Speaker Design Measurement - Frequency Discrepancy with Analog Microphone

[Ramius2]

Hi,

I have been using REW for room measurement for some time but have just joined here and have just started my first real speaker design effort (ported 2 way). I am designing in VituixCad and for the last few days I have been running impedance and response measurements using REW – closely following the Vituix guide. As recommended I am using an external souncard with loopback and dual connection and (Sonarworks) calibrated analog microphone. Speaker is 1.5m off floor in a large cleared room (3m ceilings). I have Nearfield (Cone and Port) and Farfield measurements for a number of variations (port length and fill).

I have a calculated Fb of around 40Hz which aligns nicely with the dip between peaks in the measured Impedance curve. The port maximum should also align with Fb however it is at 52Hz. This causes a significant bump in the bass response when combined with the cone measurement. This ~12Hz discrepancy between Fb and port maximum is consistent in all cases with the Sonarworks mic.

I then checked measurements with my old UMIK-1 microphone. The port frequency is now as expected. Moreover the shape of the curve is now very close to modelling while the measurements with the Sonarworks microphone show exaggerated peaks and unexpected inflections.

I have no idea what might be causing this – REW settings, some physical interaction or bad hardware. Any input appreciated – it is difficult to have confidence in the measurements or to know how to proceed while this issue is unresolved.

Chart of Nearfield Port Measurements. Green trace is USB Mic - others are variations using analog Sonarworks mic.

 

[John Mulcahy]

Are you using the WASAPI Exclusive drivers with the external interface, i.e. device names starting EXCL?

 

[Ramius2]

Hi John
Thanks for the reply and great software BTW.
No I don't think so. The driver installed under preferences is FlexASIO.

 

[Ramius2]

OK - quick read on Google - I gather WASAPI are windows drivers. Answer is still no.
Am continuing to read myself in.

 

[John Mulcahy]

I wondered if there was some sample rate issue, but FlexASIO would handle the sample rate (as would REW's Java WASAPI Exclusive drivers). The mains hum glitches are in the same place for all the traces though, so that suggests there's no sample rate problem. If you post the mdat file for the measurements that makes it easier to check for anything else.

 

[Ramius2]

Hoping this works....

 

[Ramius2]

FYI - I cheaped out on the soundcard (TEYUN Q-12), in case that could be part of the problem...

 

[John Mulcahy]

The interface looks to be picking up a strong 50 Hz hum component, which isn't helping, but that's not enough to account for the peak at 52 Hz. Nothing leaping out at me from the measurements, sorry.

 

[Ramius2]

OK thanks - I am going to play with some more measurements at different distances, then try a different computer and if nothing jumps out try a different interface. I i find something i will post it.
Cheers