TS Parameter changes during break-in

[DeZZar]

So I have no doubt this topic has been discussed many times here but given I am about to run this comparison for myself thought I would share!

Re: How does a driver measure vs manufacturer specs and what effect does "break-in" have?

Grabbed myself a fresh set of drivers for my current build and thought I would measure them straight out of the box and then let them play for a while and see what changes could be observed, and in the case of the woofer, how do these changes compare to the modelled response.

Here is a table of the measured params straight out of the box vs the manufacturer specs (measured with DatsV3)


The key electrical params seem to be the ones to suffer the most right out the box so we'll see how these change over time. Plan is to run the woofers at something like 30Hz at a reasonable volume for the next 24 hours - allow them to cool completely and then measure again. Rinse and repeat for 48 hours.

I wont run the tweeter right now as I need to find a solution for letting it play for this long without a) driving me insane and b) having all the dogs in the neighbourhood howling...some sort of "soundproof box" is needed I think...

I don't know if this is a complete fluke or not - but it turns out the woofer models just as good with the above measured params as it does with the manufacturer specs despite the significant differences!
Pink line below is the brand new driver vs the blue which is based on advertised params. I think I actually prefer the new one! Ha.


So I'll set the woofers playing and report back as I go.

[fpitas]

Yes, sometimes there's not a huge difference practically speaking. But every driver is different, since some suspensions are very "tight" from the factory, but some aren't. With regular music it may take a while to break in fully. Ken Kantor at Peerless told me he used 1/3 octave noise centered at Fs, at a high enough level to get the cone moving almost Xmax.

[DeZZar]

Nice - did Ken mention anything about how long he would run driver?

[fpitas]

Wow that convo was years ago. I *think* it was a couple hours. Of course, he knew exactly what his drivers needed to break them in. Other drivers may need more or less time.

[billfitzmaurice]

The main source of parameter change is softening of the spider and surround. If they're soft to begin with you won't see as much change as you will if they're stiff. Stiffer spiders and surrounds tend to be found on pro-sound drivers. I break my drivers in overnight, with a sine wave signal below Fs for maximum movement with minimal voltage.

[fpitas]

Yes, and some materials "re-stiffen" when they sit motionless for any time. Luckily, once broken in, they rapidly soften again in a few minutes playing.

[billfitzmaurice]

Yes, they do, so I never take a measurement with a 'cold' woofer. I often see a 10% drop in Fs after break in, with it rising a few Hz again after sitting for more than a few hours. The next time I take a measurement just a few minutes of pink noise at a few volts first is usually all it takes to bring it back down again.

[DeZZar]

I break my drivers in overnight, with a sine wave signal below Fs for maximum movement with minimal voltage.

I'll try that. I've never specifically recorded before and after or paid much attention to the difference. When these new drivers showed up I thought - I should really measure before and after to quantify the difference - if any - with these drivers at least.

[billfitzmaurice]

I always measure them when I get them to weed out any obviously defective units, then break them in overnight to be sure they come to spec.

[Chris Roemer]

I've got a general ROT I use to "model" a woofer's F3 (when/if I see a new one and WinISD isn't handy).
In a (near-"optimal") vented box, a driver's F3 can typically match its Fs in a box the same size as its Vas, IF the driver(woofer) has a Qts very near 0.40.
(Ex: an 8" w/a Vas of 1.5cf, an Fs of 30Hz, and a Qts of 0.40, will (more often than not) be able to hit 30Hz in a vented box that's 1.5cf.)

In GENERAL, for every 0.01 a driver's Qts is above 0.40, its F3 will drop 1Hz, and the box size will increase about +6% (and vice-versa).
(Ex: a 6" w/a Vas of 0.5cf, an Fs of 45Hz, and a Qts of 0.45 - most likely can achieve an F3 near 40Hz in a vented box close to 0.65cf.)

Using your 180s as an example:
pre breakin: Qts 0.49, Fs = 46, Vas = 18.7L - I'll say it can do 37Hz(46-9) in about a 29L(18.7*1.54) vented box, or approx 1.0cf
mfr "spec" (and possibly post breakin): Qts 0.39, Fs = 38, Vas=27.3 - I'll guess it'll do 39Hz(38+1) in about a 26L(27.3-6%) Vb, close to 0.9cf.

As a driver breaks in (as the spider(mainly) "softens" - usually by the "glue" cracking), more often than not you'll see the Fs drop, the Qts will drop, and the Vas will rise, and in many/most instances, the box model will not significantly change. When I get new woofers/mids (not tweeters - obviously), I typically break them in "by hand" CAREFULLY pushing them in/out (w/my fingers) to near Xmax several times each way. I can't say that their parms change any more (significantly) after that.

[DeZZar]

Quick update - I decided to re-measure after just one hour as I was curious to establish at what point the greatest change is measured. Here's the results. Driver running @ 30hz just shy of x-max for 1 hour.



Fs and Qms stand out as the ones that have gone through the greatest change in the first hour.

Mms I'm not so sure about - I thought this was essentially a static value? I don't think the weight of the air (atmospheric pressure) in my house has changed that much!

[Steve Lee]

Keep posting your results, DeZZar.

I am very interested in your findings.

My gut tells me that breaking glued physical component connections by way of brute forced hand pushing of the membranes to break them in is is not optimal and will lead to higher distortion levels over time than letting drivers break-in (become more elastic) through normal exercise via music/sound reproduction but, Hey - what do I know about anything?

