Standard for the measurement of a driver's FR

Ant_222

Member
2006-08-03 8:41 pm
Hello, all

What is the standard setup to measure the frequency response of a driver regardless of enclosure? In other words, what physical and electrical configuration does one have to employ in order to measure anything similar to what one sees in driver datasheets? A baffle of normative dimensions, I suppose?
 

Ant_222

Member
2006-08-03 8:41 pm
There isn't one. There are several IEC rated baffles &c., but there are no industry standards that everyone is obliged to adhere to, and measurement setups, equipment and conditions vary significantly between manufacturers, the majority of whom never state exactly what they used or did.
So driver manufacturers baffle their customers as much as they do their drivers. Another obscure matter is the signal source used to measure the T/S parameters: what output impedance, voltage, and current? In my experience, it has a strong effect on the measured impedance and therefore the Q-factor. For example, my Visaton AL-130 in an E. J. Jordan (PDF) aperiodic enclousure shows the desired impedance plateau from 31 to 54 Hz only at a specific input voltage of 0.73 V at the driver terminals measured at 440 Hz...
 
So driver manufacturers baffle their customers as much as they do their drivers.

Up to a point; depends how much their marketing department is involved, although that is a separate issue. As far as lack of an obbligatory industry standard is concerned, that isn't actually their fault. There never was one, nor is it likely to happen since it is not a closed industry with central management. IEC is as close as you'll get to it, and many measure on those baffles / under the stated conditions, but not all, it's not forced, and it doesn't cover everything anyway.

Another obscure matter is the signal source used to measure the T/S parameters: what output impedance, voltage, and current. In my experience, it has a strong effect on the measured impedance and therefore the Q-factor. For example, my Visaton AL-130 in an E. J. Jordan (PDF) aperiodic enclousure shows the desired impedance plateau from 31 to 54 Hz only at a specific input voltage of 0.73 V at the driver terminals measured at 440 Hz...

Yes it does. There is no industry standard, never has been and unfortunately never will be. There are values that are often applied e.g. normalised voltage 2.83v but low voltages are often used by small scale manufacturers who may not be able to afford the more expensive production testing equipment (or in some cases have much interest in doing so). This is also often compounded by what data manufacturers actually use in generating the published information: it might be an average of each parameter from a batch, an averaged data set to ensure consistancy within the supplied table, a set taken from a single random sample, an average of data from several batches... the list goes on.
 

Ant_222

Member
2006-08-03 8:41 pm
Up to a point; depends how much their marketing department is involved, although that is a separate issue. As far as lack of an obbligatory industry standard is concerned, that isn't actually their fault. There never was one, nor is it likely to happen since it is not a closed industry with central management.
I don't think a closed industry is a requirement. One purpose of standards is to facilitate mutual understanding between engineers by means of common points of reference.

I do not deplore the absence of one standard to rule them all, but I should appreciate if manufacturers published the conditions under which they measure their products, perhaps in small type if PDF paper is expensive. It is a matter of mere reproducibility—a fundamental principle in experimental discipline.
There are values that are often applied e.g. normalised voltage 2.83v
I see neither rhyme nor reason in normalising voltage, for drivers differ in power handling and sensitivity. One does not test with the same signal a sensitive paper driver intended for domestic use with a tube 3W SE amplifier and a tough rubber-suspension woofer designed for a movie theater sound system with a powerful transistor amplifier...
low voltages are often used by small scale manufacturers who may not be able to afford the more expensive production testing equipment (or in some cases have much interest in doing so).
Is that a problem? A decent Hi-Fi transistor amplifer should suffice as a signal source, because the inconsistency and inconstancy of driver parameters far outweighs any accuracy gain offered by dedicated equipment, does it not?
This is also often compounded by what data manufacturers actually use in generating the published information: it might be an average of each parameter from a batch, an averaged data set to ensure consistancy within the supplied table, a set taken from a single random sample, an average of data from several batches... the list goes on.

Agreed, but these discrepancies are of secondary magnitude and importance.
 

S4m

Member
2013-11-06 2:48 pm
Hannover
Hi Ant. You mentioned visaton. I don't find the information on their homepage but If i remember correctly they use a 1000l enclosure for measurements. If you Like to get the data from them use the tool BoxSim. All of their drivers are in a library and the data can be used for comparison
 

Ant_222

Member
2006-08-03 8:41 pm
Found in the Visaton forum: Messbedingungen bei Visaton - Visaton Diskussionsforum
The process and the environment is explained...in German. The previous mentioned 1000l are 500l in reality and they use different baffles/enclosure for different drivers eg tweeter in a Infinity baffle.

