Designing a 2-way bookshelf speaker

[tcpip]

That's not to say what the designer decides is important is "right"! It may be important to the designer but what good is it unless it sounds better.

Yes, the acid test. And then, if you don't hear any difference, you feel embarrassed to admit it to your audiophile friends. Or if the tweak challenges your basic understanding of laws of physics, you feel embarrassed to ask, "But how exactly does it work??". Been there, done that. So I just go away and try to build some speakers. [emoji5]

This is not to say that felt on the front baffle defies laws of physics: it's in fact a well documented and well understood measure to tackle some classes of tweeter problems.

 

[keith_correa]

Been there, done that. So I just go away and try to build some speakers. [emoji5]

Amen to that!

This is not to say that felt on the front baffle defies laws of physics: it's in fact a well documented and well understood measure to tackle some classes of tweeter problems.

You might find this, this and this interesting then.

 

[Kannan]

I had tried the felt tweak on my towers with discouraging results and they were off within the first few minutes of listen.

That said, I have a suede dashmat covering the entire dashboard of my car which substantially absorbs the massive sound reflections from the windscreen as the midrange and tweeters are mounted in the far left and right corners of the dashboard.

 

[tcpip]

You might find this, this and this interesting then.

I read the three articles you cited, thanks.

But I found the approach of the author confusing. I'm referring specifically to the first link here. In it, he starts with a raw baffle without felt, then he progressively adds felt here and there, and shows how each addition of felt improves things. The implication here is that the raw baffle gives a pretty bad SPL curve, which is hard to work with and will colour the sound, therefore improvements are badly needed.

But his starting point is insane. If you see his raw baffle, it's one of the worst baffle designs one can start with.

• It has no edge roundover or bevelling
• It has a horrifying one-inch step between the tweeter and midbass which will create the mother of all edge diffractions.

If he had at least once mounted the tweeter on what we would call a "normal" baffle and included the SPL measurement from it for comparison, most of us would have been able to see that there is no need for felt treatment at all. The SPL of his expensive tweeter would be quite even and acceptable without any felt treatment.

 

[Hari Iyer]

In my recent speaker built with Dayton drivers I had used both a 54 mm step for the tweeter and felt around the tweeter for taming diffraction. The step was needed to time align the drivers as I am using a first order crossover. The felt was used to cover the entire baffle encompassing the tweeter. The step was sloped to integrate and not obstruct with the tweeter.

I could only measure the impedance and phase of this speaker after completing as I don't have the microphones to do the SPL measurements.

 

[tcpip]

... The step was sloped to integrate and not obstruct with the tweeter...

All sensible designers who feel the need for a step do this. The design which Keith Correa had pointed to was using a sharp, vertical step.

 

[keith_correa]

The step on the vertical plane will not do much because the baffle continues below it. But yes, there would be diffraction due to that step. Will that diffraction be audible? Debatable!

Rounding edges is another strawman. Like you already know, for the effect of the rounding to be audible, the radius should be at least 1.5". Most rounding that is done is purely cosmetic. Inaudible!

 

[tcpip]

The step on the vertical plane will not do much because the baffle continues below it. But yes, there would be diffraction due to that step. Will that diffraction be audible? Debatable!

If it's a sharp 90-degree step an inch high, the diffraction will give SPL bumps of more than 2 dB, over maybe an octave here and there. Such bumps are audible. Many graphs available of tweeter SPL over badly designed baffles will show you how large and wide these bumps are.

Rounding edges is another strawman. Like you already know, for the effect of the rounding to be audible, the radius should be at least 1.5". Most rounding that is done is purely cosmetic. Inaudible!

The radius is inversely proportional to the frequency down to which the rounding smoothens out edge diffraction. The larger the rouding radius, the lower down you'll get the benefit of the rounding. This doesn't mean that rounding smaller than 1.5" is useless -- even 0.75" rounding will smoothen out unevennesses, but in the higher frequencies.

An excellent article from one of the Grand Masters Of The Game: http://www.linkwitzlab.com/diffraction.htm

 

[keith_correa]

The radius is inversely proportional to the frequency down to which the rounding smoothens out edge diffraction. The larger the rouding radius, the lower down you'll get the benefit of the rounding. This doesn't mean that rounding smaller than 1.5" is useless -- even 0.75" rounding will smoothen out unevennesses, but in the higher frequencies.

