getting to test many speakers for FR

[mcmahon48]

I am going to be testing a bunch of different speakers with the following equipment
virtins spectrum analyzer 3.1
Dell notebook computer
Soundblaster audigy nx
Tek DMM252
precision potientometer 10k
17 watt mono amp
audio resource eq100 electrotet microphone
high density pink noise
high density white noise


So now what other things would help in getting good results with out a lot of expense

I will be saving the result from my testing and testing conditions
to be shared onced I get started.

plus I will later show my tmtm design and results later

[mcmahon48]

what is the F3 frequency about for I just go about doubling the resonant frequency for to start the frequency to work with and use a zoebel to control the impedance to the amp

[djarchow]

What are the goals of your testing? Are you measuring the drivers for crossover design work?

You mention using pink noise and a spectrum analyzer, but this will give you the frequency response of the driver combined with the room response. Normally you would want to measure using a signal like MLS, a gated sine, chirp etc so that you can gate out reflections from the room. Software such as Arta, SoundEasy, Praxis etc will let you do this.

Also if you are tryng to measure the driver including impedance for crossover design, you shouldn't be using a zoble during the measurement. If you amp can't drive the driver without one, I would find another amp.

Regards,

Dennis

[bjorno]

Hi,

To backup your FR measurements: IMO, I recommend you to determine driver T/S parameters too:

http://www.sound.westhost.com./tsp.htm

In addition while your driver under test is placed in a box, you could measure the FR and from it determine the Q of the box i.e. Otc and do simple calculation to test if FR measurement is complying with some of the previous found T/S data.

1.Measure the driver in box FR at about ½ ” microphone distance at the box system resonance (impedance peak and no stuffing’s).

2.Compare SPL at a frequency where the driver shows to be ‘mass controlled’ i.e. where the FR flattens out:
Then the difference in dB SPL is = 20 x log (Q) or Q = 10^((+ or -) dB difference/20).

If the driver is peaking, the difference must be positive but if negative the FR is gently sloping towards lower frequencies (happens when Qtc lower than 0.707).

3.From the equation Qtc =Qts x sqrt (1+Vas/Vb) where Qts, Vas and Vb is previously calculated, check if the former FR determined Q is close enough to Qtc.


b

[mcmahon48]

My testing procdures are: testing will be done outside at night so there will be little air disturbance and little to none reflection. Plus air density will be good for get good results. I am doing this for I have many speakers that I can not find a FR graph, and of course without knowing where the best frequency response characteristic are you cant build the best speaker combination.
My main focus is the lower midrange and up. Anyone have schematics for a good 10 to 15 mono amp that has ideal testing characteristics with integrated power supply I have lm317 and lm327 voltage regulators

[Ron E]

If you want the measurements to be useful to anyone other than you, you might want to measure with Speaker Workshop or ARTA instead of all that other gear. With Speaker Workshop you can get usable frequency (and phase) response and Thiele Small by recording the impedance.

Presumably you will mount the speakers on some sort of baffle, you should record the size and the speakers' position on it, also record the temperature of measurement and any other conditions.

I don't know what that software system you have can measure, they are rather cryptic about it. Maybe you can do a sample measurement and we can lead you in the right direction.

[mcmahon48]

Everytime time I tried to install speaker workshop it just get hung up in the setup section and I can go no further

[mcmahon48]

it took a little while but I did get it up and running

[mcmahon48]

Here are the specs I have come across so far I have just started to gather the data And should you know the formulas for *
Re
Ro*
Zmax*
Xmax*
Z' reduced*
Fs
F3
Qms
Qes
Qts
Vas
Sd*
Cms*
Bl*
Le@1khz
Bv

[Ron E]

I have made some contributions to:
http://en.wikipedia.org/wiki/Thiele/Small
...most notably the equations you seek.

I believe Ro is Zmax/Re, not very useful.
Z'reduced ?

F3 is not measurable unless you build a box suitable for each driver.
Xmax is not measurable with simple equipment. Maybe you could make a DC displacement vs voltage curve

Sd is most easily measured as the diameter to the peak of the surround on each side, the diameter of the cone plus half of the surround. Some might say this is a bit large, saying you should use the diameter of the cone plus 1/3 of the surround.

Cms and Bl can be measured directly but it is easier to infer them from other measurements.

Le is a useless measurement other than giving an order of magnitude for advertising purposes. An impedance curve is far more valuable.