# Power factor correction simulation

#### Ande

Joined Feb 3, 2017
44
I'm attempting to design a capacitive pfc curcuit for a sub system of my final year project however, I'm having trouble with the simulation but on multisim. I keep getting convergence errors. The diagram is attached. Is there anything I'm doing wrong? The NO1 represents a relay normally open contact.
Secondly, is there a method that I can use to reduce capacitor in rush when the normally open contacts close for the first time?

#### Attachments

• 60 KB Views: 12

#### WBahn

Joined Mar 31, 2012
26,295
MOD NOTE: Moved to Homework Help.

#### WBahn

Joined Mar 31, 2012
26,295
It's likely that your problem is that you have nodes that do not have a DC path to ground during simulation initialization.

You should be able to deal with this by putting initial condition constraints on all such nodes.

#### MrAl

Joined Jun 17, 2014
7,849
I'm attempting to design a capacitive pfc curcuit for a sub system of my final year project however, I'm having trouble with the simulation but on multisim. I keep getting convergence errors. The diagram is attached. Is there anything I'm doing wrong? The NO1 represents a relay normally open contact.
Secondly, is there a method that I can use to reduce capacitor in rush when the normally open contacts close for the first time?
Hi,

If it really is a convergence error and not an ungrounded component error then you might get round this by inserting a small resistance in series with each capacitor that represents the approximate ESR of that capacitor.
Simulators dont like waveforms that go up or down too fast because the derivatives top out the floating point numerical capacity, so they generate an error rather than use a fictitious top end amplitude. The solution is you have to slow the time constant(s) down.
This could also happen when one time constant is exceptionally faster than another, which puts it into the category of a "stiff" equation. Again slowing down the fastest one usually does the trick, and it is usually not unrealistic to do so.

As far as limiting the inrush current in a real circuit, you might look into near zero crossing detection. That way you can energize the relay some time just before the zero crossing so when it closes the voltage is nearly zero.