open-source-silicon.dev

EF_DACR1001.gds

image.png

<@U016EM8L91B> <@U017X0NM2E7> Excuse me, can you please let me know how to fix this DRC error? Is this an error caused by a MAGIC bug?, Can we waive this kind of DRC error?
Can you please guide me on how to fix it, please?
Noted: The layout is a 10-bit DAC, which is developed by reusing the pre-designed S-string 8-bit DAC.

FYI, this is *not* the way to make a DAC.  8 bits is about the limit on a tapped-resistor design.  Expanding such a design will likely get you 8 bits plus two bits of noise.  But expanding such a design by repeating blocks will get you much less than 8 bits, with massive discontinuities at the upper two bits. (Also attn. <@U03UV5NUV6Z>)

Otherwise, I think that the dnwell-to-nwell spacing rule implementation might end up being dependent on artefacts that show up in a hierarchy.  I will review the rule implementation.

<@U016EM8L91B> Thanks for you fast response.
• Noted that the layout is clean. However, once I use it hierarchy, it shows this error. Is there a reason for that? 
• Do you recommend me to use another structure instead of R-string structure to build a 10-DAC circuit! for example using C2C DAC or Charge-redistribution DAC  structures. Which structure is better to build 10-bit DAC or more from your point of view, please?

As far as I know a C2C DAC _is_ a charge-redistribution DAC.

Yes, right. Do you recommend to use the C2C DAC structure to build a 10-DAC circuit instead of using R-string structure ?

Yes.  The mismatch on capacitors (especially MiM capacitors) is much more tightly controlled than the mismatch on resistors.

<@U03ED7UNVFH>:  Back to the original discussion of the false-positive DRC error:  I confirmed that this is a bug in magic, a pretty long-standing one, and probably accounts for some other mysterious false-positive DRC errors as well.  It's an obscure problem caused by failing to clip the search area for chained DRC rules.  Once I understood that it was searching outside the area of valid layout, it was easy to understand why it was wrong and how it needed fixing.  I have corrected the error in magic version 8.3.426.

<@U016EM8L91B>  Thank you. Totally appreciated.