Driver Break in - Fact or Fiction?

[phaseshift]

I have noticed a debate about speaker break in on this forum (Breaking in a speaker thread) and other forums. As someone that has been around this for many years, particularly at the factory level, I would like to offer my observations and some facts to clear the air.

I think where the myth of break-in occurs is at a factory production level.

Most speaker drivers are tested to verify their T-S parameters and frequency response before assembly. This requires some testing at the factory. Plus, once a speaker is built, there is usually a test period to verify the crossover and drivers are all working as planned. Granted, this is not a full 40 hour burn in, but the drivers have at least seen some use before going to the consumer.

It is true that the drivers do se some testing prior to final pack out, but it is not really accurate to say that this breaks them in.

Generally, a driver will get a sweep test a couple of times during it’s life and nowadays, a response check as well. Sweep voltage depends on the customer or manufacturer’s preference, but you would find a lot of drivers swept at about 4 to 5 volts from their target resonance to a bit higher up to around 5K. Tweeters a little different, but again, a reasonable range; mainly looking for loose particles or a RVC condition (voice coil rub). This sweep will typically be over in 3 or 4 seconds max- that is it. You would again have this sweep test as the driver is installed into the system and again at the system level. The amount of change you see in the suspension is minimal- there are just not enough cycles or at a high enough power level to really impact the stiffness of the damper or the compliance of the surround.

Testing – for instance a MLS or other stimulus – response test is also at a very low level and is completed in a few seconds- again, you just do not have the power or duration there to impact the compliance of the driver. When it comes to testing tweeters, you will often see differences as well- often due to dispersion of the ferrofluid in the gap. I have done a very extensive study on this as a QA – manufacturing project for a high end and very well known OEM who we supply automation equipment to and the ferrofluid balance was the biggest factor for changes in the tweeter over the first few hours of it’s life. (FYI- When I proposed this concept before the study was done- I even thought it to be an absurd notion but the facts were that it was actually the thing that caused a lot of problems at the manufacturing level.) For drivers that do actually get a parameter test run on them before shipment (very uncommon except among more expensive subwoofer drivers), the signal level is very small and the timeframe is very short. There are a couple of platforms out there that will do rapid T/S parameter testing using a fixed Mmd based calculation. MLSSA, Clio QC and I believe Sound Check will do that sort of testing. You can use a Klippel to do high level testing, but this takes quite a while and is certainly not a production level test for any driver that I know of.

For those that do not believe in the break in time for a driver, here is a way to prove it to yourself once and for all…

Break in is very real and anyone with basic testing skills and equipment can verify that it does happen. How? First of all, you can look at the driver Fo – right out of the box. Measure this about 10 times and then average the number. Now put a signal through it at moderate levels for about an hour. Let the speaker cool down for one hour. Measure the Fo again, get the average and compare. Want more proof? Go for the full Monte T/S parameter testing using a similar routine… Multiple tests, average the numbers and then run a signal on the driver for an hour or so. Cool down time, re-test, re-average and compare. Still not convinced? Set up the driver in an anechoic environment, lather, rinse, repeat as before and observe the differences in the response, particularly near resonance.

What breaks in?? Mostly the damper, but to some degree, the surround as well. In general, the surround has a far smaller impact on the mechanical model of the speaker, but it is there a little bit. Mostly look at the damper and how much it changes over the first few hundred cycles at or near Xmech and resonance. There may be other factors as well, but they are going to be minor in comparison to the damper compliance change and the overall changes in the mechanical model of the speaker as related to that. Obviously, in a design that does not have a damper, the changes you observe will be different. Tweeters and other single suspension devices typically operate at very low excursion, higher frequencies where the mass and shape are the bigger factors.

Electrical break in- I have also looked at electrical break in of drivers and I will say that it ranged from minimal to a non-issue on a typical driver. There are slight changes in the coil after heating and the magnets can loose some of their zip if overheated; especially cheap-o and low Y factor stuff (it is my experience that is true anyhow…). This is mainly a factor in the cheaper stuff that most of you would not have any desire to fool with, but is all over the word, particularly in LCD / Plasma TV sets and other consumer goods where the heat generated by the driver is not the big issue- the device that the driver is located in is the overriding factor. You would also see differences in the planar drivers such as the Neo 3, but I would expect them to be absolutely minimal at the electrical level unless the driver was abused or overheated.

