I can simulate either half of this circuit successfully, but when I connect them at C5 and try to run the ac analysis from the input to the LNA, I get Fatal errror: Matrix is singular.
I had issues earlier with the LNA simulation, but KSKelving showed me how I needed to increase the gminsteps to allow convergence (gmin =25). Now, when connecting the filters to the LNA, I’m back to getting the singular matrix error. AC analysis of the filter by itself (not stuffing the LNA front end and adding an AC source to the left of C5) simulates without issues by itself.
I appreciate the help, plus I’d like some pointers on how to debug these types of errors for myself in the future without having to resort to the the forum.. Thank you.
Well, if you work on AD devices, you might consider using LTspice as you can get help from the AD official forum. Op-amp models can be challenging, as sometimes they are not well-written. I modified your circuit by reintroducing the negative supply voltage into the ADA4841 and adding a 100k load at the AD8429 output. The error you received is related to solving the DC solution. A common issue is that some op-amps can operate with a single supply, but that doesn’t necessarily mean their model is designed to support single supply operation. Therefore, a major part of your challenge may be related to the model itself.
I remember your previous post. When I saw that LTspice needs gmin stepping to resolve its own model, I started to worry if this is a well-written model. However, I am not an expert in this area; I am just sharing my feelings.
Any idea why each half would simulate independently, but not when the complete circuit was simulated? Any idea why I didn’t need the negative supply to simulate it successfully in the individual case?
Finally, when I get these singular matrix-type errors, what am I looking for to debug the reason or cause for the singular matrix. Thank you for your help— I’m trying to learn to fish…