Designing a 2-way bookshelf speaker

[captrajesh]

In a 2-way speaker, the tweeter often takes over at 2.5KHz or so, which means that the midbass is pumping the enclosure space full of all sorts of frequencies upto this point. So, the standing waves will have a lot of midrange and HF waves. These need to be attenuated or else they'll break through the cone and distort the sound in strange ways.

You post is understood except what's quoted above. Whatever the reflections, it won't get into the HF territory isn't it?

Simple way is .. Make a Vented Enclosure don't do any kind of stuffing, Put your ear on the vent and you can hear the Mid's & Highs coming out through the vents, Now add the stuffing on the walls of Enclosure & you will find a lot of difference. This i am saying because i have experienced it practically. I don't have proof to prove it by doing some calculation. However i cant say if the same thing is applicable for Sealed Enclosures. May be what you wrote is applicable for Sealed Enclosure.

I understood where you are coming from. Thanks for sharing your experiences. May be I should do these experiments as well.:thumbsup:

 

[tcpip]

What you posted was understood except what's quoted above. Whatever the reflections, it won't get into the HF territory isn't it?

By HF, if you mean tweeter, then yes even that range is possible, but from the midbass.

If a dimension of the enclosure encourages standing waves of, say, 500 Hz, then it will also be a perfect fit for standing waves of 1000 Hz, 1500 Hz and so on. Therefore, it's possible that some weak output of the midbass driver at 3 KHz will trigger standing waves of 3 KHz, 6 KHz and so on. (If a midbass is crossed over at 2.5 KHz, it means that with a second-order slope, it's just 12 dB down at 5 KHz. Plenty of opportunity to trigger HF standing waves.)

Sadik's comments about midrange noise coming out of the port tube is spot on. In fact some designers insist on putting the port on the rear wall so that any midrange noise coming out of it will be inaudible to the listener. [emoji1]

 

[tcpip]

I have successfully used felt behind the midwoofer to prevent the mids and high mids escaping from the cone and adding destructively with the midrange sound. The trick is to allow escape of low frequency and trap only the mids and mid-highs.

Where do you place the felt? On the rear wall directly opposite the back of the driver? Or somewhere in the middle of the enclosure? Where do you get felt from?

 

[Hari Iyer]

Where do you place the felt? On the rear wall directly opposite the back of the driver? Or somewhere in the middle of the enclosure? Where do you get felt from?

I mount the felt immediately on the woofer rear after cutting a small recess for the magnet. The felt loosely rest on the woofer frame at the back. The rest of the cabinet is stuffed with Recron as per my calculations.

Felts are bought by my carpenter, they are also available at LT road in some speakers shop.

 

[Kannan]

I mount the felt immediately on the woofer rear after cutting a small recess for the magnet. The felt loosely rest on the woofer frame at the back. .

Any particular reason for doing this, when the woofer is playing close to its xmax it needs sufficient breathing space in the rear. Isn't blocking it off detrimental to the driver performance

 

[tcpip]

Any particular reason for doing this, when the woofer is playing close to its xmax it needs sufficient breathing space in the rear. Isn't blocking it off detrimental to the driver performance

I too feel that this is unusual, but then Hari, being quite thorough in his approach, has probably checked impedance of the driver in box after adding the felt. If the felt is adding any new resonances, it'll show up immediately as a kink in the impedance curve. If there's no such kink, and the LF response is as expected, then aal eej wail -- he's hit upon a good way to smoothen the rear resonances.

In fact -- though I realise that some members here would like to give measurements and "calculations" a wide berth [emoji1] -- it's relatively simple to detect standing waves and resonances without listening to anything just by taking an impedance measurement of the driver in the box and see if there are any new kinks which were not there in the free-air impedance curve. Tuning the amount and position of the stuffing quickly reflects in changes in the impedance curve and helps to eliminate those resonances. This does not imply that listening is a bad idea; I just want to say that a lot of these things are simple enough to get detected by measurements alone, and there's no unique magic in the listening-based detection approach.

 

[GeorgeO]

I bought felt from a shop in the automotive market area this month.

 

[keith_correa]

Internal Standing Wave Calculator for Loudspeaker Enclosures and Rooms

 

[captrajesh]

If a dimension of the enclosure encourages standing waves of, say, 500 Hz, then it will also be a perfect fit for standing waves of 1000 Hz, 1500 Hz and so on. Therefore, it's possible that some weak output of the midbass driver at 3 KHz will trigger standing waves of 3 KHz, 6 KHz and so on. (If a midbass is crossed over at 2.5 KHz, it means that with a second-order slope, it's just 12 dB down at 5 KHz. Plenty of opportunity to trigger HF standing waves.)

Agreed that there could be a possibility of existence of standing waves higher frequencies inside the cabinet.

If a woofer is crossed over say at 2.5 khz through a second order xover, it still produces 5 khz 12 dB down and 10 khz at 24 dB down so on and so forth; agreed. But considering that there is suficient damping to lap up all those sounds, would it be significant enough to smear the original sound signal is the moot point.

 

[tcpip]

But considering that there is suficient damping to lap up all those sounds, would it be significant enough to smear the original sound signal is the moot point.

Ah, I understand. And I agree that with proper stuffing, it can be taken care of. I was earlier just responding to Sadiq's post that midrange and HF frequencies are present. I agree that they are. The stuffing needs to be done properly to eliminate them.

 

[captrajesh]

Thanks Tarun. I stand corrected.

