I've wondered for a long time about that old bone of contention "does the speaker load ever appear less than the DC resistance."
IIRC it was Matti Otala who first said an 8 ohm speaker could transiently appear to be 1 ohm
AES Convention Papers Forum >> Peak Current Requirement of Commercial Loudspeaker Systems
which led to Harman-Kardon's "high current" receivers and amplifiers.
But then a later paper-which I very unfortunately do not recall the authors of-said "well, Otala's data was only from simulations, we find no actual evidence of the loudspeakers transient impedance ever being lower than the D.C. resistance"
I'm wondering if anyone has ever tried measuring this* or knows of any other evidence about this?
*I had a scheme where I wanted to rig up a data-output voltmeter across the speaker, another across a 0.1 or so ohm resistor in series to the speaker, then play actual music. Each second or whatever, record both values into a computer, which would then divide the two to find an impedance magnitude. But then I changed jobs and that never happened.
Another approach might be to calibrate a storage oscilloscope running an X-Y I-V plot-the slope of the line for a sine wave would be the impedance, so presumably you could look at that and see if the slope ever goes past the DCR value (which you could calibrate with an actual resistor).
Both of these idea may not be quite right-I haven't thought about this for a long time.
IIRC it was Matti Otala who first said an 8 ohm speaker could transiently appear to be 1 ohm
AES Convention Papers Forum >> Peak Current Requirement of Commercial Loudspeaker Systems
which led to Harman-Kardon's "high current" receivers and amplifiers.
But then a later paper-which I very unfortunately do not recall the authors of-said "well, Otala's data was only from simulations, we find no actual evidence of the loudspeakers transient impedance ever being lower than the D.C. resistance"
I'm wondering if anyone has ever tried measuring this* or knows of any other evidence about this?
*I had a scheme where I wanted to rig up a data-output voltmeter across the speaker, another across a 0.1 or so ohm resistor in series to the speaker, then play actual music. Each second or whatever, record both values into a computer, which would then divide the two to find an impedance magnitude. But then I changed jobs and that never happened.
Another approach might be to calibrate a storage oscilloscope running an X-Y I-V plot-the slope of the line for a sine wave would be the impedance, so presumably you could look at that and see if the slope ever goes past the DCR value (which you could calibrate with an actual resistor).
Both of these idea may not be quite right-I haven't thought about this for a long time.
An incorrectly configured crossover network which can resolve to a series L,C circuit with minimal series resistance in some cases ,can be nearly a short circuit at certain frequencies. Impedance/phase graphs a are important when designing speaker systems. good design practice is to make combinations as resistive as possible.
What about EPDR (equivalent peak dissipation resistance)? Is it relevant to this discussion?
Apparently large phase angles mean that the EPDR can fall below the DC resistance at more than one frequency.
Apparently large phase angles mean that the EPDR can fall below the DC resistance at more than one frequency.
I meant to say 'loudspeaker system', but that got lost when revising my post!
EPDR is something that is always measured in Hi-Fi News loudspeaker reviews and can result in some uncomfortably low impedance figures at significant frequencies.
EPDR is something that is always measured in Hi-Fi News loudspeaker reviews and can result in some uncomfortably low impedance figures at significant frequencies.
Hah! Good one! I guess I should have said I meant dynamic drivers. But all comers welcome!
Well, uh, either I guess. I was thinking more of single drivers, but Ottala's paper was for systems. I *think* EPDR is what we're talking about. Basically any instant that the voltage divided by current is less than DCR if I'm figuring correctly.
DC resistance refers to DC.
Impedance refers to AC.
Yes. Impedance is *always* lower than DC impedance for dynamic drivers.
Just take a look at any arbitrary speaker impedance plot from stereophile or whatever...
Impedance refers to AC.
Yes. Impedance is *always* lower than DC impedance for dynamic drivers.
Just take a look at any arbitrary speaker impedance plot from stereophile or whatever...
For a solo voice coil driver the impedance can never be lower than the DC resistance. But in the past many bad crossover designs did have very low impedance notches that were well below the drivers DC resistance. And the capacitance of an electrostatic driver drops its impedance with frequency.
I've done a little research into EPDR and will attempt to explain it here. If I'm wrong, perhaps someone can explain it better to me! 🙂
EPDR is the resistive load that would give rise to the same amplifier peak output stage power dissipation as the speaker itself.
The EPDR minima are significantly lower than indicated by a speaker's minimum modulus of impedance, and they occur at different frequencies.
In dynamic speakers the minimum EPDR generally occurs quite close to peaks in the modulus of impedance curve that relate to driver fundamental resonances.
The modulus of impedance can be well above minimum, but if the phase angle is large, the EPDR will be low.
EPDR is the resistive load that would give rise to the same amplifier peak output stage power dissipation as the speaker itself.
The EPDR minima are significantly lower than indicated by a speaker's minimum modulus of impedance, and they occur at different frequencies.
In dynamic speakers the minimum EPDR generally occurs quite close to peaks in the modulus of impedance curve that relate to driver fundamental resonances.
The modulus of impedance can be well above minimum, but if the phase angle is large, the EPDR will be low.
I vaguely remember an AES article by Peter Baxandall about how to measure the performance of the protection circuits of audio amplifiers. I think it also had some information about what combinations of voltage and current are likely to occur while playing music.
This must be the article.
https://linearaudio.nl/sites/linearaudio.net/files/bax i-v prot.pdf
Interesting, I'll have a read later. 😎
this is a very tricky question .... try to figure out what will happen if 2 drivers are mounted in a single box ... if one is working as a speaker and the other one as a microphone .... and it may happen ... if you have a open back midrange in the same box with the woofer ... and connected with a cross-over .... a nightmare .....
I think you may be referring to back emfs - induced voltages pushing backwards, from the speaker to the amplifier.
Your 'microphone' and 'open back midrange' scenarios will produce back emfs which form part of the load on the amplifier and contribute to the modulus of impedance and phase angle of the speaker system.
I don't follow what EPDR is and what it has to do with loudspeaker impedance.
Which is precisely why one should never do that.
EPDR has to do with the amplifier power dissipation into a real life speaker load at different frequencies.
The EPDR equals the value of the fixed resistor that would result in the same amplifier power dissipation as the speaker load at any particular frequency.
The EPDR changes because of the changing phase angle between the voltage and the current in the reactive loudspeaker load.
Earl may understand this, but here is some useful background reading for anyone still uncertain:
http://sound.whsites.net/patd.htm
Last edited:
There is so much confusion here around real part of impedance, magnitude of impedance, resistance, etc. I don't know how someone coming here to learn something can make sense of it all.