Questions on frequency response of TL speaker

[Kirkg]

Hi all-

I have designed and built a pair of quarter-wave 10:1 tapered transmission line speakers with 10” woofers and dome tweeters. Here's a link to some questions I had on stuffing the speakers: Stuffing a transmission line speaker- unexpected results. This link also gives pictures and more details on the speakers.

I measured the output of one of the TL speakers driving only the woofer in an anechoic chamber. I used a signal generator and NAD amp to drive the speaker. I used a Dayton Audio iMM-6 microphone and the Audio Tool Android app. to measure the output.

The first attached picture shows the background noise in the chamber. The 2nd attached picture shows the output of the TL speaker at 5' from the front of the woofer on axis. For this picture- ignore the green bars- just look at the blue line segments. I plan on initially crossing the woofer over at ~2kHz when I use a dome tweeter in a 2-way. Later, when I get my ESL's done I will replace the tweeter and crossover at a lower frequency. Note that there are dips in the frequency response at ~600Hz and 2kHz.

The 3rd attached picture show output of the TL speaker from the rear opening which is on the backside of the speaker. Again, just look at the blue line segments. I expected to see the output dropping off faster and at a lower frequency.

Any comments would be appreciated. Any suggestions on how I can smooth out the frequency response?

 

[TMM]

The 3rd attached picture show output of the TL speaker from the rear opening which is on the backside of the speaker. Again, just look at the blue line segments. I expected to see the output dropping off faster and at a lower frequency.

Probably because you are also measuring the sound from the front of the speaker which has diffracted around the box.

 

[weltersys]

1)I used a signal generator and NAD amp to drive the speaker.
2)I used a Dayton Audio iMM-6 microphone and the Audio Tool Android app. to measure the output.
3)The 2nd attached picture shows the output of the TL speaker at 5' from the front of the woofer on axis. The 3rd attached picture show output of the TL speaker from the rear opening which is on the backside of the speaker. I expected to see the output dropping off faster and at a lower frequency.
4) Any suggestions on how I can smooth out the frequency response?

Kirkg,

1) What type of signal?
2) This looks like a pink noise RTA (Real Time Analyzer) screen. Have you measured any "flat" speakers to determine if your mic/app. is flat? Have you measured the signal generator itself with the app?
3) The HF drop off to the rear looks reasonable, but the roll-off with frequency does not look quite consistent with the front measurement, which suggests to me that there may be a fair amount of level "bouncing" going on, with your screen captures perhaps not having hit the average level. If that is true, the dips at 630 Hz and 2kHz may not be quite as deep as they appear to be. Using 1/6th octave resolution would be advisable to better determine the frequency and depth of the dips.
Better yet would be to use REW (or any dual FFT type measurement system) with no smoothing.

If the measurement is correct, the frequency response could be smoothed using a 1/3 octave equalizer or parametric EQ.

Art

 

[Kirkg]

Hi- thanks for the comments. Can sound from the front of the speaker wrap around to the rear? I know the wavelength of sound in air is quite long- approx. 22 feet at 50Hz. In the anechoic chamber there should be no bouncing of sound off the walls.

Regarding the measurement I had the signal generator slowly sweep between 10Hz and 5kHz. I checked the output of the signal generator and amp and the voltage was constant across the frequencies. I then had the spectrum analyzer app record the peak amplitude at each frequency. Unfortunately, I don't have a speaker with a calibrated output. I used the 1/3rd octave resolution as it is the only resolution with a calibration file for the mic I have.

 

[pkitt]

Absolutely! Sound wraps around all speakers, starting at a frequency determined by the baffle's width. It's called the Baffle Step Frequency and corrected in a crossover design by BSC (Baffle Step Compensation). It's not really a sudden "step" but a gradual increase in the amount of wrap-around loss ending at a maximum of 6 dB SPL at a very low frequency. For instance, a 12-inch wide baffle has a Baffle Step Frequency of 380 Hz. To properly measure the system output from a TL, and a ported box, the individual outputs of driver and terminus are measured separately, then summed together after normalizing their magnitudes due to their different radiating areas and correcting for their phase differences.
Paul

Hi- thanks for the comments. Can sound from the front of the speaker wrap around to the rear? I know the wavelength of sound in air is quite long- approx. 22 feet at 50Hz. In the anechoic chamber there should be no bouncing of sound off the walls.

Regarding the measurement I had the signal generator slowly sweep between 10Hz and 5kHz. I checked the output of the signal generator and amp and the voltage was constant across the frequencies. I then had the spectrum analyzer app record the peak amplitude at each frequency. Unfortunately, I don't have a speaker with a calibrated output. I used the 1/3rd octave resolution as it is the only resolution with a calibration file for the mic I have.