[opnly bafld]

Rudolf,
Thanks again for your time and effort, it is much appreciated. 

Lin

[Bob in St. Louis]

Thank for trying my suggestion Sir!!

Bob

[D OB G]

Hi Rudolf, and thanks for your contribution to some hard research.



This is my attempt at simulating the 'driver up' configuration (at 260 cm), noting that The Edge doesn't take floor reflections into account.

[D OB G]

Obviously the screen capture didn't come through.
Can anyone help me.
The image insertion read: .
Is this the right format?

[D OB G]

It seems that even writing the text creates an image that won't insert.

I'll try to fool it by trying to show what I inserted this way:

"square bracket"img"square bracket"edge visaton 2 w250.jpg"square bracket"/img"square bracket"

What's wrong (the "square brackets" point the right way)?

[JohninCR]

Rudolph,

Interesting stuff.  What surprises me is that the measurement distance only seems to show a difference in the general slope of the bass response with the woofer up away from the floor.  Can you tell us some general info about the room size and approximate speaker placement for your measurements.  I think I need that to wrap my brain around your results.  Again, thanks for the effort and for sharing.

John

[Daygloworange]

Yes, great effort. 

Thanks for taking the time to post your measurements. Very interesting.

Did you run impedance measurements for the various configurations as well?

Cheers

[D OB G]

In the absense of the Edge graph, it is impossible to show a comparison, but the most notable difference is that the response is "flat" down to around 300 Hz, and -3 dB at about 170 Hz as it develops the 6 dB per octave loss (when I say flat, a 3 dB peak occurs at 700 Hz, and a 2 dB peak around 2 kHz).

This makes an interesting comparison with your "real world" measurements showing a drop off at around 500? Hz, and -3 dB at, say, 350? Hz.

I can only invite those interested to try The Edge themselves, but it appears that it is almost twice as optimistic as these real measurements show.

David

[Rudolf]

What surprises me is that the measurement distance only seems to show a difference in the general slope of the bass response with the woofer up away from the floor.  Can you tell us some general info about the room size and approximate speaker placement for your measurements.

John,
the room is roughly 5,3 x 4,2 m with the OB standing 1 m in front of the front wall (4,2 m wide). There is a 30 cm deep open rack on the wall behind the OB. The baffle is placed at 2/5 of the room width. Indeed the measurements show less difference between the driver positions than one might expect.

[D OB G]

I've finally remembered how to post an image.

So as far as I can tell, the Edge simulation suggests lower frequency extension at Fe than actual measurements show (and given that it is a simulation of a baffle in free air, i.e. not resting on the floor, you might expect it to be a bit higher).

[D OB G]

Here is the correct Sd



This graph especially begs the question: where is Fe?

What is correct, that there is a reference level with a large peak, or that Fe is the top of the peak, with some unexplained? higher frequency baffle drop-off?

(The second scenario doesn't? seem to accord with your graphs, except in so far as the driver itself has a high frequency roll-off).

David

[Rudolf]

My measurements give me the opportunity to compare the real world with some simulations. David has already shown, that EDGE on itself does not help much for lower frequencies (albeit simulating in EDGE with a mirror image would be nearer to reality). MJKs worksheets http://www.quarter-wave.com/Models/MathCad_Models.html are a much better way.
First I simulated (red) my H-frame from the same 60 cm distance as the measurement (black):



Next I simulated the OB in the nearfield at 20 cm distance. The conformity is close enough:



For a comparison I simulated the OB at two distances. Light colour is for the simulation:



It´s very obvious: While at 20 cm simulation and measurement are rather close, the divergence at 120 cm distance is sincere. The later measurement is almost in the middle of the room, so accounting for floor and front wall only - as MJK does - can´t reflect reality completely. But neither can any other program.

A last simulation I want to show is the cone excursion of both baffles:


The dotted blue line is for the infinite baffle.
My conclusion: MJKs worksheets will help you very much to design the best dipole for your needs, but the room will be a major factor in what you will hear in the sweet spot.

One day later:
This is how the MJK worksheet see the difference between the driver mounted low and high on the OB. The large dip at 430 Hz is the floor reflection for the driver mounted at 56 cm.