 

[jmascreen]

There is a Gadiwalas atThane who sells Recron brand polyester fiber. I searched after romaing for 3 weeks

I have seen Aquastuff available at amazon.in. They are not very expensive and are the best to use.

@ jmascreen.
The density of stuffing required for your box is decided by the simulation software.
You will need to convert that to grams reqd by multiplying the volume of your cabinet with the density. Care needs to be taken for unit of measure conversion as you are dealing with multiple units of measure to arrive at the grams reqd.

Thanks Hari, I am looking more towards material density rather than stuffing.

How different it will be with egg crates like material used for wall lining to foam based?

 

[Hari Iyer]

Thanks Hari, I am looking more towards material density rather than stuffing.

How different it will be with egg crates like material used for wall lining to foam based?

I don't have practical experience with egg crates and hence will not be able to comment on this. If you have access to this material you can try.

You will know if you got it correct by measuring the impedance of the completed Speaker system. First measure the free air impedance of the woofer. Note the magnitude of impedance at resonance.. Then build your box, stuff them with egg crates. Then again measure the impedance of the speaker system and note the magnitude of impedance at resonance. If the impedance at resonance of the line drops to 40% to 60% of the free air impedance and there are no wrinkles in the impedance cruves after resonance to say 5,kHz then the egg crates stuffing is working for you else not.

 

[tcpip]

I am happy to report that the speaker building activity has resumed, after a gap of more than a month. Basically, my carpenter disappeared back to his home village due to some personal emergency. I've now been able to find another carpenter, and I showed him the router day before yesterday, taught him to cut holes. Today, he's started work on the front baffle of the MTM, and work is proceeding well.

Life is back on track again.

 

[Hari Iyer]

my carpenter disappeared back to his home village

Hmmm, sounds like my carpenter hyeah:

 

[rajesh srinivas]

I am happy to report that the speaker building activity has resumed, after a gap of more than a month. Basically, my carpenter disappeared back to his home village due to some personal emergency. I've now been able to find another carpenter, and I showed him the router day before yesterday, taught him to cut holes. Today, he's started work on the front baffle of the MTM, and work is proceeding well.

Life is back on track again.

Good news keep us updated with the progress...

Since its a long gap kindly summaries with the driver being used and the box vol..

Rajesh..

 

[tcpip]

Since its a long gap kindly summaries with the driver being used and the box vol

The idea of my keeping a log in this thread is that I'm sharing the journey. This approach is a first for me. If you want a summary, you'll have to wait till everything is done and I write up the design on my website. [emoji5]

 

[HormusPeston]

Thanks for sharing.
Cheers,
~HP

 

[tcpip]

Some photos of the Bhairav (MTM) enclosure under construction.

These show the front baffle. A small amount of bevelling has been done from the inner side, upto a depth of about 6-8mm. More bevelling would not be advisable, because I'd land up losing material to let the mounting screws get a bite into the MDF. This is because the recess from the outer surface is already about 10mm deep, which leaves about 15mm MDF to play with, and the bevelling takes away some more mm of material.

Needless to say, no bevelling needed for tweeters.

This is the stack of larger braces, ready for fitting. The smaller braces are simple rectangular pieces of ply.

And here is one sidewall with all braces, top and bottom piece, fitted, with Araldite and screws.

After the photo was taken, the second sidewall was fixed on top of this assembly and fixed with Araldite and screws again.

Two observations about bracing for this specific enclosure, after the box was assembled and I got a chance to inspect it for rigidity:

• Bracing is needed, even with this small 18" box, even when I'm making it with 25mm MDF, which many consider overkill. (Ready-made Parts Express enclosures of this size are made of thinner MDF and are braced less than I do.) There is clear, audible difference in the inertness of panels which are braced vs the top and bottom panels which are not braced.
• The bracing I've used here is more complex than it needs to be. On hindsight, a simpler set of braces would have done an equally good job. Instead of the two large and two small braces, I think now that a set of three small braces would have done an equally good job. I'll change the design diagrams to reflect the simpler bracing.

In the meantime, I wanted to get the binding post assemblies ready for the carpenter to make cutouts in the rear panel. I discovered that DIYAudioCart is selling the PE binding posts and plates which I've used in the past and loved so much. But then a very Indian thing happened: when I entered the number of pieces I want (3-4 pairs of speakers), they ran out of stock. So, I'm importing the very same items from PE US now. You may of course use any other terminal block or binding posts.

 

[tcpip]

The last three days have gone in completing the boxes (sides, top, bottom, braces) and in re-doing the front baffles. We (my carpenter and I) have built and discarded two pairs of baffles, and are now on our third. Basically, my carpenter took this much time to understand what sort of precautions he needs to take to use a router, and what degree of accuracy is needed in speaker building. The first set of baffles were discarded because the recess he had cut for the midbass units were larger than the drivers by about 1-1.5mm on all sides. I thought hard about it, and finally told him to re-do it. In the entire process, one router bit broke and was replaced. All part of the game.

We're now almost done on our third set of baffles, and I'm hoping it'll now be usable. He is now using a procedure which takes much more precautions to ensure that the router does not jump or kick and thus remove a chunk of material from any part of the circles. He's also learned to measure the settings much more carefully before he cuts.

The nice thing which happened yesterday was the rounding of the vertical edges of the baffles, using a new 1" rounding bit I've procured. This rounding is so even that it'd probably take days to do by hand what the router did in about 10 minutes.