LTSpice transient analysis of inverter issues

Thread Starter

Joseph1991

Joined Apr 5, 2022
3
I'm having issues with trying to step through from 0 to 500nF in steps of 50nF. The simulation shows the first three steps and then starts choking up. Lowering the steps to 200nF lets me get results faster but then I'd have to separately simulate from 200nF to 400nF and then 400nF to 500nF. I'm trying to calculate rise and fall times as well as the 50% times so it would be inconvenient trying to stitch all the results together.
Is there any way to speed this simulation up? It's only using 1 thread and 550MB of my RAM. The simulation is that of an inverter using PMOS and NMOS transistors and I'm calculating Rise and Fall times at 20% and 80% voltages as well as finding when 50% voltage is reached.

I'm on:
Windows 10 LTSC, Ryzen 5 3600, 16 GB RAM DDR4 2400, 1TB SATA SSD, RX560 4GB
 

Attachments

crutschow

Joined Mar 14, 2008
29,785
Try using the Skip initial operating point solution command in the transient analysis.
Sometime Spice has a problem trying to find the initial bias conditions, and that's not needed for this simulation since that's primarily for determining the bias voltages in an analog circuit (similar to finding the DC op pnt).
That's usually the first thing I try if I have a simulation hang-up.

1650240176517.png

1650240093160.png
 

Thread Starter

Joseph1991

Joined Apr 5, 2022
3
Try using the Skip initial operating point solution command in the transient analysis.
Sometime Spice has a problem trying to find the initial bias conditions, and that's not needed for this simulation since that's primarily for determining the bias voltages in an analog circuit (similar to finding the DC op pnt).
That's usually the first thing I try if I have a simulation hang-up.
This definitely did the operations very quick but the results are wrong. They need to look more like in the "Right.PNG" or even "RoughlyRight" photos.
Both the rise and the fall are needed and even the results for the rise that I got don't match up compared to both when I painstakingly measured the 20% and 80% Voltages with cursors and manually changing the capacitance as well as the results that I got for just the 0n to 200n using the SPICE directives.
 

Attachments

Thread Starter

Joseph1991

Joined Apr 5, 2022
3
Alright I figured it out. Someone else told me that I should switch from the "Normal" to the "Alternate" solver and that fixed the problem. Mods can close this thread now.
 

Attachments

Top