I am designing a constant current source to charge my battery in a simulink model. I am using a BJT as a constant current source in one case and a MOSFET mirror in the other case. I receive a constant current output till the load isn't connected, but as soon as a load is added, the current shows a ripple for the initial time period.
Akash Sharma I wonder about the path of the constant current without load. Where does it flow if not through a load?
I assume that the circuit with load is oscillating; in this case as a first step, I would try to use transistors with lower gain.
Joerg Fricke I am using an ammeter to measure the current and I can see the response in the scope block connected to the ammeter.
On adding a series resistance with the ammeter, there is an initial spike in the output current.
Akash Sharma If I understand correctly, in the first case the load has a small (or even zero) resistance (the ammeter), in the second case the resistance is considerably higher.
If you try to (temporally) zoom into the spike, does it look like just one (unipolar) peak or more like a damped oscillation?
I guess you have what is called overshoot.
Current sources usually being closed feedback systems may suffer from that.
Usually, it can be avoided if limiting the feedback bandwidth by using bypass capacitor (attached is an example).
Akash Sharma What strikes me about the signal is the fact that it contains two flat steps of equal duration (the second one only barely discernible). This doesn't look like the behavior of an analog circuit, so I guess it is an artefact of the simulation.
I'm not familiar with Simulink. Can you adjust the temporal step width somehow? I would expect the overshoot to become shorter and virtually to disappear if you decrease the duration of the simulation steps further and further. (However, the simulation time might increase linearly.)
- Badges
- Science topic