https://open-source-silicon.dev logo
Join Slack
Powered by
Hello I'm doing layout on PNP device and encounter...
# lvs
z
Hello I'm doing layout on PNP device and encountered property errors when running LVS:
Copy code
Circuit 2 parallel/series network does not match Circuit 1
Circuit 1 instance /XM_bjt_out_0//sky130_fd_pr__rf_pnp_05v5_W3p40L3p40_0[0|0]/sky130_fd_pr__pnp_05v5_W3p40L3p40:0 network:
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
Circuit 2 parallel/series network does not match Circuit 1
Circuit 1 instance /XM_bjt_out_0//sky130_fd_pr__rf_pnp_05v5_W3p40L3p40_0[0|0]/sky130_fd_pr__pnp_05v5_W3p40L3p40:0 network:
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
  M = 1
Circuit 2 parallel/series network does not match Circuit 1
Circuit 1 instance /XM_bjt_out_0//sky130_fd_pr__rf_pnp_05v5_W3p40L3p40_0[0|0]/sky130_fd_pr__pnp_05v5_W3p40L3p40:0 network:....
All other LVS netlist/instance checks are passed so I'm really confused on why this is happening. Does anyone know what causes the issue? Thank you in advance.
BGR_lvs_magic.spicecompare_outBGR_lvs.spiceBGR_lvs.mag
t
I'd guess that the relevant line in the compare_out file is
Copy code
sky130_fd_pr__pnp_05v5_W3p40L3p40 (50->4)  |sky130_fd_pr__pnp_05v5_W3p40L3p40 (37->4)
meaning that the layout started with 50 PNP instances and knocked them down to 4 by parallel combination, while the schematic had 36 PNP instances and knocked them down to 4 by parallel combination. The "M = 1" entries are definitely confusing and I should take a look at how that output could be improved, but the number of those entries is, I think, related to the difference between layout and schematic PNP instances. It matched a bunch of them but the "M = 1" entries are the set of PNP devices in the layout that were left over after matching all of the others.
Ideally netgen should report this as a mismatch of emitter area between the two sides. The problem is that the device model is specific to that one specific size of PNP device, so there's no emitter area passed to the model (subcircuit). From netgen's perspective, it's a pure black box. That doesn't excuse netgen from not producing more helpful output, though.
z
THank you!! I found in my schematic the BJT mult is missing. I've hence correct the issue!! Many thanks Tim as always
3 Views
Open in Slack
PreviousNext
Powered by