Others here may have done studies related to the electrical changes in coil drive devices (drivers and tweeters) and they may have different experiences or comments. Particularly in a case like a large, high power sub with a multi-layer coil and enough Bl to bend space a little bit. I have not had the opportunity to delve into that sort of driver on a study like this, but it would be interesting to hear form some that have. Manville from JL is probably the most likely candidate that I know of on here- they have a lot of really heavy duty drivers and I believe that JL actually puts a lot of brain power into their products. It would be interesting to hear from them on this topic.


-------------------

Edit - Adding comments on materials 9-28-07 –


When you look at damper compliance or stiffness, it is usually measured at very small levels only. Typically, an engineer would look at the deflection at a certain weight; say a few grams. Obviously, this will depend on the application- an earth-buster woofer that moves 35mm is going to have a completely different overall compliance model than a 5.25” midrange. The metrics are the same, but the amounts are different.

I have seen the discussion lead to a question- What damper material is best? What speaker has the least break in shift?

Well, that depends on what you are looking for… Best lifespan, best cost for size, flammability, stability, adhesive bonding… there is a long list to consider.

Here are 3 popular materials and within them subcategories which tweak or add options to the particular variety, but for simplicity, lets stick to the 3 basic materials used in hi-fi drivers. This is my experience as an engineer – some can be proven by science if you want to delve into that; most is based on experience with design and development of hundreds of drivers and systems.

1. Cotton – This is the old stand by and is used in a lot of lower power drivers. It is light weight and forms easily and predictably when laced with the phenol or other bonding agents. The benefits are that it is certainly easy to bond to (adhesives hold your speakers together- not screws) and the performance is predictable. It is also cheap. Downsides include a high level of flammability and in a lot of opinions, moderate sonic performance. Not bad, but you can do better and the flammability is a big issue these days. Also, it is known to wear out over time. This is especially problematic on higher excursion drivers. It is seldom to see plain cotton dampers on drivers these days unless you are opening up your dead clock radio.
a. Poly-Cotton blend- Adds some poly to the cotton; you probably guessed that. My experience is that the sonic performance is better and the compliance model has better stability through the deflection. However it is still flammable and now costs more.
b. Treated cotton- can be a host of treatments, but most are really there as a flame retardant. Well, that is my experience anyhow. Note that with the treatments, you tend to have bonding issues which is a big problem.
2. Conex – While I have used a lot of this stuff, I am not sure if Conex is a trademarked name or if it is just what everyone calls this material. Basically it is a mix of Nomex and cotton. ON the plus side, you have a lot better stability in your overall compliance model and it is partially fire retardant. My experience is that the acoustics and control using this flavor of material is better, but mainly in a longer excursion driver. You also have to be more picky about your damper geometry and landings or else you can get into problems. Adhesive bonding is very good as long as you do not treat it with a flame retardant.
a. Varieties are mostly due to mix ratio- 50-50 seems to be very popular, but some like to use a higher Nomex percentage.
b. The cost is significantly more than a cotton damper, but not to bad. A lot of speakers in the consumer market have Conex dampers in them; many with flame retardant treatments to get them to self-extinguish.
3. Nomex- A synthetic material and works very well for dampers, but is tough on the budget. It is practicality fire proof in a speaker and have great forming and deflection properties.
a. My experience is that the sonics and predictability you get out of this material is about as good as anything out there and it is readily available, stable and easy to work with. Adhesives bond to it well, but I prefer epoxy which normally has a bit more time to wick in before going out of the green stage. Changes over the lifespan are the best of the lot with this material, but wil will wear out as any material will. However, I have run long term tests where the damper was the only difference on a driver and it lasted more than double that of a conex damped driver. Yes, damper failure was what we were looking for; the test was shaped tone burst on a batch of 10” woofers, nothing too special. Double the lifespan is significant.
b. The downside is cost- it is far more expensive than a similarly sized chunk of cotton or Conex material and costs do add up quickly. Also, some engineers do not like it as it is a synthetic material; I have head someone rant on about how it as not acoustically adequate and that a cotton / hemp blend was the ultimate because it is organic…. To each his own, but as a child of the microwave age, I am open minded to synthetics being OK and often better depending on the application.

