I've built a speaker cabinet for a pretty good speaker with a big magnet. I'm wondering if it needs a back? My Dad once told me it improved the sound quality. In fact, he said, ideally there should be a small vacuum inside. What is the reasoning here?

 

Hello,

What kind of speakers are you making?
Is it a woofer or for wide range?

Greetings,
Bertus

 

Hi-fi speakers need an enclosed cabinet (with or without bass reflex ports) or the woofers response will be underdamped, leading to a muddy, booming sound. The woofers might be more prone to damage by over-excursion too. Google Thiele-Small; also infinite baffle.

I've never heard of running a low cabinet internal air pressure before - assuming it was possible to seal a standard cabinet that well (it isn't) the major effect would be an undesirable lowering of the headroom on negative excursions.

Guitar speakers are generally designed for cabinets with partly open backs, and have a stiffer foam surround to provide greater intrinsic damping. They're also much more resilient to being overdriven.

 

The enclosure for a speaker must prevent the sound wave from the rear of the speaker from mixing with the sound wave from the front of the speaker. If the sound waves mix then bass frequencies are cancelled. Guitar speakers don't care about deep bass frequencies so they are frequently open at the back.

The size and type (sealed or ported) of the enclosure must match the spec's for the speaker.

 

I've built a speaker cabinet for a pretty good speaker with a big magnet. I'm wondering if it needs a back? My Dad once told me it improved the sound quality. In fact, he said, ideally there should be a small vacuum inside. What is the reasoning here?

You might want to look up internet references to "infinite baffle" to discover what a speaker cabinet actually does.

If you have a speaker just sitting in free air, most of the sound will be cancelled. Assuming the speaker is moving forward, the increase on air pressure in front of the con simply spills around the edge of the cone into the low pressure area BEHIND the cone.

To eliminate this leakage between the front and the back of the cone, you need to isolate them with some kind of a wall. The larger the wall (the baffle), the more ideally the speaker will behave. Since most of us don't have the space for infinite sized baffle, we use a compromise...the speaker cabinet. The larger the speaker cabinet is, the better it will work at baffling low frequencies.

A sealed cabinet will also work as an infinite baffle, but creates entirely new problems, so it's not the simple solution it seems. So, most speaker designs came about as an effort to approach an infinite baffle with a reasonable sized enclosure.

Eric

 

Very interesting stuff here, from all of you. Would an infinite baffle approximation be had if you mounted a speaker on a wall facing one room while you let the back of it point into the adjacent room?

The concept of audio vibrations cancelling one another is strange to me. Is this similar to electrical signals of the same amplitude and 180 degrees out of phase cancelling?

 

Very interesting stuff here, from all of you. Would an infinite baffle approximation be had if you mounted a speaker on a wall facing one room while you let the back of it point into the adjacent room?

The concept of audio vibrations cancelling one another is strange to me. Is this similar to electrical signals of the same amplitude and 180 degrees out of phase cancelling?

Indeed! In the 1950s there were quite a few systems described in the audio magazines at the time that mounted speakers directly into the wall of the "sound room"....back when a lot of audiophiles had sprawling suburban homes for such extravagance.

There is indeed a close correlation between acoustic phase cancellation and electical wave cancellation. In fact, you can easily see changes in speaker electrical impedance by changing the mechanical environment! A speaker in free air will generally have a very sharp resonance and high impedance at that resonant frequency. As you put it in a cabinet, you will see the impedance drop (as it is now doing some real WORK, in that the forward and reverse pressures no longer cancel), and the Q of the circuit will drop as well. The frequency of resonance generally drops, as well.

Eric

 

I'll be dog gone, Eric, that's wild! So the the mechanics of air pressure vary so does the response of speaker? At least I understand you saying that. After all a speaker is a transducer responding to its environment, not a constant forcing function.

