Passive driver fb discussion

[audio_04]

Hello everyone,

I'm currently doing acoustic measurements for my engineering thesis in an anechoic chamber using a Brüel & Kjær Pulse system, and I've run into some strange low-frequency behavior that has me stuck.

The design is a compact active 2-way monitor with a Dayton DSA135-8 (5") woofer coupled to a passive radiator (PR). I'm taking near-field measurements (mic < 1 cm from the dust cap) using a Log Sine Sweep to avoid the 50 Hz mains hum of the lab.

The issue: I cannot observe the classic attenuation dip (notch) in the woofer's response at the tuning frequency (fb).

Theoretically, at fb, the woofer's excursion should be minimal as the PR takes over the acoustic load (just like in a bass-reflex ported system). However, my woofer's LF response is almost flat and rolls off smoothly without any pronounced dip.

I've attached two images for comparison:

this is from a practice project we did in class with a bass-reflex system, meassuring nearfield responses near the woofer and the vent.

this is the response I get near the woofer, where there's no attenuation.

Checks I've already done:

• I've thoroughly checked for air leaks and sealed everything (connectors, rear panel, etc.) to ensure a proper acoustic suspension.
• Hypothesis: The rear center mass of my PR is a fixed metallic piece (looks like a structural rivet/post). Since the enclosure is very compact, the woofer's magnet is physically very close to the back of the PR. Could magnetic attraction be stiffening or "locking" the PR's suspension, preventing it from resonating properly?

Is it normal for this notch to be harder to detect in PR systems compared to vented boxes? Are there any other physical factors that could be decoupling the system so the woofer doesn't unload at fb?

Any insights or measurement suggestions would be greatly appreciated.

 

[NTK]

Amir's measurement of the AsciLab F6B, which has a passive radiator, showed the dip.

 

[Jon AA]

Too much weight on the radiator or area mismatch? Have you done an impedance sweep?

 

[DavidR]

How have you determined the tuning Fb?

 

[audio_04]

How have you determined the tuning Fb?

In WinISD. Fb should be around 60 Hz.

I think it might be that there's magnetic attraction to the PR. The rear center mass of my PR is a fixed metallic piece, easy for adding mass. Since the enclosure is very compact, the woofer's magnet is physically very close to the back of the PR.

 

[audio_04]

Too much weight on the radiator or area mismatch? Have you done an impedance sweep?

How can I perform an impedance sweep? The PR doesn’t have terminals to connect it to the signal generator or to the FFT analyzer.

 

[Redacted]

Hang the speaker by a cord from the ceiling - the table has an effect.
Do a quick test at different heights just to check you are not adding to room resonances (yes, I know the room is supposed to be anechoic!)

 

[audio_04]

Hang the speaker by a cord from the ceiling - the table has an effect.
Do a quick test at different heights just to check you are not adding to room resonances (yes, I know the room is supposed to be anechoic!)

The measurements where taken from a distance of 1cm from the woofer. There should be no room resonances in the measure.

 

[Redacted]

The measurements where taken from a distance of 1cm from the woofer. There should be no room resonances in the measure.

Mmm. 'Should be'. And yet the post is about an anomaly. And mic <1cm from the dust cap? The table is still not ideal.
If the driver under test is passive, the problem may be an unexpected resonance in the box.

 

[DavidR]

A close-mic measurement will swamp any room modes and likely any reflections from almost anything. An impedance sweep suggested is not made on the PR, it's made on the driver and will show roughly what the Fb is. If the PR isn't moving, as appears to be the case, you'll see the tuning of just the woofer.

Describe the metal you placed on the PR. How heavy is it? Either it is as you suggest, the magnet is preventing much movement due to the metal on the PR or the metal is so heavy on its own coupled with the mass of the PR that the true Fb is far too low and will show little response in a measurement.

To verify just cover/seal the PR from the outside to make it essentially a closed box.

 

[Redacted]

"A close-mic measurement will swamp any room modes"

... the table top is a few cm from the mic. Albeit not in the pic shown. Can we get back to common sense?

 

[DavidR]

"A close-mic measurement will swamp any room modes"

... the table top is a few cm from the mic. Albeit not in the pic shown. Can we get back to common sense?

It's based on the physics, not common sense. It will still swamp reflections such as that almost entirely and will be immaterial anyway. Were it not so you would see a significant dip in response related to the wavelength based on that. Specifically it will be centered on the frequency with a 1/4 wavelength equal to the distance from driver center to reflection surface. That will have no affect on the PR tuning area.

Guessing from the photo and the fact that the driver is 5.3" in frame diameter according to Parts-Express web site and guessing that the rim is about 1.5" above the baffle bottom edge we get 4.15" from driver center to edge for 1/4 wavelength. f = c/lamba where c is speed of sound and lamba is wavelength. For 1/4 wavelength f = 13550 /(4 x 4.15") = 816Hz.

