Second driver as MFB sensor?

[Mr Evil]

I have two identical drivers I want to use in a subwoofer. They would need a prohibitively large box used together normally, so I thought an isobaric configuration would be nice. I'm also interested in motional feedback for the potentially low distortion and tweakable frequency response.

One thing that's always bothered me about MFB is the sensor. Searching revealed plenty of discussion on the subject and the limitations of dual voice coils, accelerometers etc.

Dual voice coils is the most appealing from the point of simplicity, but it needs a special driver. So what would happen if the second driver in an isobaric configuration ws used as feedback sensor instead of being powered?

Ideally it would yield the same output as a second voice coil in one driver, without the disadvantage of inductive coupling. It seems clear that it wouldn't be so effective at higher frequencies due to the finite stiffness of the coupling air mass, but for a subwoofer it might just work. It would also mean doubled moving mass. However it has the advantage of being very easy to implement.

So, has anyone thought of this before, or even tried it? Any particular reasons why it might be a good or bad idea?

[markp]

I think it will be too loosely coupled. The 'passive' driver will also need to be the external driver in a isobaric setup so it wont have the impact of a driven unit. If it is the internal driver it will not track the active woofer closely enough to be usefull as a feedback source.

[Svante]

If we are talking subwoofers here, I think the main drawback would be the reduced efficiency.Compared to a single driver it would drop to a fourth, compared to the isobarik driver system it would drop to half. But as xmax is still the same, the maximum acoustic output for low frequencies will probably be the same, as long as the driver can stand the quadrupled electrical power for that acoustic output.

I think the best way to do MFB is an accelerometer on the cone, there has been a lot of errors in the debates on this recently. I will not enter these discussions again, but just note that since the sound pressure generated by a loudspeaker cone is proportional to the acceleration of the cone at low frequencies. Some people argue that it is proportional to the cone velocity, but they are plain wrong.

So, the accelerometer is the logical choice.

[johnnyx]

If you build an isobaric enclosure, designed as such, then you could try MFB and see. If it doesn't work you've still got an isobaric sub. The coupling would probably be ok, good enough to try. I thought about this myself. If the enclosure is vented, I would close the vent when trying MFB.
I didn't find the lack of efficiency to be the problem, but the lack of excursion is. There are now some good dual vc units with large Xmax I'd like to try; maybe then the reduced efficiency would become a problem.

(I agree that an accelerometer is probably the best method, as used by Velodyne in their subs. One day I'll try that too.
It's the emf that is proportional to velocity, like an electric motor)

[Mr Evil]

I come here looking for quick fixes and find I have to actually get off my **** and do some real work?!

Ok, I took a couple of pairs of speakers and did some measurements. I put them in a sealed box, in isobarik configuration with the inner driver powered and the outer one connected only to my oscilloscope. I measured the voltage out from the outer driver's coil, both magnitude and phase relative to the input to the inner driver, from 20Hz to 300Hz.

I've attached the results from the first pair (some unidentified old Wharfedale 8" drivers) to this post.

The results suggest that the outer driver is acting more like a passive radiator, showing a large resonant peak. I think it would be hard to use the signal for feedback.

[Mr Evil]

Here's the results from the second pair (an unidentified pair of 8" Pioneer drivers).

Interestingly the peak appears to be at exactly the same frequency as from the previous pair of drivers. I'm not sure if it's just coincidence that the two pairs have similar characteristics or if there is some more fundamental reason for a peak at 80Hz.

[moamps]

Originally posted by Mr Evil
The results suggest that the outer driver is acting more like a passive radiator, showing a large resonant peak. I think it would be hard to use the signal for feedback. [/B]

Hi,
Your results are expected.IMO.
From V=B*l*v, you get voltage proportional to expected cone velocity. Svante said something about this.

Interestingly the peak appears to be at exactly the same frequency as from the previous pair of drivers.

Sealed box limitations?

Regards
Milan

[MarkMcK]

Starting with post number five, I really like this thread. It is great because you discovered your own error. You did it yourself.

The problem was modeling the system as an isobaric. Without driving both drivers in push pull you do not have an isobaric. Instead you have a dual chamber, passive radiator enclosure design.

My advice is to not waste any more time on this project unless you want to learn more about complex vented enclosure systems. If you do, I encourage you to go for it. If not, then it is far better to try selling your drivers to some poor unsuspecting soul and then using the money to buy drivers more appropriate for whatever subwoofer design you have in mind.

Good designing and good building,

Mark

[Mr Evil]

Originally posted by MarkMcK
...buy drivers more appropriate for whatever subwoofer design you have in mind...

I don't have any design in mind. I just want to use what I have to hand and try something new if possible.

Of course in retrospect it seems clear what is happening and that it wouldn't work for MFB... or would it?

It occurs to me that this could be applied to solve two problems inherent to MFB, namely the excessive excursion limits if bass is extended far, and the difficulty of applying it to anything but a sealed box.

There could be two sources of MFB: One from the main driver and the second from a passive radiator. Sum them and it would be possible to control output from both driver and passive radiator at the same time.

I would test it, but I would run into budget limitations (which is at about 0 at the moment).

[MarkMcK]

Sorry, I thought you had figured out that it would not work for MFB on your own. The two woofer fluid coupled design has too many variables to be practical. Without quality instrumentation and lots of effort, who knows what you may end up correcting for with the MFB signal.

If you just have to make one MFB subwoofer from these drivers it would be far better to remove one of the cones and mechanically couple the two voice coils. The drivers would have to be precisely aligned, but you have a bit of mass to work with the cone being gone.

It would not work as well as one dual VC woofer, but at least you would have a chance of pulling it off with minimal instrumentation. I would also advise holding the MFB to 6 to 10 db.

Mark