Creating a reliable VCO Model

[stantheman286]

Hi everyone,

I've been searching the forums, and there's a lot of good VCO information, but I haven't been able to solve my issues yet.

I'm working on a control system for a DC-AC inverter, which controls a voltage that sets the amplitude of the sine wave in our sine-triangle PWM circuit.  Unfortunately, I haven't been able to find a model for our VCO (XR2206).  I'm interested in just controlling the amplitude, not the frequency of the sine wave.

I have tried to create a VCO model using a Wien Bridge (see attached circuit).  By itself, it works fairly well (there's a little bit of saturation, but that's expected).  However, as soon as I hook my control circuit model up to my VCO, I get convergence and timestep errors.  None of the tricks I've seen in the NI documentation for timestep too small errors have worked so far.  I've tried to buffer the control output, but it still causes errors.  If the control is run by itself with a load, there are no issues.

Any advice on fixing the VCO model or a good one to use?

I'd really appreciate it 🙂

Thanks!

Matt

[Tayyab_R]

Hi,

The best thing to do when you get a timestep too small error is to check which component is causing the error. What you can do is go to Simulate >> Simulation error log/audit report. This will show you where the error is originating from. If you see xu1, that means refdes u1. I checked this and it seems that the virtual 5T opamp was having issue. I replaced that with a real opamp just for checking and the simulation started to run. Please take a look at the attached modified circuit. You should be able to simulate the complete system.

Hope this helps.

[stantheman286]

Hey Tayyab,

Thanks for the suggestion.  Yea, I actually went through each piece of my VCO and also discovered that it was the virtual devices that were causing the problem, so I switched to using real opamps instead of virtual opamps and VCVS's.  I guess that's why my teacher warned us to use real components 🙂

Thanks!

Matt