|
Magpie-DX
Before I work on any woofers, I always listen to her without any crossover. That will give me an idea of what I’m up against. What I heard was horrific. The SB17NRX cone breakup (Fig 1) was like screeching glass. Very bright and brittle. I would understand if it’s a metal cone but a paper cone? I’m looking at +15 dB at 6kHz. It is not an issue if I’m using her strictly for bass but in a 2-way, the cone breakup can be problematic.
After numerous attempts, I finally found the sound I was looking for. The Blue plot in Fig 2 is with a rather elaborate low pass network. There’s no way to save cost on the crossover with this SB17NRX. I had to EQ out the cone break-up. Now, the tip is about -10dB below 85dB. Hopefully, she will not cause any cancellations when I integrate in the tweeter.
The Red plot in Fig 3 is the Peerless DX25TG59-04 with her high pass network. She is crossing at 2kHz with the SB17NRX. That’s good because 2kHz is my targeted crossover frequency.
Fig 4 shows the summation of the two drivers. There are no cancellations in the passband, indicating nothing is lost during the summing. What really caught my attention is the flatness from 3kHz~6kHz. That band in the tweeter is actually not that smooth. What happened is the cone break-up of the SB17NRX pushed up the valleys.
The frequency response of the Magpie-DX is incredibly flat from 1kHz onwards. Measurements below 500Hz includes room reflections. Ignore the deep notch at 150Hz. It’s a floor bounce in my setup.
The null response in Fig 6 shows the two drivers phase alignment leaves plenty to be desired. I’m not too concerned about this since I didn’t pick up any phasing issues during auditioning. I can tidy it up later if I want to. Right now, my priority is to lock in the sound of the Magie-DX. I don’t want to be distracted with technicalities.
The Blue plot in Fig 7 is the Nearfield response of the SB17NRX. In this measurement, the microphone is about 1/4″ away from the woofer cone. Because it’s so close, room reflections are not recorded.
Fig 8 is the final frequency response of the Magpie-DX. This is what the response looks like without room reflections.
The jagged tip at the bottom (Fig 7) reveals the phase misalignment of the SB17NRX and the DX25TG59. The SB17NRX hits the apex with a healthy sharp tip at 350 microsec. This is quite fast for a 6-1/2″ woofer. Normally, the tip is at 400 microsec.
The Waterfall in Fig 10 recorded some artifacts in the treble. They are insignificant. The Spectrogram in Fig 11 verifies the treble is very clean from 1.5kHz upwards. However, there is a bleed at about 1.1kHz. Fortunately, it is dissipated by 7 msec. I did not pick up any smearing during playback.
The Magpie-DX registered an Excess Group Delay of -2.59 msec at 48Hz. This is where a sealed box beats a bass reflex.
The Magpie-DX distortion is impressively low. The 2nd (Red plot) and 3rd (Violet plot) harmonics are -60dB below the fundamental. The Blue plot is the sum of all the harmonics from 2nd~5th. Sealed Box Alignment
Fig 14 is the box modeling of the SB17NRX in a sealed box. F3 is at 70Hz. This may give the impression that there’s very little low bass. In reality, that’s not true. If we look at the frequency response of the Magpie-DX in Fig 5, you can see the bass extends to 50Hz. This is possible because a sealed box has a gentler roll-off. Sound of Magpie-DXThe Magpie-DX is for midrange lovers. Vocal clarity is superb. Voices pop out of the mix. There’s no shrillness in female voices. Sibilance is well controlled. Doesn’t spit at you. No harshness in the treble. Bass is tight and dynamic. Guitars sound particularly beautiful. The strings in Tino Izzo One-Blue Desires is crystal clear. Relax to the music in Paradise Cafe by Antar. And if you are a fan of a singer/guitarist, there’s no better than John Denver. In Perhaps Love, his singing did not recede into the background. He was situated in the front of stage. And his guitar can be clearly heard, even the bass. About the SB17NRX2C35-4Working with this SB17NRX has been a challenge. I initially tried her in an 18 liters bass reflex but the outcome was terribly disappointing to say the least. I even resorted to aperiodic but the bass was still bloated and lacked definition. I think it would have worked had the T/S been the same as the manufacturer’s. What I extracted with DATS was quite way off. As a last resort, I abandoned the bass reflex and loaded her in a sealed box. Only then did the bass sounded right. After that, it was a question of finding the sweet spot, which is the vocals. I am glad I persisted with the SB17NRX. She was about to be consigned to the dud bin but actually ended up with a speaker I am very proud of. I plan on upgrading the tweeter at a later date. Perhaps with one of the SB Acoustics models. Not that the DX25TG59-04 is bad. In fact, I think very highly of the DX25. It’s just that in speaker design, any marginal improvements makes a huge difference to those that are seriously into sound. Converting the Magpie-DX to a Bass Reflex
I managed to convert my 18 liters sealed box to a bass reflex. The result is a deeper bass and more punch. But it comes at a cost. The bass is slightly bloated and not as tight as the sealed box. For this conversion, I used a 1.5″ diameter PVC tube cut to a length of 4″. The alignment is modeled in Fig 15. There is no loss in the midrange and treble. Only the character of the bass changes. Unless otherwise stated, all measurements were made in Full Space (4 pi) with the mic at 36 ins, tweeter axis. Impulse Window=5ms. No smoothing applied. |
February 11, 2021Projects
Fig 1 – Black plot=SB17NRX2C35-4 RAW Response • Baffle Width=9-1/2″
Fig 2 – Blue plot=SB17NRX2C35-4 with Low Pass network
Fig 3 – Red plot=Peerless DX25TG59-04 High Pass
Fig 4 – Black plot=Magpie-DX Passband
Fig 5 – Magpie-DX Frequency Response
Fig 6 – Magpie-DX Null
Fig 7 – Magpie-DX Nearfield below 500Hz
Fig 8 – Magpie-DX Nearfield Frequency Response
Fig 9 – Step Response
Fig 10 – Waterfall
Fig 11 – Spectrogram
Fig 12- Excess Group Delay
Fig 13 – Magpie-DX Distortion
Fig 14 – SB17NRX box modelling
Fig 15 – Bass Reflex Modeling