Can I get a hand with this inverter loop simulation: <https://github.com/mattvenn/instrumented_adder...
m
Can I get a hand with this inverter loop simulation: https://github.com/mattvenn/instrumented_adder/blob/sim-inverter-loop/docs/sim.md ngspice fails to converge
h
I would first look into this warning
Warning: singular matrix:  check nodes reset and reset
Do you have an idea why this could be?
Why does it say “reset and reset”? are there two nets with same name somehow confused?
m
where was that warning?!
yes we just spotted reset name was incorrect
oscillating!
image.png
🙌 1
h
The warning was right in the spice output file that you linked: https://github.com/mattvenn/instrumented_adder/blob/sim-inverter-loop/spice/spice.log If sim fails that is the first thing I check, a somehow malformed netlist.
1
m
thanks Harald
1
s
@Matt Venn i see in the adder netlist some nodes have escapes on brackets , some haven't, like:
outputs[1] _15_/X
buffers\[5\]/A
So it seems the naming is a bit inconsistent. However this is not the problem, i think. Another issue to consider, in the
instrumented_adder.spice
file the
reset
pin is only declared in the .subckt line and never used, so it seems no resettable flops are used in the design.
m
Thanks Stefan, yes we figured it out, the rst line wasn't connected
s
@Matt Venn sorry, the reset issue is a false flag, it is used, indeed.
m
I fixed it last night, its working as expected now
👍 1