Correcting Baffle Step (CHP-70-P Gen2 with Audax TW010E1)

Baffle Step Compensation

I saved the baffle step compensation (BSC) for last because I wanted to see whether I can do without it. Presently, the CHP-70 with the TW010E1 are still in Time Aligned mode, that is the front panel is tilted backwards. With my setup unchanged, I will now attempt to correct the midrange emphasis caused by the baffle step.

The Black plot in Fig 1 is the frequency response of the CHP-70 with the front baffle slanted backwards. The Blue plot is with the BSC network installed.  

Fig 1 – CHP-70 with Baffle Step (Blue plot)

As can be seen, the baffle step reduction is about -3dB~-5dB. At 3kHz, it’s actually -6dB less.  

Fig 2 – CHP-70 with BSC and TW010E1 with L-pad

The Blue plot in Fig 2 is the CHP-70 with the BSC, followed by a low pass filter and a LCR. So far, I have yet to use a Zobel but it doesn’t look like I need one.The Red plot is the TW010E1 tweeter with just a capacitor. I had to install an L-pad to lower the sensitivity of the tweeter to match that of the CHP-70 with the BSC.

Fig 3 – Final Frequency Response with BSC

Fig 3 is the summed response of the CHP-70 with the BSC and the TW010E1 with L-pad. The tweeter is wired in reversed phase. Compare this to Fig 4, which is without the BSC, the reduction from 1kHz~4kHz is apparent. 

Fig 4 – Final Frequency Response without BSC

Fig 5 – Null Response with CHP-70 BSC and TW010E1 L-pad

To check on how well the two drivers are crossing, I rewired the tweeter back to normal phase. The resulting notch is quite deep, indicating the two drivers are crossing properly. It’s not as pretty as I would like it to be but it’ll do for the time being. I am more interested in how it sounds than perfect measurements. 

Fig 6 – Full bandwidth response

Once the CHP-70 and the TW010E1 are sorted out, I added in the bass. For this, I used a 24dB/oct electronic crossover to merge my bandpass subwoofer, the Toucan-SF, to the CHP-70. After some adjustments, I found 250Hz works best. Note that at this frequency, the deep notch caused by the floor bounce is absent (Fig 6).

Fig 7 – Subwoofer Null

Fig 7 is the response when I flip the subwoofer wires back to absolute phase. A notch appeared at 200Hz. This shows the Toucan-SF sub and the CHP-70 are crossing properly.

Fig 8 – Time aligned Step Response with BSC

For the step response, I disconnected my bandpass subwoofer. What I’m aiming for is the time alignment between the CHP-70 and the TW010E1. With the BSC in place, the step response (Fig 8) is virtually similar to without. However, there is a slight difference. In this step response, the tip is at 17 usec whereas the previous one is 23 usec. So, with the BSC, the time alignment is marginally better.

Fig 9 – Spectrogram of CHP-70 with TW010E1

The spectrogram in Fig 9 reveals the CHP-70 ringing from 1kHz~3kHz. They are not severe, lasting up to 6 msec before dying off. I doubt they are audible. In the treble, whatever ringing from the tweeter is inconsequential because they do not last more than 1 msec. We can see the width difference from 4kHz upwards compared to what’s below 4kHz.

Does it sound better

Most certainly, the BSC made a huge impact. Without it, I found the midrange emphasis distracting at times. At low volume, it’s fine but the minute I pump it up, the speaker sounds shouty. With the BSC in place, the music is much more refined. One must be careful in tuning the BSC though. If it’s overdone, the speaker will sound dull.

I am currently listening to this system with my Toucan-SF bandpass sub doing the bass. If I want to use a passive crossover for the subwoofer, I will have to pad down the CHP-70 and TW010E1 satellite by 1.5dB. That is something I can work on later. What is important is I managed to develop a speaker I’m proud of with drivers that are not too costly. I am so pleased with this system that I will build proper boxes for them this spring. 

Part 1 – CHP-70 with Audax TW010E1
Part 2 – Time Alignment
Part 3 – Correcting Baffle Step

Enclosure used is 1.7 liters Sealed Box stuffed with Polyfill.
Unless otherwise stated, all measurements were made with the mic at 36 ins, tweeter axis. Impulse Window=5ms. No smoothing applied.