Breaking in speakers
- Thread starter Intrepid
- Start date
Register to hide this ad
Intrepid said:
Breaking in boils down to playing them for a while (exact duration unspecified) before listening too critically to reach a premature conclusion about whether or not you like the way they sound.
The basic premise is that the moving parts of the driver need to be flexed in order to properly limber up and reveal their true character. Similarly, the electronics in the crossover may be affected by applying signal voltage.
Some folks believe that the break-in process can be accelerated by playing the speaker at louder levels, and/or with a wider range of frequencies.
Break-in effect would likely be most apparent with brand new speakers, or speakers that have had parts replaced or repaired (such as new capacitors, surrounds refoamed, etc...)
There is also another school of thought that believes the break-in effect is mostly psycho-acoustic...meaning that your ears and brain need time to adjust to the sound of a new speaker.
Last edited:
hpsenicka said:
I had always been of this school until I read an article in a recent (maybe last month's?) issue of Affordable Audio. In it, they were measuring the various response levels of a driver at different points in a break-in cycle. Sure enough, as the driver was played for longer periods of time, the measured characteristics of the driver changed a meaningful amount. Pretty interesting article, you might go and download it and check it out.
David
I "got to know" what break in or "use" could mean recently by measuring the Thiele Small parameters of speakers, something I had never done before.
Identical used speakers measure very differently in key parameters and if I am to believe the current theories for speakers these drivers would give vastly different results when used in cabinets.
The results were so disturbing to me (since I suspected errors in my set up) that I did a bunch over again and left one speaker hooked up to my set up for the better part of the day. Its published resonance is 200Hz.
I measured repeatedly 270Hz initially which decreased to around 250Hz after several hours of "work".
This, I think, could be explained by a limbering up of the speaker cone support.
The next day the speaker started out again at around 275Hz!!
So they do change thru use. I don't know if they all change this much, it seems it would make speaker design a total crapshot, but some people say that's what it is exactly.
Uwe
Identical used speakers measure very differently in key parameters and if I am to believe the current theories for speakers these drivers would give vastly different results when used in cabinets.
The results were so disturbing to me (since I suspected errors in my set up) that I did a bunch over again and left one speaker hooked up to my set up for the better part of the day. Its published resonance is 200Hz.
I measured repeatedly 270Hz initially which decreased to around 250Hz after several hours of "work".
This, I think, could be explained by a limbering up of the speaker cone support.
The next day the speaker started out again at around 275Hz!!
So they do change thru use. I don't know if they all change this much, it seems it would make speaker design a total crapshot, but some people say that's what it is exactly.
Uwe
With break in periods as long as you mention isn't it wrong to design enclosures with the specs of new speakers.
I wondered about this very problem in another post when I found that the speaker parameters, at least as far as I could determine them, were moving targets at best.
Uwe
I wondered about this very problem in another post when I found that the speaker parameters, at least as far as I could determine them, were moving targets at best.
Uwe
ken kantor
Fowl Humor
Othello,
This is a can of worms far larger than most hobbyists, (or pros, for that matter), want to digest. Several large speaker manufacturers, Tymphany included, have mathematicians and physicists pretty much dedicated to studying it full time. It is a far from simple problem.
First of all, one must remember that Thiele-Small parameters are only characteristics of an electromechanical MODEL of a loudspeaker. For the most part, they are not actual, physical properties of the driver itself. This is a tricky concept to grasp fully, but it is fundamental. Some parameters have stable and clear physical meanings: Mmd, for example. Others, such as Vas, are pure abstractions. They have a definition, but not a direct congruency. Still other parameters, such as Fs and Qts are in a gray area. They seem to have an intuitive meaning, but they can be mathematically defined in several ways, all correct and all yielding slightly different results. Again, this is a tricky notion to really internalize, but it is true.
Thiele-Small analysis only has real meaning in a linear system, which a speaker is not. To further complicate matters, even after deciding how to exactly define these parameters, measured and derived values will be affected by:
1- The short-term output history of the driver.
2- The long-term output history of the driver.
3- The overall temperature of the driver.
4- The temperature gradient within the driver.
5- Orientation of the driver with respect to gravitational forces.
6- The type of drive stimulus used.
7- The level of the drive stimulus used.
8- Drive source impedance characteristics.