[scorpion]

Rudolf,

Extremly good and interesting comparisons. I was going to remark that your 120 cm measurements fall a bit off the line if
compared to MJK's Math-Cad sheet and also if compared to the 60 cm and 260 cm measurements for the plain OB.
In the 60 and the 260 cm measurements the response is flat (relatively at least) above 200-250 Hz up to the upper roll off as one would expect, while the 120 cm
result is behaving quite differently. That the results for the up and down placed units doesn't differ more is a bit surprising considering that you have
good help of the floor. Baffle dimensions can as I have put forward be quite small if floor reflection is accounted for.

However that the results with the 45x70 OB and the H-OB does'nt differ more is not so surprising as dimensions are fairly equal.

Also which perhaps has not been quite put forward is that the Visaton W 250 is quite a capable bass unit for OB use.

Another thought, with the help of Martin's models one could simulate the ideal OB bass unit. I wonder if someone has given that angle a thought.
I think such an excercise could qualify as serious scientific research. There will be some combinations of TSP-parameters that will give superior OB Bass result, and these will not only be dependent of Qts value. There is more to it ! 

/Erling

[Rudolf]

Erling,
thanks for your friendly remarks. And yes, there are measurements that are not like expected, like that H-frame and OB comparison at 120 cm. The first two results in my last post make me even more confident, that sims with the MJK worksheets will deliver clearer results than that mess of reflections and room modes which I measure with JustOct.
With regard to the OB and H-frame comparison I expected a close race too, since the effective dipole radii would be roughly equal. What I still want to show is that you can get equal performance to a plain OB with much less visual impact. Your U-baffle does support that as well.

I see that we really are in the same boat when it comes to price-performance-ratios. Those Visaton drivers are just basic drivers, but with reliable values. No quality outliers, no strange noise. Nevertheless I am on the verge of buying some Perless SLS drivers to replace the Visatons.

Rudolf

[Rudolf]

I did another comparison between a plain baffle and a "tapered" H-frame:



The baffle is 80x30 cm with the wings base protruding 21 cm to the front and back. The baffle is tilted back 5°. Driver is a Peerless SLS 10. The camera perspective might suggest otherwise, but the baffle is rectangular with the wings perpendicular to it.

I measured the baffle without wings (black) first and then added the wings (red). Distance 1 m. The diagrams are smoothed 1/3 octave:
 


By and large I win about 6 dB by adding the wings, which is the same as adding a second SLS 10 at floor level.

I then turned the H-Frame 180° and measured the response at the back side. With the tilted baffle the wings inclose a slightly bigger volume in the back than in the front. So the back SPL (blue) is a bit more than the front radiation, but symmetrical enough for the planned operation up to 200 Hz:



Finally I want to build H-frames with equally tapered wings, but with a SLS 12 at the bottom, SLS 10 above and a 7" midrange with 1" tweeter on top, resulting in a tapered baffle too.

[scorpion]

Rudolf,

Interesting design and measurement. Am I right to conclude that the side panels are triangular with a base of (21+21+baffle thickness) cms and about 80 cm high ?

/Erling

[ttan98]

Rudolf,

Just a point of clarifications, the actual woofer implementation would be a pair of woofers mounted on your "tapered" H-frame, one firing forward and the other firing towards the rear.

If that the case then you would see the rear of one woofer facing you? The general appearance may not look too good.

I may be wrong if this the case please explain.

[Rudolf]

Am I right to conclude that the side panels are triangular with a base of (21+21+baffle thickness) cms and about 80 cm high?

That´s exactly right, Erling. Base length is 44 cm total. This H-frame is just a test bed. Later I will cut off the upper 20 cm of the triangles and put in horizontal dividers between and above the SLS to see if that makes any difference (I believe the difference will be quite small).
The basic design is no invention of mine. I first saw it here:
http://www.diyaudio.com/forums/showthread.php?postid=752224#post752224


 
I will taper the final baffle in the same way. In the end the midrange driver will see a smaller baffle than the bass drivers and no protruding wings to the sides.

[Rudolf]

Just a point of clarifications, the actual woofer implementation would be a pair of woofers mounted on your "tapered" H-frame, one firing forward and the other firing towards the rear.

I may mount the lower woofer with the basket to the front, yes.

If that the case then you would see the rear of one woofer facing you? The general appearance may not look too good.

In that case I intend to experiment with some front cover. I´m not sure yet, whether that will add too much visual "volume" to the contraption. I will have to do it to find out.

[scorpion]

Rudolf,

I suppose you already have seen it otherwise you will be interested, a new MJK-study of OB, U- and H-frame performance:
http://www.quarter-wave.com/OBs/OB_Theory.html .

/Erling