I know that but ALL my points were about audibility/inaudibility. Never said that rounding will not smoothen the response. BTW, here's a sim of the baffle diffraction responses of a tweeter with the 1st being a 90 deg. edge and the subsequent responses with radii in 0.25" increments except the one at the bottom which is with a radius of 2". Draw your own conclusions about audibility/inaudibility. And, enough discussion about diffraction. Let's move on. hyeah:

 

[tcpip]

The first graph shows a peak of 2dB at 1KHz. The last one, with the big-radius rounding, shows this peak dropping to 1dB. This change may not be audible, specially if you take into account that the graph is a simulation of pure on-axis outcome, whereas what we hear in-room is the "power response" which will blend a lot of the off-axis response with the on-axis response.

The difference may not be audible.

 

[keith_correa]

If it's easy for you to lay hands on, and if you are inclined to, see if you can mass load the inside enclosure walls and the outside wall of the port tube [if it's a tube] with floppy 2+ mm silicone sheets. For low cost designs like these, they may mitigate the need for extensive bracing altogether. Just a thought.

 

[tcpip]

If it's easy for you to lay hands on, and if you are inclined to, see if you can mass load the inside enclosure walls and the outside wall of the port tube [if it's a tube] with floppy 2+ mm silicone sheets. For low cost designs like these, they may mitigate the need for extensive bracing altogether. Just a thought.

Give me more details? I know of EVA sheets, no idea of silicone. And in my experience, even lining with kadappa stone has given poorer results compared to bracing.

That said, these two designs will have small panels, and I'll use my usual 25mm MDF. So, bracing needs to be minimal, if at all. Only large panels (floorstander side panels) badly need careful bracing, in my experience.

 

[keith_correa]

Buy Silicone Kitchen Baking Mat Non Stick Oven Proof Sheet Home Tray Liner 15"x11" Online at Low Prices in India - Amazon.in

Not sure but I think these are it. EVA works well too. It's worth using EVA if you can't find silicone sheets.

 

[tcpip]

Buy Silicone Kitchen Baking Mat Non Stick Oven Proof Sheet Home Tray Liner 15"x11" Online at Low Prices in India - Amazon.in
...

These are too light.... 150gm for something almost one sqft in size. I had used EVA sheets which were 4kg/sqm or roughly 400gm per sqft. And I had put 3 layers of that material. In the Asawari 1.

Will these have much impact in mass damping the panels?

 

[keith_correa]

Like I said, I'm not sure of the one whose link I provided. The sheets I was referring to were very dense!

Isn't EVA the material used in craft foam sheets? I thought it was. But I don't think they weigh 400 gms psf so maybe I thought wrong.

A loony acquaintance poured very diluted wall putty over the inner walls. Lined all inner walls with it. Said it worked quite well.

 

[Hari Iyer]

The diffraction imo also depends on the filter order and crossover frequency. If the crossover frequency is higher, you need to add felt only nearby the tweeter.

 

[keith_correa]

The diffraction imo also depends on the filter order and crossover frequency. If the crossover frequency is higher, you need to add felt only nearby the tweeter.

There are many ways to skin a cat. :lol:

 

[tcpip]

Isn't EVA the material used in craft foam sheets? I thought it was. But I don't think they weigh 400 gms psf so maybe I thought wrong.

Look at this!

https://www.amazon.com/EVA-density-Foam-Sheet-x0-5/dp/B01LR0776Q

One kilo per sq foot!

I'd used sheets which were perhaps 3mm thick but were 400gm/sqft.

The light EVA sheets people see are EVA foam, deliberately designed to be light. My stuff looked like black rubber sheets.

 

[Hari Iyer]

This is what i did to tackel diffraction issue. There is anoffset of 54mm between woofer and tweeter, and the tweeter area is having low felt material to absprb mid to high frequency diffraction.

 

[jmascreen]

The Vifa XT25SC90 just came in. Kept beside the "normal sized" Peerless India tweeter, it looks unbelievably tiny. It'll be fun playing with this tiny tweeter for the MTM.

Any clue about the positive and negative terminal in peerless tweeter?