 

I'll be dog gone, Eric, that's wild! So the the mechanics of air pressure vary so does the response of speaker? At least I understand you saying that. After all a speaker is a transducer responding to its environment, not a constant forcing function.

Hi Paul:

Yes, this is very easy to demonstrate. All you need is a signal generator, a 10k resistor, a raw speaker, and an AC voltmeter or oscilloscope, if you have one. (I prefer the oscilloscope).

Wire the 10k resistor in series with the speaker. Feed the resistor and speaker with a volt or two from the signal generator. Suspend the speaker in free air with a clamp or a vice. Look at the voltage across the speaker with the oscilloscope as you sweep the signal generator frequency. You will see a sharp peak in the voltage at the free air resonant frequency. It will probably be a bit louder at this point. Now take a soggy towel and dangle it in front of the speaker, and watch what happens to the voltage!

Eric

 

Eric has given you some pretty good pointers, but I haven't seen an answer to this question

In fact, he said, ideally there should be a small vacuum inside. What is the reasoning here?

A vacuum would be a bad idea. I suppose your dad thought to counter the build up of pressure inside a closed cabinet.

A speaker cone forms part of the walls of a cabinet so if you think about a speaker cone moving backwards into a cabinet it must reduce the internal dimension of that cabinet. this is another way of saying it must compress the air inside. This is another way of saying it must raise the internal air pressure. So called 'soft suspension' speakers rely in this increase in pressure to support the cone, which can even be damaged if the enclose looses its airtightness and so stops the pressure increase. Soft suspension speakers are used for large power purposes, particular for bass.

 

A sealed enclosure causes the resonant frequency of a speaker to increase from its free-air resonant frequency.
If the enclosure has the correct size for the spec's of the speaker then the frequency response will be flat down to the resonant frequency.
If the enclosure size is too big for the spec's of the speaker then the low bass frequencies will not be loud enough.
If the enclosure size is too small for the spec's of the speaker then the upper bass frequencies will sound too loud and sound boomy.

 

Hi Paul:

Yes, this is very easy to demonstrate. All you need is a signal generator, a 10k resistor, a raw speaker, and an AC voltmeter or oscilloscope, if you have one. (I prefer the oscilloscope).

Wire the 10k resistor in series with the speaker. Feed the resistor and speaker with a volt or two from the signal generator. Suspend the speaker in free air with a clamp or a vice. Look at the voltage across the speaker with the oscilloscope as you sweep the signal generator frequency. You will see a sharp peak in the voltage at the free air resonant frequency. It will probably be a bit louder at this point. Now take a soggy towel and dangle it in front of the speaker, and watch what happens to the voltage!

Eric

I shall do what you said. There's nothing like experimenting to impress the need for compensation on me. This is very interesting. I think we are dealing here with the variance of physical parameters aquired by transducers. And such devices tend to vary with physical conditions. They are not ideal. But they reflect a variety of conditions. Thanks, Eric.

 

A sealed enclosure causes the resonant frequency of a speaker to increase from its free-air resonant frequency.
If the enclosure has the correct size for the spec's of the speaker then the frequency response will be flat down to the resonant frequency.
If the enclosure size is too big for the spec's of the speaker then the low bass frequencies will not be loud enough.
If the enclosure size is too small for the spec's of the speaker then the upper bass frequencies will sound too loud and sound boomy.

Audioguru, what determines the resonant frequency of a speaker? I never even knew they had a resonant frequency. You make it sound like they have inductance -- which I know they do -- in parallel with a capacitance, which seems to exist everywhere. And so you seem to have a good point. A speaker has a resonant frequency. So what is the significance? What in the world differs here in respect to the size of the cabinet?

 

The resonant frequency of a speaker is determined physically (not inductance and capacitance). The weight of its cone (and coil) and the flexibility of its suspension (including the springiness of the air inside its enclosure.

The frequency response of a speaker drops below its resonant frequency so for good bass response you want the resonant frequency to be low.

