why it's just the right hand cell

rpm.png

<@U016EM8L91B> Here’s the resulting gds RPM layer.
The synthesis for RPM includes `grow-min` but `grow-min` says that it should only be used on simple shapes.
``` layer RPM
        bloat-all xhrpoly xpc
        grow 200
        grow-min 1270
        grow 420
        shrink 420
        calma 86 20```
&gt; The problem with *grow-min* is that it only works well for simple, isolated rectangles.
Maybe the pcell generator shouldn’t make resistors with bends.

<@U0172QZ342D> you only included the top level mag file. Can’t see what’s in the resistor pcell.
You might try `cif see RPM` to see what gds data will be generated in magic.

<@U0172QZ342D>:  "Parent and child disagree on CIF" is not _necessarily_ an error.  It means that the GDS writing rules (namely statements like `grow-min`) are inconsistent in how they are applied to a cell when the cell is by itself vs. when the cell is inside a larger cell.  The result _can_ be an error but usually isn't.  If you run the klayout DRC deck and it doesn't identify a DRC error related to RPM then the layout should be okay.

ok, I fixed a bunch of other errors that I had, and got rid of the gds error by moving the resistor