Strange OB dipole behaviour ...

[paulspencer]

Well, I finally tried a fully open baffle setup, and the results have me puzzled. Firstly, I'll say I'm impressed. After a number of hours of listening, and watching King Kong, I notice the following:

* much bigger sweet spot
* midrange is more transparent
* bass is more accurate than I thought my drivers were capable of, most noticeable is that it is subjectively much more "quick" and it makes me aware of how much was added to the sound. In particular I notice this on things like kick drum - very convincing as if it is actually in the room - previously bass sounded exaggerated, but not due to frequency response

Before:
TL mains with Vifa P17 in MTM with two push pull sealed AV12 subs in the corner, EQ'd flat with Ultracurve (including room modes which aren't that bad in this room)

After:
The same drivers placed on 400mm wide x 500 tall main panels with 50 - 150mm wings and subs push pull H frame approx 400mm wide and deep

I notice the tonal balance doesn't sound the same, though the response is calibrated flat. The bass sounds more lean and quick, also the midrange does as well. I put this down to box coloration, and perhaps time domain performance.

Now here's the mystery:
The EQ required to compensate for dipole roll-off is ALMOST non existant!

Here are the settings used by Ultracurve to calibrate it flat. The response in the middle shows the "before", on the top a monopole sub is used and on the bottom is the full OB setup. Note, this is the EQ settings, so the actual response is close to the inverse.

Around 35 Hz there is a room mode peak which is EQ'd out in all of them.

So my question to OB gurus:

How do you explain the absence of EQ required? The main panel should have an fequal around 200 Hz, and similar for the H frame.

Another Q:
When calculating Fequal for a H frame, should D be taken as merely half the depth, or should I add half the baffle width?

[paulspencer]

This is what it looks like (ugly I know)

[paulspencer]

As you can see, the construction quality and finish is outstanding! ... 5kg weights keep them from falling, and I'm using my existing passive 2.5 way crossovers

Also as you can see, those panels were precisely calculated for target fequal after hundreds of hours of research, simulation and testing ... *coughs* ... well, actually they were what I had lying around ...

[ashok]

Hi Paul,
Great to see the MTM VifaP17's up in an open baffle . I have been planning this for a couple of years ! However I keep looking for more info on other implementations and their results.

After reading your post I've decided to go ahead and build one myself. I've had 4 P17's stashed away just for this kind of a design. My tweeters however are Vifa , D27AG , the XT25 ring radiator and a Dayton Planar PT2. Not quite sure what I should use.
What crossover frequency did you use with the P17 ?

I did want to use a narrower baffle but most things I read indicate that a wider baffle is preferable . I have been toying with a partially wooden and partially perspex baffle so that it "looks" smaller than it really is. Lots of designs around with this scheme.

Your system looks unfinished but OK. Certainly not ugly as you said. Some self adhesive wood finish vinyl will give it a nice appearance. The "stand" could be improved of course ...black ?
Cheers.

[paulspencer]

Ashok,

You might want to also have a go at the Vifa P17 tweak which I'm in the process of doing. It involves putting glue rings on the back of the cone and adding a phase plug. I've seen documentation for this and the improved frequency and time domain performance. I haven't had a chance for a proper AB comparison, but there does seem to be some improvement. I wrote about this in another thread I started recently.

I'm using my existing 2.5 way xo which is 2nd order and xo @ 3.5k. Ideally this is a bit high WRT off axis lobing but it's good for distortion and power handling. The 0.5 midbass has a first order LP @ 200 Hz.

You could try a 2.5 way where you have WMTMW. I'd try this where the outer woofers have a first order LP. This way you won't have the off axis lobing problem. I'd go with the ring radiator and experiment with a small waveguide to allow it to go a bit lower and take some strain off as they say it tends to want a 3rd order xo higher up than most would prefer to use.

Regarding baffle width, I did come across some guys who measured different widths compared to Linkwitz Phoenix design. Apparently its a trade off between lowering Fequal and polar response. If I stick with OB and make some decent ones then they won't be as wide.

I've been thinking of putting a small H or U frame below each main panel, or I could put the main panels on stands. I might also experiment with a corner dipole woofer.

Next I want to try:
* U frame woofer
* waveguides on the main panels, similar to Zaph's

[DorinD]

Regarding the mistery lack of EQ for OB I am thinking that 12dB/oct gain boost due to the room dimensions shorter than ¼ of a wavelength partialy compensate for those 6dB/oct roll-off the OB?!

[paulspencer]

That would mean the TL/subwoofer setup would show room gain, yet it doesn't!

What confuses things more is that dipoles are meant to couple differently to room modes and gain as I recall ...

[kneadle]

Originally posted by paulspencer
That would mean the TL/subwoofer setup would show room gain, yet it doesn't!

What confuses things more is that dipoles are meant to couple differently to room modes and gain as I recall ...

Hi Paul,

I'm curious, too. What are your room dimensions?

My under-informed guess is that the dipole behavior causes the room to behave differently.

I'm hoping to see an informed guess...

Dave

[paulspencer]

3.9m wide x 4.9m deep x 2.9m ceiling height

12'10" wide x 16' deep x 9'6"

[nunayafb]

How much space is behind the entertainment center? and how far are the speakers from the E.C. and the wall?