Use of an integrator/differentiator block requires selection of a time constant for the block. How can one go about selecting the optimal time constant T?
E.g. a 5kHz, 2Vpp, duty 0.5 triangular signal is differentiated. It passes through a 50kHz LPF (a requirement). The following are outputs (Vopp) of the differentiator block with respective Time (T) constant:
(1). T = 1, Vopp = 20,000
(2). T = 0.1, Vopp = 2,000
(3). T = 0.01, Vopp = 200
(4). T = 0.001, Vopp = 20
My confusion is about which one gives the correct answer of the differentiation of a triangular signal?
You will be surprised: ALL answers are correct (except IMO for a factor of 2. But this would be negligible for the understanding of the following).
Your signal: 2 Vpp @ 5 Khz - symmetrical is rising resp. falling with 2 V/0.1 ms
= 20 V/ms = 200 V/10 ms = ...
OK - peak-to-peak I'd expect 4xx V, but maybe you'd have to set the time constant to 2xx to get this value :)
I have no experience with PSIM but from the description at wikipedia my best guess is that it is fast, but not suitable for analog circuit simulation. To overcome this they obviously added this time constant 'workaround'.
Consequently the output is not what you expect, but the same is valid for a real world hardware implementation which is 'adjusted' with resistor(s) and capacitor(s). (The same is valid in principle for integrators.)
Now it's up to you to choose what you can handle best.
In a time discrete system I would choose the time constant to equal the sampling time first, but this is a deliberate decision to be evolved during further refinements.
U. Dreher, you are right there, the value should be 2Vp (0 to 2V or -1 to 1) and not 2Vpp (my mistake in typing).
- Badges
- Science topic
- Similar topics
- Control Systems Engineering
- Filtering