Difficulty to measure speaker impedance with REW

[HERVIEU]

After measuring successfully the speaker impedance with REW three months ago, I tried again recently while respecting the procedure, but the result is totally abnormal.I notice that the measure of a non inductive resistor gives a straight line at the good value and that the speakers measured are alone without any filter.
I put here attached the file that shows the phenomena.I
Thank you for your response.

 

[sm52]

Before measuring the impedance, it is better to do a calibration every time. Especially after a few months.

 

[HERVIEU]

Yes, calibration has been performed according to REW recomanded procedure. Is there any bug with Windows 10 ?

 

[sm52]

Something is interfering with the measurement. You have selected the UR22C mix R output. Maybe the output is not a clean signal, maybe the sound card mixer is adding some kind of signal.

 

[John Mulcahy]

Make sure the mix knob is turned all the way to DAW.

 

[HERVIEU]

Thank you for your post.

I just arrived to obtain realist measures by selecting the noise filter at "low level". It is a fact that without such noise filter, the measure is definitely abnormal ...

 

[John Mulcahy]

The noise filter should only remove noise, though it might also reduce the effect of feedback from monitoring being active. Make sure there is no monitoring, on the UR22C turning the mix knob all the way to DAW should achieve that.

 

[HERVIEU]

Thanks for your attention. My interrogation lies in the fact that all my measures have been made with the knob in DAW position and as soon as the noise filter has been selected, the measure came correct. It should be interested to understand how such filter operates. Perhaps my UR22C knob is not perfect on the fact that even on DAW position, there is still input signal mixed ??? I observed the signal on the two input of the UR22C with the scope function at various generator frequencies but not any feddback effect was visible, but off-course, such observation cannot be accurate ....

 

[John Mulcahy]

The noise filter is a tracking filter that follows the sweep and so excludes content that is not at the current sweep frequency. That would attenuate the effects of a monitoring path that had a significant delay since the delayed signal would be at a lower frequency than the current sweep frequency.

 

[HERVIEU]

Thanks for the explanation about the noise filter. I think that question is solved.