Impedance compensation needed for active speakers?

[tcpip]

Another newbie question...

Considering how the non-linear impedance of a driver will result in non-uniform power delivery from amps (specially solid-state ones), is it necessary/preferable to add passive impedance compensation circuits between the power amp and the driver even if I'm going with line-level xo?

I know most full-range lovers can't imagine having anything but straight wire between power amp and driver, but I can't figure out how you're expected to get a uniform power output from the driver without impedance compensation. Maybe it's a simple theoretical confusion I have? I'd be really grateful for any light on the matter. Thanks.

[Jennice]

The sensitivity is, historically, stated as xx dB at 1 W, 1 meter.
However, due to this problem you mention, many specify the driver sensitivity as xx dB at 2.83 VOLT, 1 meter. 2.83V gives 1W at 8 ohm, but the dB/V rating is not sensitive to driver impedance.
This is how you get away with the "simple" solution of source, amp, driver-unit for full range speakers.

Jennice

[richie00boy]

Impedance compensation is pretty pointless for voltage driven sources, as 99.9% of amps are. Unless the compensation provides some means of making the 'power factor' a bit nicer and the amp struggles anyway, the sound should be unchanged.

[phase_accurate]

The only passive impedance correction of which I expect being useful in active speakers - is a Zobel compensating for the impedance-rise due to the driver's Le.

Regards

Charles

[tcpip]

Originally posted by phase_accurate
The only passive impedance correction of which I expect being useful in active speakers - is a Zobel compensating for the impedance-rise due to the driver's Le.

I'm trying to digest all the responses... I'm not sure I understood the meaning of that earlier post's reference to power factor. I guess changes in phase have something to do with the power output.
And when you said the rising impedance is the only thing that one needs to bother about, what about the resonance peak? Shouldn't that worry the designer, specially for tweeters, which usually do receive some signals at their Fs, even after the xo cuts it down? Won't a resonance peak compensation simply make the tweeter output flatter at the Fs?

[phase_accurate]

If you are using an amp that resembles a voltage source (as mentioned by richie00boy) you won't have any response aberrations due to the impedance-rise. Some amps however work better in purely resistive loads, that's where the power factor comes in.
I wouldn't compensate the fs but the rising impedance due to Le (Lvc) only.

If you use a tube-amp(s) things might look differently however.

Regards

Charles

[Svante]

If you use a normal solid state amplifier with no crossover between the amp and speaker, impedance compensation is not needed.

Impedance compensation can only be needed when the source (=amp+filter+cables) has an output impedance that is comparable with that of the driver.

Some say that a resistive load is better for the amplifier than a reactive load, but in normal cases, the load rather becomes more difficult for the amp if a compensating network is used even though it becomes resistive.

[soongsc]

With an XL-280 amp, there is an audible difference whether impedance compensation is used or not. I would not recommend the RC zobel unless you are trying to reduce high frequency content.

[forr]

A Zobel network (something around 100 nF + 4.7 Ohm) is usually necessary to ensure the stability of power devices used as followers at the output stage of most solid state amplifiers.

When directly driven from the ouput of such amps, there is no need to compensate for the rise of louspeaker impedance at high frequencies :
- it would be a waste of power
- a solid state amp is less linear when it delivers more current (when the load impedance decreases)

Less obvious : with an inductive load at high frequencies, as the current is lagging voltage, the phase margin of a standard Miller compensated amplifier is probably better than with a purely resistive load.

~~~~~~ Forr

§§§

[rcw]

In order to be "unconditionally stable" amplifiers already have an output "Thiele" network, or some varyation of it.
The idea of this is to isolate the amplifiers feedback loop from the reactive speaker load, in the case of no form of passive crossover this is more than adequate.