Two notes. First, the impact is distributed since the driver is not a point source nor is the reflection area of the box below, so the response will be spread out, but centered on 815Hz.

Second, a close-mic measured as described by the OP will only be valid up to about 400Hz at best due to phase issues in the driver response at the mic location so the reflected wave is immaterial.

The sole problem is related to the PR.

 

[DavidR]

Is it normal for this notch to be harder to detect in PR systems compared to vented boxes? Are there any other physical factors that could be decoupling the system so the woofer doesn't unload at fb?

A PR response tends to be a bit more broad and somewhat lower due to damping in the PR. They also exhibit a secondary resonance based on the surround as well. Ports have little damping at Fb, less loss. The tuning is more distinct as your practice response shows. The lower a PR tuning, the more broad the response with smaller impact on SPL. Tune it too low and you get very little output from the PR. Simliar with a port I think, but I've never actually built a ported system, only re-used boxes. I have made PR systems.

Also, a PR system does not "unload" as a ported system does. A PR system will eventually asymptote to rougly the equivalent of the closed box without the PR. At low frequencies the PR cannot produce enough volume velocity so it's more like a lossy closed box in a way.

 

[Salt]

- Frequency Response of the passive radiator?
- Impedance 20 Hz to 20 kHz?

 

[DavidR]

- Frequency Response of the passive radiator?
- Impedance 20 Hz to 20 kHz?

I don't understand your questions. A PR has an Fp and other T/S parameters, but no inherent frequency response unless you're asking for the direct PR measurement in the system. That's a good point. The PR output and its impact on system impedance will be minimal by about 1kHz.

 

[Salt]

I don't understand your questions. A PR has an Fp and other T/S parameters, but no inherent frequency response unless you're asking for the direct PR measurement in the system. That's a good point. The PR output and its impact on system impedance will be minimal by about 1kHz.

Measure the FR close to the PR and the impedance of the whole speaker was the message.

 

[Redacted]

I think this needs more real physics, more postulation and test, and less making it up to suit.....

 

[Salt]

As example (how it can go wrong ):

This is a 10" woofer in 36 L and BR tuned to 43 Hz, transducer (red) and BR (green) measured seperately:

What we see here predominantly is the room modes of the basement; besides the significant dip of the bass proves that the BR resonance frequency fits.

But what happened here? This measurement was taken via DATS, that was in use for the impedance plot and TSP ). This system is driven via USB's 5 V via Laptop ....

After using an Hypex PA for measurement, the result looks as to be expected (all measurements not aligned, no gating, simply shot):

So, everything matters, the equipment as well as the environmental circumstances (like tables in an anechoic room).

 

[DavidR]

Measuring equipment that is not adequate for the task is, I would submit, a separate issue.

I have measured several PR systems of mine, one being a large woofer with a PR not far from the floor. My basement also has significant room modes. I have never had any room mode nor floor reflections present in my measurements in the PR range. Internal resonances, yes. The larger one had 10" driver and 10" PR (it moves a lot), tuned to 19Hz. I also measured the primary woofer close-mic. It's baffle mounted 19" above the floor. There was no room mode influence in that, either. That's the beauty of the Keele method.

To be clear, were you measurements all close-mic, as in 10mm or closer?

I still do not believe that the table in use by the OP has any signficant influence anywere near the region of the box tuning, the OP's concern with lack of PR influence. It certainly is not going to somehow cancel the PR influence as appears to be the case.

The close-mic measurement he describes is proper, <1cm. Whether or not there is room mode influence there should be the classic PR dip in the woofer close-mic measurment. The only explanation is that the PR is simply not moving or is moving minimally. The table reflection can be ignore given that it will almost certainly be approximately 816Hz measured close-mic at the woofer.

I would also say that if a far field measurement were taken on that table in an anechoic table that the impact would still be well above the box tuning frequency. We're talking about extremely long wavelengths. The tuning being (at least by design) about 60Hz may just be in the range of the anechoic chamber room mode, but the impact on a measurement will vary depending on both the speaker placement and the mic location.

But none of that should have significant impact on a 10mm close-mic measurement.

It will still be good for the OP to take a close-mic measurement of the PR response and measure the impedance.

 

[Jon AA]

How can I perform an impedance sweep? The PR doesn’t have terminals to connect it to the signal generator or to the FFT analyzer.

You do it on the whole speaker as assembled. It's a really fast way to see if something is going on. A DATS V3 is fine for that use. On Amazon

You'll get one of these:

That's a ported speaker tuned to about 38 Hz. My guess is your speaker won't show the "double hump" as your simulation shows in the second pic because something is off with your PR. It's very fast and easy to do--you could pull the PR, remove some weights, slap it back in and do another sweep in a couple of minutes.

What's the ratio of the area of the PR to that of the driver?