9- Etc.
(As an aside, when I first studied loudspeaker physics in college, we did not use modelling. We used the actual, physical parameters of the speaker to calculate from. This meant we had to literally tear apart a sample of a driver to understand it. We had to put it in a vacuum chamber to calculate air load. Use force gauges and micrometers to measure compliance and BL. Cut off the diaphragm to weigh it, etc. After this, each enclosure design took days or weeks of calculation.)
OK, so where does this leave us? Well, no doubt, Thiele-Small modelling methods are a very valuable to loudspeaker designers, and they can yield useful results. A good enclosure design includes a sensitivity analysis (perturbation, not SPL), to assure that some variation in parameters is tolerable. Of course, reputable driver suppliers publish specs after break-in, and measured under conditions that anticipate actual use.
The ideal thing for you to do is build a simple test box. After you measure its parameters, put it in the test box and measure the response to see if it meets your predications. This test box doesn't have to be fancy, or even exactly the same type as your final design. It only needs to be reasonable for the driver. For example, a sealed box of 50 or 100 liters can check out just about any 12" woofer. If the predictions match, you can proceed with your final design. If they don't, you need to refine your parameter test methods until they do.
Other than this, try testing the driver under different conditions; different drive levels, orientations, amount of burn-in, etc. Look for trends and averages in the data. Clamp the driver as firmly as possible to a stable test jig. Avoid room noise and footstep vibrations. I can elaborate on the test methods I find the most useful, if there is interest, and at an earlier hour....
Hope this helps!
-k
http://kkantor.spaces.live.com
www.tymphany.com
This is a can of worms far larger than most hobbyists, (or pros, for that matter), want to digest. Several large speaker manufacturers, Tymphany included, have mathematicians and physicists pretty much dedicated to studying it full time. It is a far from simple problem.
First of all, one must remember that Thiele-Small parameters are only characteristics of an electromechanical MODEL of a loudspeaker. For the most part, they are not actual, physical properties of the driver itself. This is a tricky concept to grasp fully, but it is fundamental. Some parameters have stable and clear physical meanings: Mmd, for example. Others, such as Vas, are pure abstractions. They have a definition, but not a direct congruency. Still other parameters, such as Fs and Qts are in a gray area. They seem to have an intuitive meaning, but they can be mathematically defined in several ways, all correct and all yielding slightly different results. Again, this is a tricky notion to really internalize, but it is true.
Thiele-Small analysis only has real meaning in a linear system, which a speaker is not. To further complicate matters, even after deciding how to exactly define these parameters, measured and derived values will be affected by:
1- The short-term output history of the driver.
2- The long-term output history of the driver.
3- The overall temperature of the driver.
4- The temperature gradient within the driver.
5- Orientation of the driver with respect to gravitational forces.
6- The type of drive stimulus used.
7- The level of the drive stimulus used.
8- Drive source impedance characteristics.
9- Etc.
(As an aside, when I first studied loudspeaker physics in college, we did not use modelling. We used the actual, physical parameters of the speaker to calculate from. This meant we had to literally tear apart a sample of a driver to understand it. We had to put it in a vacuum chamber to calculate air load. Use force gauges and micrometers to measure compliance and BL. Cut off the diaphragm to weigh it, etc. After this, each enclosure design took days or weeks of calculation.)
OK, so where does this leave us? Well, no doubt, Thiele-Small modelling methods are a very valuable to loudspeaker designers, and they can yield useful results. A good enclosure design includes a sensitivity analysis (perturbation, not SPL), to assure that some variation in parameters is tolerable. Of course, reputable driver suppliers publish specs after break-in, and measured under conditions that anticipate actual use.
The ideal thing for you to do is build a simple test box. After you measure its parameters, put it in the test box and measure the response to see if it meets your predications. This test box doesn't have to be fancy, or even exactly the same type as your final design. It only needs to be reasonable for the driver. For example, a sealed box of 50 or 100 liters can check out just about any 12" woofer. If the predictions match, you can proceed with your final design. If they don't, you need to refine your parameter test methods until they do.
Other than this, try testing the driver under different conditions; different drive levels, orientations, amount of burn-in, etc. Look for trends and averages in the data. Clamp the driver as firmly as possible to a stable test jig. Avoid room noise and footstep vibrations. I can elaborate on the test methods I find the most useful, if there is interest, and at an earlier hour....
