Ground VS Closed Loop Battery Circuit Simulation

Thread Starter

CineDroid

Joined Jan 25, 2020
20
I read everywhere that in a closed circuit running on positive voltage, powered by a battery, the negative pole of the battery is effectively ground. I'm playing with buck and boost converters in Falstad and I realize I get wildly different results depending on wether I'm using a +xxV source and ground pins or I'm drawing the actual leads to the negative pole of the battery.

For the past week and a half I've been pulling hair thinking I couldn't just even get an ideal buck converter to work but tonight I stopped using ground and made an actual closed loop, and I finally got the expected voltage on the load side.

Then I did the same experiment with the ideal boost converter that I got working last week and realized that in a closed loop the voltage on the load side is 10V higher than when using ground!

I'm very confused now :/
 

Attachments

crutschow

Joined Mar 14, 2008
27,394
Simulators do not like floating nodes.
All nodes must have some path to ground, even if it's through a very high resistance (to simulate isolation).
 

ci139

Joined Jul 11, 2016
1,696
The behaviour of the Falstad also depends on the iteration time constant - and it's mostly 10-s of times off the actual time.
MENU > Options > Other options... -- it does not "play right" with the above-equal to 1....5 ms time steps (especially if there is small duration events to react to)

To make the Falstad more realistic add small resistors in series with capacitors . . .
I found that the small signal transistor model (such as for 2N390x , BC337 , BC847) gets better if you add some 9Ω in series with the emitter and 3.6pF from base to collector . . . ← makes a complex circuitry a lot slower but also prevents above multi-MHz oscillations

- also the wave-forms of the oscillations (square-wave) get weird in Falstad -- partially hanging or ramping linearly or suddenly jumping in frequency/wave-form

it's a good online tool to explain the principles of the circuits - but not too good for the SMPS statistics
 
Last edited:
Top