Why filters with current source behave differently?

[Claude Abraham]

A "true" CCS or CVS delivers a fixed output of I/V in spite of changing load impedance, over a limited range of compliance. An "ideal" CVS/CCS delivers fixed output over unlimited range of compliance. This is unattainable in practice because the energy conversion that drives the source is limited. The CVS turbines that power the commercial grid have immense power ability, but nonetheless limited. If a transmission line touches ground due to a vehicle striking a utility pole, the voltage of that line wrt earth drops. The generators have resistance, synchronous reactance, and the turbines can only exert a limited amount of force. The t-lines have R & L so the current capability is limited. A line to earth short can result in kilo-amps of fault current but not mega-amps.

But within a limited range of compliance, the power grid is a true CVS, but not an ideal one. A photo-diode terminated in a low Z, is a true CCS, not ideal. If the load impedance varies from 1.0 ohm, to 0.10 ohm, to 0.010 ohm, to 1.0 milliohm, etc., the PD delievers a constant current proportional to incident light, while V varies directly w/ Zload. It does, however, have a finite range of impedance over which it can maintain constant current.

In the high Z terminated mode, the PD can look like a CVS. If Zload is varied from 100 kohm, to 1.0 Mohm, to 10 Mohm, etc., the V stays pretty fixed while I varies inversely w/ Zload. So I hope I've clarified the matter a bit.

The CCS & CVS in the ideal realm are mental abstractions. They don't exist but conceptualizing them helps us gain insight into electrical science. In the realm of the actual non-ideal physical world with imperfections all around, a "true" CVS or CCS can be attained, but with the understanding that its compliance is limited. It can behave as a true CVS/CCS over a limited range of terminating impedance values. Thus the PD can attain "true" CCS or CVS behavior if the terminating impedance is within its compliance range, as well as the incident light being sufficient in power magnitude.

Inductors/capacitors are "true" CCS/CVS resp., albeit, not ideal sources. Hopefully this clears things up at least a little. BR.

 

[MrAl]

Hi,

If we are to state that an inductor is a current source then we must also state that it is only a current source for a very limited time. The definition:
V=L*di/dt

and hence:
di=V*dt/L

is only true when dt is small such that it forces di to be small, di being the change in current which we ideally want to be zero. So the inductor is only a true current source for an infinitesimally short time period, which we allow to be a little larger in the approximation.

 

[Claude Abraham]

Hi,

If we are to state that an inductor is a current source then we must also state that it is only a current source for a very limited time. The definition:
V=L*di/dt

and hence:
di=V*dt/L

is only true when dt is small such that it forces di to be small, di being the change in current which we ideally want to be zero. So the inductor is only a true current source for an infinitesimally short time period, which we allow to be a little larger in the approximation.

Of course you are correct, but it deserves mention that capacitors are voltage sources only for a limited time as well. Those skilled in the art of electrical science know that if one is using an inductor in a constant current LED driver, that the inductor value must be chosen to get the right time constant, i.e. switching period, load impedance affect the rate of de-energizing. If the ripple is kept under control, the inductor will maintain constant current to a good enough approximation. After the current droops a pre-determined amount, the inductor will be refreshed by the input power source.

Likewise, a cap used in a linear full-wave rectified supply exhibits ripple voltage. But we can choose the cap value so that the time constant results in a limited droop. After the cap voltage has dropped a small amount, the next ac cycle from the input supply refreshes said cap. I realize that L & C have time constants but we don't use them w/o refreshing them. If the time constant is chosen to produce an amount of droop that is acceptable, we refresh on the next cycle. This way, the L/C maintains steady I/V with a small amount of incurred ripple. We just refresh every cycle and all is well.

 

[Winterstone]

To Claude Abraham:

Thank you for the very detailed answer.
However - the only thing I wanted to know is your definition of "true" because there is - as far as I know - no general definition of this term in conjunction with electronic properties.
With one sentence: "True" is identical to "real" with deviations (if compared with "ideal") that are within acceptable limits (depending on the particular case).
I think, such a clarification is necessary in order to avoid misinterpretations.

I am aware that no ideal CCS and CVS exist - and that is no surprise at all because in electronics there is no real part or device that behaves "ideal" (like it should).
Nevertheless, we speak of linear circuits (but they are not) and about squarewave signals (which are not square waves). That`s the common way of thinking like an engineer.
More than that - most of the formulas applicable in the field of electronics are not correct by 100%.

Regards
W.