I measured my speakers with... themselves

[daftcombo]

Hi,

I wanted to measure one of my Focal Aria 906 yesterday with a Behringer ECM8000 calibrated microphone and a Focusrite Scarlett 2i4 2nd Gen as DAC/ADC, using REW.

The mic was positionned 2 meters away from the speakers, on the tweeter axis.

I obtained a super nice curve, looking a lot like this one:

(Sorry, I don't have the actual REW curve where I am at the moment.)

What surprised me was that there were no peaks or dips like I usually obtain when I measure a speaker in my nasty room.

There was a big problem though: distortion was about 200%!!

Searching for an explanation, it took me a few seconds to realize that I had forgotten to switch the phantom power of the Scarlett 2i4 on. So basically the mic didn't record anything. The only explanation I can think of is that the speakers themselves acted as a microphone and somehow passed the signal back to the ADC of the Scarlett.

My question is: wouldn't forgetting to plug the mic be a great way to measure speakers without room interferences?? Without needing semi-anechoic conditions?

Cheers.

 

[exaudio]

@daftcombo, even without the phantom power switched on--as long as your mic was plugged into your mixer/ADC it could still be recording. If I understand correctly, the transducer in the mic will still generates tiny voltages, but without the phantom power that tiny signal doesn't get any amplification. It's probably still enough to get picked up by your Focusrite though. If your mic wasn't even plugged in then maybe REW was actually recording from another source like a built-in mic in your laptop/pc. Sadly I don't think there's a free lunch when it comes to recording without room interference.

 

[daftcombo]

I came back home so here are pics.

With phantom power off:

And with phantom power on:

Both 1/6 smoothing!!!

 

[RayDunzl]

Thought:

The condenser mic may not have finished losing the charge it received from the prior use.

Test:

Apply phantom power and take a reading.

Turn off phantom power.

At intervals of your choice, repeat the measurement.

Apply rigorous SWH Test Protocol* and report back.

I may try it myself.

*See What Happens

 

[LTig]

I came back home so here are pics.

With phantom power off:

View attachment 48136

And with phantom power on:

View attachment 48138

Both 1/6 smoothing!!!

In this case I would always listen with phantom power off, SQ must be much much better...

 

[daftcombo]

What bothers me the most is the measurement with the phantom power off being so close to the one made by Soundstage (see my first post)! Check the dip around 400 Hz and the small peak around 1.5 kHz.

I just repeated the "phantom off" version with the mic randomly laying on the ground, and got the exact same curve. So it definitely isn't recording.

 

[RayDunzl]

I may try it myself.

My mic died right away. ECM8000 into a Focusrite.

Phantom Power On (TV in room, Adam Schiff still trying to convince me)

Power off

 

[scott wurcer]

Thought:

The condenser mic may not have finished losing the charge it received from the prior use.
*See What Happens

Electrets don't work that way, the charge is built in (unless you mean the charge on the coupling caps). Turning off phantom power does not disconnect the input in general there is the possibility of feedthrough. I did some experiments and these capsules work at 10uA with some output.

 

[RayDunzl]

Electrets don't work that way, the charge is built in (unless you mean the charge on the coupling caps).

I see.

I'm past my "open everything up" years...

This guy isn't:

The other business end of an ECM8000

Looks like it needs a little power to do anything, though, even if not applied to the capsule...

 

[scott wurcer]

I see.

I'm past my "open everything up" years...

This guy isn't:

The other business end of an ECM8000

View attachment 48148

View attachment 48149

Looks like it needs a little power to do anything, though, even if not applied to the capsule...

Pretty standard pseudo-balanced circuit, C5 is not big enough to supply current for long after the phantom power is turned off so there should be no chance of any signal feedthrough.