Keep going . . .

[Steve Lee]

I've got a general ROT I use to "model" a woofer's F3 (when/if I see a new one and WinISD isn't handy).
In a (near-"optimal") vented box, a driver's F3 can typically match its Fs in a box the same size as its Vas, IF the driver(woofer) has a Qts very near 0.40.
(Ex: an 8" w/a Vas of 1.5cf, an Fs of 30Hz, and a Qts of 0.40, will (more often than not) be able to hit 30Hz in a vented box that's 1.5cf.)

In GENERAL, for every 0.01 a driver's Qts is above 0.40, its F3 will drop 1Hz, and the box size will increase about +6% (and vice-versa).
(Ex: a 6" w/a Vas of 0.5cf, an Fs of 45Hz, and a Qts of 0.45 - most likely can achieve an F3 near 40Hz in a vented box close to 0.65cf.)

Using your 180s as an example:
pre breakin: Qts 0.49, Fs = 46, Vas = 18.7L - I'll say it can do 37Hz(46-9) in about a 29L(18.7*1.54) vented box, or approx 1.0cf
mfr "spec" (and possibly post breakin): Qts 0.39, Fs = 38, Vas=27.3 - I'll guess it'll do 39Hz(38+1) in about a 26L(27.3-6%) Vb, close to 0.9cf.

As a driver breaks in (as the spider(mainly) "softens" - usually by the "glue" cracking), more often than not you'll see the Fs drop, the Qts will drop, and the Vas will rise, and in many/most instances, the box model will not significantly change. When I get new woofers/mids (not tweeters - obviously), I typically break them in "by hand" CAREFULLY pushing them in/out (w/my fingers) to near Xmax several times each way. I can't say that their parms change any more (significantly) after that.

I love your analysis as a ROT, Chris.

It would be really reassuring if your calcs and DeZZar's measurements hold true to one another given the significant differences in your collective methodologies . . .

[fpitas]

I love your analysis as a ROT, Chris.

It would be really reassuring if your calcs and DeZZar's measurements hold true to one another given the significant differences in your collective methodologies . . .

Unless you're in a hurry, or trying to QA and ship a lot of drivers, a few days of normal use will normally break in a driver sufficiently. And as DeZZar showed above, it usually doesn't make a big difference in the response.

[billfitzmaurice]

True. I break mine in because they're being used in design prototypes, so I have to be sure that they reach the advertised specs after break in. It doesn't happen often but I've had drivers that didn't even come close to what they should have, and didn't function as they should have either.

[1100xxben]

If you're trying to compare apples to apples with TS parameters, then you need to make sure the coil/motor cool back down to the same temperature before measuring them again. A good measure of your test setup reliability/repeatability should be to monitor Re, as that should never change (assuming the coil temperature is the same). You measured an increase of 9.54%, which means the coil temperature was up about 24 degrees C.

That being said... I also wouldn't trust the Re measurement from any typical system that is measuring the impedance sweep such as DATS or a jig with REW/LIMP, etc. It is estimating the DCR, and most systems use a 2-wire method. The better way to get highly accurate DCR measurements is to use a 4-wire resistance meter.

Also, as you have seen and others have commented, the LF response of the system is not dependent on one single parameter, but the full TS parameter set. As the suspension softens up with use, the parameters will all shift and the response typically does not change significantly.

[fpitas]

11 posts and no one has mentioned capacitor or inductor break in? We need better trolls here.

[PWR RYD]

Don't forget your resistors too.

[fpitas]

Resistors sound warmer after breaking in. But cables are night and day!

[fpitas]

I bet the current through the resistor smooths off those sharp edges you can see in the schematic symbol.

[DeZZar]

...you need to make sure the coil/motor cool back down to the same temperature before measuring them again...

You are absolutely right - I left it cool for just half an hour which is quite likely not long enough even though its winter down here!


So here we are now at a total of 7 hours. The driver ran for 6 hours straight near xmax again today with the following results after cooling down for a couple of hours.


The key Q params are getting much closer now with MMS and VAS still being a fair margin off spec. Vas seems to have gone back the other way although by a pretty inconsequential amount that is probably down to DATS error or a slightly different placement of the added mass. (I'm using a series of nuts and blu-tac around the cone as close to centre as I can get them - not really any alternative for these as the dust cap is fixed, so cant place anything exactly centre).

In terms of modelling in the box designed for this driver....this where we are. Starting to get a bit of a rise +1.5db around 40Hz which is wrecking my group delay but power handling is slightly better!


Will report back again after the next 6 hours!

[billfitzmaurice]

Pertinent to the topic: https://www.gr-research.com/driver-burn-in.html
https://www.gr-research.com/burn-in-myths.html

[fpitas]

That's an interesting article, Bill. I didn't realize how long it took to break in some drivers. I wonder if compression drivers have a break-in period.

[fpitas]

You might be able to "de-hump" it by tuning the port slightly lower.

[DeZZar]

Yeah my real world tuning is closer to 37-38 Hz instead of 39 so it will reduce it a little if this is how it turns out in the end.

[billfitzmaurice]

The hump probably means the box is too small. Don't sweat group delay. In order for it to be audible the speaker would have to be so bad that you wouldn't want to listen to it anyway.