A heavy cone causes a low resonant frequency but it requires a high power to move it so it is inefficient.

The correct size of the speaker's enclosure and the extremely low output impedance of the amplifier damp the resonance of a speaker so it does not sound boomy.

 

The resonant frequency of a speaker is determined physically (not inductance and capacitance). The weight of its cone (and coil) and the flexibility of its suspension (including the springiness of the air inside its enclosure.

The frequency response of a speaker drops below its resonant frequency so for good bass response you want the resonant frequency to be low.

A heavy cone causes a low resonant frequency but it requires a high power to move it so it is inefficient.

The correct size of the speaker's enclosure and the extremely low output impedance of the amplifier damp the resonance of a speaker so it does not sound boomy.

http://en.wikibooks.org/wiki/Engineering_Acoustics/Moving_Coil_Loudspeaker

Here's a really great link that shows the equivalent electrical and mechanical model of a speaker. As you can see, the mechanical resonances manifest themselves as an equivalent electrical resonances (as well as a TRUE electrical resonance usually situated far above any acoustic interest!)

In the late 50s and early 60s, people began to take a serious analytical approach to speaker cabinet design...much before then, it was a fairly trial-and-error approach. I have an old RCA Radiotron Engineers Handbook from that era that goes into tremendous detail on this. You could really make a career out of this...actually many people did! It's always been a subject of deep fascination for me...so simple in appearance..but incredibly subtle and complex....so many diffferent disciplines involved.

Eric

 

If you are interested I gave the maths of speaker action in post#8 of this thread.

http://forum.allaboutcircuits.com/showthread.php?t=20775&highlight=loudspeaker&page=2

 

Excellent, Studiot! You guys on this forum have reawakened my interest in electronics. I see now that audio amps and speakers and such are far more sophisticated a topic than I had thought previously.

I ran into one poster on the thread you linked me to that brought up a good question that I still wonder about. How do you model a speaker when you're designing an amplifier and don't want to hear the noise? I used an 8 ohm resistor (made up of a bunch of resistors in order to provide the power), but when I hooked up a real speaker I got a 3.7 MHz sinewave that mixed with the input and created noise. A poster showed me how to make a circuit (zobel) that got rid of it. But how do you model a speaker without having to hear it?

 

Excellent, Studiot! You guys on this forum have reawakened my interest in electronics. I see now that audio amps and speakers and such are far more sophisticated a topic than I had thought previously.

I ran into one poster on the thread you linked me to that brought up a good question that I still wonder about. How do you model a speaker when you're designing an amplifier and don't want to hear the noise? I used an 8 ohm resistor (made up of a bunch of resistors in order to provide the power), but when I hooked up a real speaker I got a 3.7 MHz sinewave that mixed with the input and created noise. A poster showed me how to make a circuit (zobel) that got rid of it. But how do you model a speaker without having to hear it?

A 3.7 MHz sine wave on the output of an audio amplifier is an oscillation caused by a defective amplifier. Fix that and you'll fix a lot of things. A Zobel is a circuit used to cancel out speaker inductance at high frequencies...it's NOT there to cure an unstable amplifier!

Eric

 

A Zobel is a circuit used to cancel out speaker inductance at high frequencies...it's NOT there to cure an unstable amplifier!

Not in a good amp for sure, but they're in the minority, and besides it's impossible to have a zobel network that works for all speaker loads. They're always a best guesstimate unless they're designed to match a specific speaker.

It's well documented that certain amp and speaker model combinations don't work well, and that's because the awkward complex load of the speaker finds the Achilles' heel in the amps stability margins. All these amps have zobel networks. KL7AJ has a very valid point, though to be fair to the amps it's very difficult to achieve intrisic stability under all likely load conditions without leaning on the get-out-of-jail-free card that is the zobel network.

 

I thought we already offered to help with controlling the underlying stability of your amp in the other thread, but you went with the Zobel.

Post some circuit details if you want to look deeper.