Parts Express Customer Service Tells me to "Burn-in" Out-of-spec Drivers

[Phenoez8]

Recently I bought some Dayton Audio ND64 and ND65 speakers for a project, and by the time I got them, they measured out of spec. Many of the T/S parameters were off. I contacted Parts Express customer support and they said I should measure them only after 48 hours of "burn-in". I gave them the benefit of the doubt and here is what happened after.

Here are the datasheets of the speakers in question. I bought the ND64-4 and ND65-8 models.

Looks really good, great for a small enclosure with passive radiator design with those low Qts values.

And here is what I got when I measured them, even after a hundred hours of burn-in.

Not exactly the speaker they described. The ND64 was still useable, and the ND65 was no longer suited for the original design plan of a reflex/passive radiator box. What's your experience with burning in speakers? Would you still trust Dayton Audio? They sure have nice datasheets for all of their products, but this time it was a letdown.

 

[Phenoez8]

To be fair, Parts Express customer service did offer me a return even after hundreds of hours of burn-in tests, so getting my money back was not a problem. I only hope they can bring out more products that can stand up to inspection in the future. They did not offer an explanation on why these parts were out-of-spec.

 

[mhardy6647]

I suspect QC of the drivers is the real culprit here.
I've been fiddling with loudspeaker drivers in a post Thiele-Small sense for about a quarter of a century. Common wisdom, especially with off-brands (e.g., Parts Express's catch-all "Dayton") has always been measure actual parameters and work with those, AFAIK.

 

[StigErik]

What kind of signal did you use for "burn-in" ? Type of signal, and how many Volts?

 

[SIY]

I suspect QC of the drivers is the real culprit here.

Maybe. OTOH, I have consistently seen drivers from very high end makers (e.g., Dynaudio) show exactly the same problem- much higher fs and Qts than the datasheet. And consistently, over multiple batches. So it seems to be a systematic issue.

 

[mhardy6647]

Maybe. OTOH, I have consistently seen drivers from very high end makers (e.g., Dynaudio) show exactly the same problem- much higher fs and Qts than the datasheet. And consistently, over multiple batches. So it seems to be a systematic issue.

That is interesting.
Maybe it's the Big Driver cartel, then?
(not being entirely facetious)

EDIT: In the context of this thread, though, I guess I also have to ask -- were those overages independent of burn-in time?

 

[MAB]

Wow, those aren't even close to spec. And dramatically change, to be clear >10% change in f(s) is not normal. Except...
I measured break-in on HiVi B3N, and saw large changes compared to drivers I normally measure. I have many of these left over from an old line-array project. I stopped using them due to high variability and overall quality issues, for instance about 10% of my samples had coil-rub, I returned them and saw the same issues on the replacements.

The driver's spec sheet says f(s) = 77Hz, I measured ~108Hz on 4 new drivers. Even after the drivers dramatically change after break in to 98Hz, is still 27% higher than spec. Aside from the driver to driver variability and overall QC issues, these drivers are not even close to advertised specs. I also had to change plans after measuring.

 

[Phenoez8]

What kind of signal did you use for "burn-in" ? Type of signal, and how many Volts?

I used pink noise low passed at 150Hz to generate plenty of excursion without overheating the voice coil. The signal level was not measured by adjusted by eye and by ear. I turned it to a point where the cone was moving about 1 to 2mm each way.

 

[Phenoez8]

Wow, those aren't even close to spec. And dramatically change, to be clear >10% change in f(s) is not normal. Except...
I measured break-in on HiVi B3N, and saw large changes compared to drivers I normally measure. I have many of these left over from an old line-array project. I stopped using them due to high variability and overall quality issues, for instance about 10% of my samples had coil-rub, I returned them and saw the same issues on the replacements.

The driver's spec sheet says f(s) = 77Hz, I measured ~108Hz on 4 new drivers. Even after the drivers dramatically change after break in to 98Hz, is still 27% higher than spec. Aside from the driver to driver variability and overall QC issues, these drivers are not even close to advertised specs. I also had to change plans after measuring.

This is really good experiment. Thank you for sharing. I also found this Klippel paper that explains some of the speaker aging behavior. https://www.klippel.de/fileadmin/kl...s/Aging of loudspeaker suspension_Klippel.pdf

 

[StigErik]

I used pink noise low passed at 150Hz to generate plenty of excursion without overheating the voice coil. The signal level was not measured by adjusted by eye and by ear. I turned it to a point where the cone was moving about 1 to 2mm each way.

I’ve heard from the SEAS factory that they recommend using 25 Hz sinewave and turn it up to get excursion around the max mechanical limit. That doesn’t require a lot of power! After a minute or so, the driver parameters should be 90% settled.

 

[somebodyelse]

In another thread there are reports of the SIG180 parameters being somewhat different from spec, with similar numbers reported by different people.
https://www.audiosciencereview.com/...s/budget-directiva-project.52024/post-1881130

 

[SIY]

I guess I also have to ask -- were those overages independent of burn-in time?

I could get the fs a few Hz lower with exercising the drivers, but never anywhere close to the datasheet values.

Cynically, I would note that the error is always in the direction of weaker magnet.

 

[Ze Frog]

Driver's do change parameters slightly upon a limited break in. Before designing a crossover I usually leave them on a test bench for 24 hours. If they aren't wildly different then it's worth a go, if they are different by a big amount then likely they will still be off. Most times I build D.I.Y I would buy multiples of for that very reason and then match them to pairs.

 

[Ze Frog]

Wow, those aren't even close to spec. And dramatically change, to be clear >10% change in f(s) is not normal. Except...
I measured break-in on HiVi B3N, and saw large changes compared to drivers I normally measure. I have many of these left over from an old line-array project. I stopped using them due to high variability and overall quality issues, for instance about 10% of my samples had coil-rub, I returned them and saw the same issues on the replacements.

The driver's spec sheet says f(s) = 77Hz, I measured ~108Hz on 4 new drivers. Even after the drivers dramatically change after break in to 98Hz, is still 27% higher than spec. Aside from the driver to driver variability and overall QC issues, these drivers are not even close to advertised specs. I also had to change plans after measuring.

Yeah, it's been like this almost forever, very rare in my experience to find a driver completely on spec regardless of price. Incidentally though, some of the most accurate to claimed I came across were actually cheap things, Visaton WS 25 E's. Not really much good for anything beyond open baffle or ripole type use though. Still, was kinda strange how they were so much closer to spec considering the cheapness.