Optimizing speaker distortion

[Niels]

Acourate is a fantastic toolbox! I have been a convinced Acourate user since 2006.

For years I had a couple of old B&W 801F speakers in my basement. Now I renovated them in order to sell them. I removed all passive filtercomponents and mounted high quality cables directly to the back. Then I made active filters in Acourate with same charisteristics as the passive crossovers (380/3.600hz, 4th order Linkwitz). Big improvement! - - still room for improvement, especially the treble irritated me...

So I measured all the driver distortions with Acourate and a microphone close to the speakers in 0, 25 and 45 degrees off-axis. It turned out that especially the treble had quite a high distortion, especially 3-order distortion, at 3.800hz it was only 25-28 db below 1-order signal(!)
I ended up changing the crossovers from 380/3.800 into 250/6.000 hz, and from Linkwitz 4th into Ulis new UB Pol11 4th order filter.

What a change! - and the total speaker distortion is about 20 db better in the high frequency range, but even the bass is clearly better after the change.

So my advice to everyone is to pay attention to distortion measurements!

Acourate is a fantastic toolbox!

best regards to all users
Niels

[Labdoc]

Niels,
I agreee re Acourate is a fantastic tool box. I have a few questions you might be able to answer:

1) "I measured all the driver distortions with Acourate "
Do you have a "how to" guide you can provide to use Acourate for distortion measurement?

Mitch Barnett describes an approach, but does not give much explanation of the measurements, at this link https://audiophilestyle.com/ca/ca-acade ... lkthrough/

Edit: The Acourate Wiki also has an entry on distortion measurement in German. Assisted by Google Translate, my impression is that the Wiki is what Mitch states in his Audiophile Style entry. So, I still need some more information about distortion measurement and interpretation.

2) How did you choose the new XO frequencies? You mention Distortion. I guess linearity range as well. Any additional criteria?

3) Did you add a protection capacitor for the Tweeters? If so, what specification?
Murray

[UliBru]

A simple hypothesis: if a speaker driver does not like to play certain frequencies it will produce distortions in this range.

So usually my approach is this: a given driver shall play from a 1st XO corner frequency to a 2nd XO frequency. Example: Niels' midrange driver = 250 to 6000 Hz. In addition it is necessary to think about the width of the XO transition area. Creating a XO from 250 Hz/6kHz with UB jPol11 4th order allows to check the overall range by simply loading XO2. It goes from 200 Hz to 7150 Hz.

Now I run a logsweep measurement WITHOUT crossover from 200 Hz to 7.2 kHz. And then I check the harmonic distortions. If some distortion is too high in the measured range I check the frequencies where the distortion is as low as possible, at least acceptable. This defines at the end the new XO frequencies.

Of course the same procedure has to be done with all drovers. And of course hopefully at the final crossovers all drivers show up a low distortion.

It is also necessary to think about the radiation of the soundwave at the XO frequencies. If the driver diameter gets into the size range of the XO wavelength it starts to beam. It makes sense to avoid a radiation picture like a Christmas tree.

Anyway this method is like a comfortable synthesis. Take the proper drivers, let them play will low distortion = assign proper XO frequencies and filter types and you get a nice speaker. It's that simple.

Tweeter protection: I run tweeters without protection since twenty years without problem. But it is necessary to use good amplifiers. They are not allowed to produce a DC current with power on/off or failure state. Otherwise you can use a capacitor or, without influencing frequency and phase response, also a fuse.

[Labdoc]

Thanks Uli. I am adding this to my Acourate project list!

[dzeekee007]

Hi,

Thanks for very interesting thread.

Uli, What valor of fuse could you advice ?

[UliBru]

The fuse depends of course on the continuous current. With a given tweeter power and its resistance you can calculate the current by
P=U*I and U=R*I as
I = sqrt(P/R)

So with e.g. P= 40W and R=4 Ohm you get an current of 3.16 A

IMO it does not make sense to use such a big fuse first. Simply start with smaller values and replace it with a higher value if the fuse blows up during normal operation.

Another example: if a 4 Ohm tweeter creates 88.5 dB at 2.83V the according power is P = U²/R = 2 W. Then the current is 0.7 A.
So you can get a feeling by thinking about how loud you will typically listen to the music.

[dzeekee007]

Uli, you are the Best !
Thanks a lot, very very useful.

Should be a fuse semi slow or ...?

-------
On occasion, (because of your great knowledge), maybe can you tell me the formula to find:
how much the power will supply the ACA amp* (balanced mono mode)
*(the same DIY amp that Mitch use for his tweeters, https://diyaudiostore.com/products/amp-camp-amp-kit),
to drive my tweeters (TC120TD5-focal) which has the following specs:

• efficiency/sensivity 93.5 dB
  nominal power 15W
  maximum power 150W
  nom. impedance 6 ohms
  DC resistance 6 ohms
  inductance 1.49 mH

for a normal listening level (about 80 db).

Thanks in advance and great day to all.

PS
BTW, my fuse's question concerned the tweeters mentioned above.

[UliBru]

The tweeter efficiency is defined as 93.5 dB per Watt.
The ACA amp specs is 8 W into 8 Ohms. so it can deliver an output voltage of 8 V and a current of 1 Ampere.
The distortion is 0.7% = -43 dB.

As the tweeter impedance is 6 Ohm it will reach 93.5 dB at 2.5 V and 0.4 A. So the amp is capable to run the tweeter from this figure. You should anyway calculate with ~95 dB instead of 80 dB.

[dzeekee007]

Thanks a much Uli,
Very helpful

Regading the fuse (tweeter's protection), should it be fast or semi-slow or ...?

[UliBru]

The fuse wire has to melt quicker than the coil wire or the insulating varnish of the tweeter.
There are two fuse parameters: nominal current and speed. The safest way is to select a low currect value and a high speed.
But this may be uncomfortable if you need to change the fuse too often. Then you can increase the current and/or lower the speed.

As I do not use a fuse or a capacitor by myself I cannot tell you about an optimal selection to start with.