Hope this helps!
-k
http://kkantor.spaces.live.com
www.tymphany.com
Ken,
Thank you for your wide ranging response.
I come to this quest by wanting to build a better pair of horn loudspeakers, reading Edgar's approach and now working thru d'Appolito's book on speaker testing.
To start the design process the knowledge of the relevant Thiele Small parameters is highly encouraged, even to several digits after the decimal point.
Now I realize the variability that you spelled out in your post, that's why I try to keep everything fairly constant as I measure and found it instructive to hold a newspaper about 2 feet away from the DUT, the resonance point shifted immediately, this is sensitive stuff.
But then it dawns on you that if the parameters shift that readily, what does that mean for its final performance which will probably be under quite varying conditions, in spite of the fact that you tried to determine voice coil resistance within a small fraction of an ohm under lab conditions.
I guess the question is, how well does this model represent real working conditions.
I like your suggestion to make your predictions based on the model and test them in a small sample box. This would probably not be a horn, but it would give you some idea about the validity of the predictions and how those predictions hold up when the working conditions start to change.
None of the authors I read sofar has dealt with this.
So yes, if you have any pointers to procedures and or articles written about this issue, I would appraciate hearing about it.
Uwe
Thank you for your wide ranging response.
I come to this quest by wanting to build a better pair of horn loudspeakers, reading Edgar's approach and now working thru d'Appolito's book on speaker testing.
To start the design process the knowledge of the relevant Thiele Small parameters is highly encouraged, even to several digits after the decimal point.
Now I realize the variability that you spelled out in your post, that's why I try to keep everything fairly constant as I measure and found it instructive to hold a newspaper about 2 feet away from the DUT, the resonance point shifted immediately, this is sensitive stuff.
But then it dawns on you that if the parameters shift that readily, what does that mean for its final performance which will probably be under quite varying conditions, in spite of the fact that you tried to determine voice coil resistance within a small fraction of an ohm under lab conditions.
I guess the question is, how well does this model represent real working conditions.
I like your suggestion to make your predictions based on the model and test them in a small sample box. This would probably not be a horn, but it would give you some idea about the validity of the predictions and how those predictions hold up when the working conditions start to change.
None of the authors I read sofar has dealt with this.
So yes, if you have any pointers to procedures and or articles written about this issue, I would appraciate hearing about it.
Uwe
jesse111
Active Member
Clmrt said:
Do you have a link to this information?
It's been suggested by some (the B&W manual for instance) that break is simply a matter of temperature and humidity acclimation of new speakers to your environment which can take a day or two. Some also feel it's a matter of your ears adjusting to the new sound of your gear.
soundmotor
super modified
I've noticed minimal subjective improvements to tweeters & midranges as far as break-in goes but it could just be that any changes in their performance were outside of my range of hearing. Subwoofers however is another story. I have heard dramatic changes in bass response in as little as 30 minutes of use.
ken kantor
Fowl Humor
Othello,
I'll write some more about the relationship between the model and the real world sometime in the next few days. I also vaguely remember doing a two-part article in AUDIO magazine on the subject about 20 years ago, that I will try to dig up and scan.
In a nutshell: some types of designs are less sensitive to real-world parameter shifts than others are. Likewise, a shift in some parameters has less sonic impact than others. Lastly, as someone posted earlier in this thread, certain parameters tend to be coupled, so that when one moves, the other moves, too. In certain cases, the net result can be less-than-expected change in the response.
-k
http//kkantor.spaces.live.com
www.tymphany.com
I'll write some more about the relationship between the model and the real world sometime in the next few days. I also vaguely remember doing a two-part article in AUDIO magazine on the subject about 20 years ago, that I will try to dig up and scan.
In a nutshell: some types of designs are less sensitive to real-world parameter shifts than others are. Likewise, a shift in some parameters has less sonic impact than others. Lastly, as someone posted earlier in this thread, certain parameters tend to be coupled, so that when one moves, the other moves, too. In certain cases, the net result can be less-than-expected change in the response.
-k
http//kkantor.spaces.live.com
www.tymphany.com
![JimJ[VT]](/forums/data/avatars/s/20/20031.jpg?1443894878){kind=avatar}
