I found this very informative article on the Quad ESL impedance:
http://www.audiomisc.co.uk/57and303/interact.html
I decided to verify the impedance plot on my speaker, so I used an audio signal generator, a series resistance and an audio voltmeter. I set up the level so that 40mV on the voltmeter corresponded to 10 ohms. Then I entered the measured values in an Excel table and I got this impedance diagram:
I wondered how accurate is the electrical model against the measured values, so I created an LTSpice model. See .asc file attached. I had to modify the electrical circuit by adding a 1.5R resistor series to the 2uF capacitor representing the treble section, so the impedance is about 2R at the treble dip.
This is the AC analysis plot:
I did not measure the phase. From the analysis it seems mostly inductive below 100Hz, then capacitive up to 20 kHz or so.
Due credits to Sheldon (quadesl.org) for the reference to the ESL Dummy Load originating from Christian Steingruber:
https://www.quadesl.com/origRefs/ampload.jpg
And thanks to Jim Lesurf for the original model that I modified a little and used as the basis of the LTSpice model.
http://www.audiomisc.co.uk/57and303/interact.html
I decided to verify the impedance plot on my speaker, so I used an audio signal generator, a series resistance and an audio voltmeter. I set up the level so that 40mV on the voltmeter corresponded to 10 ohms. Then I entered the measured values in an Excel table and I got this impedance diagram:
I wondered how accurate is the electrical model against the measured values, so I created an LTSpice model. See .asc file attached. I had to modify the electrical circuit by adding a 1.5R resistor series to the 2uF capacitor representing the treble section, so the impedance is about 2R at the treble dip.
This is the AC analysis plot:
I did not measure the phase. From the analysis it seems mostly inductive below 100Hz, then capacitive up to 20 kHz or so.
Due credits to Sheldon (quadesl.org) for the reference to the ESL Dummy Load originating from Christian Steingruber:
https://www.quadesl.com/origRefs/ampload.jpg
And thanks to Jim Lesurf for the original model that I modified a little and used as the basis of the LTSpice model.
I did what you did to get the impedance plot that has been on my site forever. I remember reading or maybe noticing myself (it's been a LONG time) that the impedance hump around 100hz is level dependent. I never explored how much it varies though.
Sheldon
Sheldon
Oops, I should have referred to it, apologies:
http://www.quadesl.com/graphics/quadGraphics/quad_impedance_graph.jpg
oh no, I wasn't looking for a reference, I just wanted to point out that the big impedance hump down low might be different at different test levels
Sheldon
There are two ways to measure impedance:
- Add a series resistor of at least 10x the maximum expected magnitude, and measure the voltage on the DUT. The measured signal will be proportional to the absolute value of impedance. Phase can be measured with two-channel instrument (soundcard + software). This setup does not make possible a measurement at high signal level.
- Add a series resistor of at least 0.1x the minimum expected magnitude, and measure the voltage on it. The measured signal will be proportional to the current. The impedance will be its reciprocal. High signal level measurement is possible.