Baffle surface for a long ribbon

[StigErik]

Originally Posted by LineSource

To reduce lobing, it is necessary to keep the bass-ribbon separation distance less than the crossover quarter wavelength. A 200Hz crossover requires <17" seperation

I solved this by using two H-baffle bass towers on each side with the ribbon in the middle. The towers are spaced more than 1 meter from each other.

[solhaga]

StigErik and LineSource:
I think I then have to reset my input criteria.
I have actually tried (and understood) all the things that you mentioned, but obviously ended with the wrong conclusion.

This is what the bare ribbon driver looks like using FEMM:


Air gap is 32 mm.

The area noted as air is actually quarter round wood at front and aluminium at the back. The aluminum is hollow and holds the return cables.

So I should have quarter rounds at the end of the baffle as well:


Then I would end up with 6 cm baffles.

Perhaps I should move the return cables to inside the U.

[solhaga]

Originally Posted by StigErik

I solved this by using two H-baffle bass towers on each side with the ribbon in the middle. The towers are spaced more than 1 meter from each other.

Yes, I saw your pictures, StigErik.
I don't have that kind of large room that you have.
I have to cope with 3,2 x 5,0 m.

I don't "want" H-baffles, for the bass, I'm aiming at OBs.
But if I don't need to isolate the bass drivers from the ribbon drivers it will be a much easier task to experiment with the bass baffles.
I just make a broad enough baffle to hold the bass drivers and then just add on whatever extensions I like.

[solhaga]

Originally Posted by Andersonix

..
What do you not like about the K340?
..

I still like them, the reason for buying and using them though was to be able to play music loud at 2 in the night not disturbing my neighbours.

[StigErik]

I have tried fairly narrow flat baffles for woofers. Didnt work all to well because of too much dipole cancellation and resulting distortion. I found out you must have large path length to get good bass, and that means H-baffles or extremely wide flat baffles....

I will recommend placing the dipole peak of the woofer baffles at or just below the crossover frequency.

[solhaga]

I got a suggestion to place the side baffles at an angle somewhere between 90 degrees (H-frame) and 0 degress (OB).
Then perhaps letting one point forward and one backward.
One could even have them in a X shape (from above that is) and thus have the foot print more forward/backward than left/right.
Having all these four baffles foldable would also allow you to experiment and also the woofer can be tucked away when not listening.

Some quick ASCII art:

[solhaga]

Actually, this was my setup a couple of years when I had my old subs:


So removing the kitchen drawer mat:


and removing the ribbon baffles:


continuing with switching woofers and ribbons and adding the quarter round wood:




resulted in this new old setup:


R.I.P.:


So now I'm ready for my new OB woofers...

[StigErik]

Originally Posted by solhaga

Yes, I saw your pictures, StigErik.
I don't have that kind of large room that you have.
I have to cope with 3,2 x 5,0 m.

Well ... in my opinion, near-field listening is never a bad idea, and much less so if the room is small.

[LineSource]

Someone needs to talk about the Elephant in the room.

In your small room, your speakers are as big as an elephant.

From what I have learned, your IMG_022 arrangement with the narrow baffle ribbon on the inside sound stage has the potential for great sound. Dipoles sound best when they are 4' or more away from every wall, and in your small room the H-frame is right next to the side wall. Personally, I would experiment with minor changes to IMG_022 and understand that the small room will limit its full potential.

I continue to learn from JohnK's work on controlled directivity dipoles like the NaO NOTE. musicanddesign.com A review might generate some interest in his research and possible construction of a smaller 4-way dipole that is better size matched to your room. You could DIY a wide ribbon dipole tweeter. JohnK just finished a new narrower NOTE bass frame. In the USA, the AE Lambda TD10D dipole midbass and TD18D dipole woofer are speaker porn. I use old Lambda TD15D dipole woofers.

Most NdFeB ribbons use a simple construction where the magnets are next to the ribbon on the inside of the pole steel, and several thin rear steel cross supports are added for strength and to close the magnetic circuit to increase the gap flux. A FEMM simulation quickly shows where stray magetic fields between the rear steel bracing can cancel a portion of the gap magnetic field.

[solhaga]

I'm very well aware of the elephant but it is an elephant by choice.
In fact, the room layout isn´t a straight 3x5 m rectangle.
On the left side of the left speaker there's a door opening into another room and on the right side of the right speaker there´s a garret, 1,2 meters wide and 70 cm deep.
I guess they both widens the room for the dipole woofers.
But sure, it could never hurt to move the away from the wall a couple of inches.

In the left picture you suggest to move the ribbon speaker close to the woofer baffle. Could you please explain why? (Doing it is not that simple as the previous changes).
I see that the ribbons are not flush to the woofer, my mistake. I have already a 0,73 ms delay for the ribbon speakers to correct for a flush mount.

Very interesting setup in the middle picture.
But AE elements are even harder to come by in Sweden. And then there´s the shipping cost and toll and taxes to pay so the total cost is nearly the double.
The modified DAC I currently have only supports 2x2 24/96 outputs and as I kind of like the USB way of solving the PC to amplifier problem I just wait until there´s some USB 3.0 DACs with more 24/96 outputs so I can still use XO and PEQ in the digital domain for all element´s amplifiers.
Also the tweeter would need some kind of transformer I reckon and it is not (currently) my domain.
All in all, I think I stick with my simpler setup for a while though.

I don't understand your last paragraph.
Do you mean that I should close the circuit or not? If you can elaborate on that I would be pleased.
When I simulated a closure of the magnetic circuit in FEMM it didn't give much due to the saturation anyway.
My concern was that little was gained compared to that it wouldn't be a true dipole any more.

Looking at the driver again, top view:

The ribbon is between the two middle dots.

A flux density plot:


The flux is uniform in vertical plane (y) for practical ribbon excursions but not in horisontal plane (x):

I have solved by using the method with different widths of the aluminium strips giving different current density; wider at the ends where the flux is stronger and narrower in the middle when the flux is smaller.
So Bxi is somewhat linear over the gap giving the same resulting force over the ribbon's surface. But I guess you already know all this.