Thanks for the information, S4m. There seems little sense in testing low-Q woofers on baffles or in 500-liter boxes, though...
 
I don't think a closed industry is a requirement. One purpose of standards is to facilitate mutual understanding between engineers by means of common points of reference.

No, nobody said it was a 'requirement', and yes, that is one of the main purposes of common standards. However, the fact is, it's not going to happen because they're selling product, not discussing engineering features.

I do not deplore the absence of one standard to rule them all, but I should appreciate if manufacturers published the conditions under which they measure their products, perhaps in small type if PDF paper is expensive. It is a matter of mere reproducibility—a fundamental principle in experimental discipline.

You're speaking as though a commercial environment was an engineering laboratory.

Right: the truth. The majority of loudspeaker drive unit manufacturers have zero interest in doing so. It is not, as far as they are concerned, of much value to them under the current commercial conditions. It would take time and effort for little or no return, and most would rather their employees were engaged in other activities. There are exceptions, but not many. By and large, you would have to force them all, which is basically impossible.

I see neither rhyme nor reason in normalising voltage, for drivers differ in power handling and sensitivity.

Yes, I know. It was a single example that is sometimes done for certain uses; nobody was advocating it for some blanket industry standard.

One does not test with the same signal a sensitive paper driver intended for domestic use with a tube 3W SE amplifier and a tough rubber-suspension woofer designed for a movie theater sound system with a powerful transistor amplifier...Is that a problem? A decent Hi-Fi transistor amplifer should suffice as a signal source, because the inconsistency and inconstancy of driver parameters far outweighs any accuracy gain offered by dedicated equipment, does it not?

With a handful of exceptions, very possibly, but if you're going to invoke engineering consistency, you'd still have to standardise requirements for that as well. Good luck. ;)

The fact is, you are certainly not the first to complain about this, and you won't be the last. But however much you, I and a small number of other enthusiasts might complain about it (and it's perfectly true: the current status quo is not at all satisfactory), the majority of manufacturers couldn't care less about what we think, because they still sell product, and only a very small proportion of their buyers are interested in said data. Commercial prioirities take precidence. If you can make it a commercial priority, then they'll start to be more interested.

Agreed, but these discrepancies are of secondary magnitude and importance.

Want to bet? To give one example, there are current threads on this forum where drivers have published T/S data that is spectacularly inconsistent, i.e. if certain values are correct, the others in the same table cannot possibly be what is given, and the variation so enormous that any attempt to design using the published, inconsistent values goes to pieces. I wouldn't call that of 'secondary magnitude and importance' myself.
 
Last edited:
Want to bet? To give one example, there are current threads on this forum where drivers have published T/S data that is spectacularly inconsistent, i.e. if certain values are correct, the others in the same table cannot possibly be what is given, and the variation so enormous that any attempt to design using the published, inconsistent values goes to pieces. I wouldn't call that of 'secondary magnitude and importance' myself.

The inconsistency of TS Parameters by certain brands have been around for decades. That is nothing new. Based on my measurements Eminence is the only brand that I would not make the effort to measure . This is due to owning a good amount of their loudspeakers. Eminence TS Parameters reflect their new driver fresh out of the box. Unfortunately, there are many that tend to turn up their nose to Eminence.

There is a difference between a loudspeaker manufacture that distributes their speakers primary to loudspeaker companies worldwide and, a loudspeaker manufacture that distributes their speakers solely to hobbyists where a large portion do not know nor care what the TS Parameters mean.
 
Last edited:

Ant_222

Member
2006-08-03 8:41 pm
Want to bet? To give one example, there are current threads on this forum where drivers have published T/S data that is spectacularly inconsistent, i.e. if certain values are correct, the others in the same table cannot possibly be what is given, and the variation so enormous that any attempt to design using the published, inconsistent values goes to pieces. I wouldn't call that of 'secondary magnitude and importance' myself.

Not having had many new expensive foreign (non-Russian) drivers, I didn't know T/S were so unreliable all over the world. I can excuse, however, my old Soviet ones, many of which sound well anyway...

With the measurement conditions, one should be able, at least, to separate systematic error from random deviation.
 
Another obscure matter is the signal source used to measure the T/S parameters: what output impedance, voltage, and current?

T/S are not single numbers, they are curves that are a function of drive, weather (temp, humidity & pressure), nature of the test jig, driver orientation.

The numbers you see are collapsed at some point from the curve. Pick a different point on the curve and you usually get different numbers.

At least with the driver’s i have used from the better manufacturers, the factory numbers typically lead me to the best boxes.

dave