Bipolar Loudspeaker with Visaton FRS8 8 ohm

[Esaú Cardoso]

Hi all,

I would like to share with you the design of the FRS8 8ohm (parallel) shown on Quarter Wavelength Loudspeaker Design Gallery

IMG_9354.JPG

projecto.JPG

The foam quantity of the lower section is a bit less. Any way, you can adjust the amount of bass by putting more or less foam in this area (more foam, less bass but more controlled; less foam, more bass but a little bommy).

The box design was implemented with the very good MJK mathcad spread sheets. The bass is plenty!

In the following graphs are the measured response (rta) and the simulated response on boxsim with the frs8 manufacturer drivers specifications.

curva resposrta originais.JPG

boxsim s filtro.JPG

The results are fairly similar. The curve below corresponds to the power response.

During the process of improvement of these loudspeakers, I found that the design suffers from diffraction problems, so I had to add a notch filter to improve power response.

crossover.JPG

With this crossover, the simulated response (boxsim) is:

boxsim c filtro.JPG

The power response corresponds to what you actually ear. The sound is full and very musical…! However, in large rooms, I found that the response is better without the notch (or reducing the resistor of the notch).

Later on the improvement process, I removed the notch filter and I tried to solve (or minimise) the diffraction problem by changing the behaviour of the front and rear baffle and improving the dispersion of the drivers. I place some absorbing material on the baffle and add a 2,5cm phase plug on each driver. The result was that the diffraction problem was actually reduced although less high frequency as obtained. Nevertheless, they have a very pleasant sound (a little enhancement of high frequency on the power amplifier will be OK).

05092011421_filtered.jpg

As one friend of mine said, they have a very “organic” sound…

Finally, I think that a narrower baffle and a deeper loudspeaker (keeping inicial WxD product) should reduce diffraction problems avoiding the need of the notch filter (in the first version of the loudspeaker), keeping the very musical characteristics and bass response of frs8.

Please be my guest to implement this design or to change it... And let me now the results..!

Esaú Cardoso

[bubuxxl]

Hey there! I recently built cyburgs needles for a friend and became acquainted with the visaton frs8 m, I really liked them! But it definitely had very limited spl to my taste...
Does this project handle more power than the needles?I´m tempted...

[Esaú Cardoso]

Originally Posted by bubuxxl

Hey there! I recently built cyburgs needles for a friend and became acquainted with the visaton frs8 m, I really liked them! But it definitely had very limited spl to my taste...
Does this project handle more power than the needles?I´m tempted...

Yes. You can get twice the power and more 6 dB in sensitivity (connected in parallel). The FR8 version can play really loud. FRS8M will play louder!

I made some simulations on the MJK mathcad sheet and compare both FRS8 and FRS8M.

2xFRS8
FRS8.JPG

2xFRS8M
FRS8M 0.JPG

I suggest you change the plans in the following: put the port 10cm bellow and change the port length to 9cm (or a bit larger).

The response will be (with 9cm):
FRS8M.JPG

The big deep in the frequency response is not audible in room.

As I mentioned in the first post, I also suggest you change the wide dimension to a lower value keeping the original product wide x deep (ex. 8cm x 17.9cm (WxD) inner dimensions)

I also tested the needles but with TB871 (I didn’t try them with FRS8). Comparing both designs, the needles have a more controlled bass (sometimes a bit shy in my opinion), while the bipoles have a more “strong” bass.

Hope I convinced you

Esaú

[phivates]

Do you mean bipole?

[Jim Griffin]

Yes he means bipole where-in both drivers are driven in phase.

Now I will take issue with Esau that you achieve 6 dB increase in sensitivity. You do get 6 dB from doubling the number of drivers and halving the impedance (3 dB from each of those factors). That is only true if the two drivers are both radiating in the same space across the frequency band.

For a bipole speaker both drivers radiate into 4 pi space only below the baffle step frequency. Above the baffle step point the drivers the speakers radiate into non-overlapping 2 pi spaces so no 3dB increase. You do still get the 3 dB increase associated with the impedance halving.

So the bottom line is 6 dB increase when both drivers are connected in parallel and radiating into 4 pi space but only 3 dB increase when the two drivers are radiating into separate 2 pi spaces.

[planet10]

The big dip in the frequency response is not audible in room

I think that a narrower baffle and a deeper loudspeaker (keeping inicial WxD product) should reduce diffraction problems avoiding the need of the notch filter

The big dip in the response deceases as you go off axis and can be deceased by making the box wider than deep... at least 3:1.

A notch filter cannot fix a diffraction problem and will only add problems.

dave

[Esaú Cardoso]

Jim, I agree with you that you will have 6dB bellow step frequency. The reason you refer is not so clear to me: ”… 6 dB from doubling the number of drivers and halving the impedance (3 dB from each of those factors)…”
In general, I see it this way: consider L as being the distance from two identical sources. Roughly, for distances from the two sources below L, both sources are considered coherent and the resulting spl is +6dB. For distances from the two sources higher than L, the sources may be considered as non coherent and the resulting sum will be +3dB.
This is also true when you have both drivers in the same panel.

Now, for a bipole arrangement in free field, you will have the normal behaviour of a monopole arrangement with an increase of 6 dB above the step frequency, compensated with the rear driver with an increase of 6 dB below the step frequency. The result is roughly a flat response over the entire frequency band.
Of course, as Dave says, this is true for a large baffle and a minimum deep shape.

But, theory is theory and some times, our ears don’t agree with theory specially when dealing with small to medium size rooms….

Although I have never tried these large or narrow baffles in bipole arrangement (only more or less squared), I did try them on monopole arrangements (3” drivers). In large baffle, an on axis flat response driver will sound with lack of high frequency (I tried this with FE87). With a narrow baffle, you will have more high frequency but you will probably have to add a notch filter to control the rising frequency response in medium to high frequency due to baffle step in that frequency range. I noticed this with TB871 and FR88EX on a baffle similar to the needles.

So, in particular to FRS8 in bipole arrangement, I believe that they will play fairly well keeping baffle dimensions narrow (needles like) and deep enough so that cancelation occurs below 400Hz (well bellow the “critical” zone of the ears). The room contribution will handle this deep.
I believe that this “proposed” arrangement is not applicable with small driver (3”) with a rising frequency response (similar to F85K) or more significant large driver. In these cases, a large baffle will do fine…

These believes are not totally supported in true bipole implementations, so I would like to ear from you experiences (or believes ).

Finally, I agree with dave in regard to the use of the noch filter when used to compensate problems of the bipole arrangement (not the driver it self). That’s why I removed the notch filter and tried to solve the diffraction problem with baffle absorbent and phase plugs.

Cheers,
Esaú

[hm]

Hello i made a few bipol double horns,
the TROMBONE with two FRS8M
look the measurements and feedbacks
on my HP.