Other factors related to the damper that will impact the break in and how much the driver changes during it’s lifespan.
1. Thread count – with all the above examples, you can have many different thread counts - High thread count dampers are ones where the thread is very fine and closely woven. Low thread count is the opposite. While you can use any thread count for your application, the tendency is that higher counts are in smaller drivers where the top end is more important and your excursion is limited. There are probably examples of this not being the case, but as a general rule, you will not find 60 thread dampers in a 15” car woofer.
2. Phenolic amount. Most dampers are impregnated with phenolic which gives them some of their stiffness and most of the formability. You can consider this to be long lasting starch on a shirt; just in this case, when you pleat it and iron the shirt, the pleat stays basically forever. As with most things in the engineering world, there is a range that works and outside of that you are either hurting yourself or wasting money or both.
3. Geometry- Is the damper a progressive roll format? Regressive roll format? Flat? Bumped? Triangular geometry? ??? These all change differently over time and again, are all part of the tradeoffs made in the design of a transducer. While none are perfect, there are certainly a lot of drivers using progressive roll geometry, particularly for woofers with a high Xmax / Xmech. The triangular (accordion) geometry is less popular on woofers and seldom seen on mids; it does wear out rather quickly as the bulk of the bending is done right at the top and bottom of the triangles. They simply wear through- that is bad. Oddly, some pro drivers use this very effectively; I suspect it works where the excursion is low and the thread count is high with a low phenolic level. Just a guess.

So, when you are looking at what can cause the differences in the speaker, it is not a matter of one material or one aspect of the design. Complex variables play a part in the changes and some drivers will have a large shift while others have a rather small shift. You would also need to look at this in terms of time- the driver that seems to have a smaller shift may actually have a very large shift over a longer period of time whereas a driver that seems to change a lot may do so in a short time and remain very stable for a long time.

When I or one of our engineers does a system design, I always make sure I am using driver that has been run in. We do this by looking at the application and running one of two procedures-
• For a woofer, we use a shaped tone burst with the center frequency similar to the driver Fo; typically for 2 to 4 hours depending on the design targets and driver design. Of course it will be measured before and after the TBT and the differences recorded. Some things change a lot; particularly related to the mechanicals- FO may drop as much as 10% and Cms may shift a similar amount depending on the design. Qt will change but not as significantly as it is a product of the mechanical and electrical models.
• For a mid or tweeter, we run EIA 426-B noise through the driver at rated power for an hour. Response and parameters are measured before and after. You see varying degrees of change depending on the driver design and parts, particularly the damper. IN general, we do not see Fo shifting down quite as much- depends on the factors mentioned here. Response will change at a couple of key points- one is near resonance and the others are where the damper or cone is going modal (out of piston mode). Here, the cone comes into play and yes, they do break in as well, but to a lesser degree and it is very dependant on material. An aluminum cone probably changes very little over it’s life whereas a paper cone with some silver coating (looks sort of like metal) will change a lot. Yes, you can see this on response curves, but you have to be careful to get your DUT and mic placement spot on.

One thing I would caution everyone about is that my break in routine for mids is not something that your wife or neighbors will like. We run them hard and hot to get the job done quickly and properly; it is LOUD. The TBT is a piece of cake as long at is free air- if you have it on the cabinet, your family will hate you. I say this because running at moderate levels will shift the performance less than you might expect in some cases. Actually, I would say more in the case of a midrange or system where the level you need to get the drivers really moving is so **** loud that you are worried about damage to your equipment or hearing. The woofers are easy to get going out to Xmax with a movie or rap / bass disc. If you want the shaped tone bursts, you can get them as well as the EIA 426B noise from ALMA International- there is a disc for sale on the website. You may also be able to do a search and find something close or the same.

[Geoff St. Germain]

Yeah, I've seen before and after measurements of LF drivers showing changes in Fs, Cms and Vas. I've seen measurements for the TC-3000 here and I saw measurements for a Peerless woofer on another site.

The only thing that gets me is breaking in cables.

[thxgoon]

Geoff St. Germain wrote:

The only thing that gets me is breaking in cables.

Only if they're properly isolated from the floor and made of pure silver whose molecules have all been oriented in the same direction lol!!

Great read phaseshift. Thanks!

[Sonnie]

Thanks for the detailed explanation Gui.

I copied this from the DIY Speakers forum and made it a sticky here as well.

[joetama]

Good read... Thanks...

[tonyvdb]

No mention of the person who originally brought this subject up (just kidding)... Great read phaseshift Thanks for the info.

[joetama]

tonyvdb wrote:

No mention of the person who originally brought this subject up (just kidding)... Great read phaseshift Thanks for the info.

I laughed...


I have a brand new driver sitting at home right now. I'm going to test it and spec it then run it and test it over time. I'll post my results here... Should be interesting...

[Sonnie]

tonyvdb wrote:

No mention of the person who originally brought this subject up (just kidding)... Great read phaseshift Thanks for the info.

I edited the first post and referenced your thread.

[JCD]

Something I came across..

Speaker Break In: Fact or Fiction

The quick and dirty -- speakers do change (e.g., the t/s parameters do change); however, the change is very subtle if not inaudible and happens very quickly.

Not endorsing the article, just putting it out there.

JCD