Bench Test for Speaker Impedance

[Andy Le Blanc]

I stumbled upon this in AudioXpress Dec. 2003. Looked like it could be useful.

You put a 1k resistor in series between an speaker and audio oscillator.

At any given freq. you set up the osc. to read 1v across the 1K resistor (points A and B),
Then read the impedance as a voltage across the speaker ( points B and C)

1mv = 1ohm, plot out the results on graph paper to get a clue as to what your
speaker is doing in it's chosen cab. Tune the amp, choose a better driver, pick a better box.

Much more valid than saying you liked the sound of this Vs. that.

[vibratoking]

On the surface it looks like it will provide some information. I see a few problems.

Driving the speaker with 1Vrms/1kOhm = 1mA. That's not enough drive to indicate what the speaker impedance will be at playing volume.

To get a plot, you have to measure alot of points in order to avoid missing any narrow peaks in the response. Speaker plots that I have seen generally show a pretty narrow peak somewhere in the response plot.

[FYL]

You may also consider buying a dedicated tool such as Dayton's WT3, currently available in a bundle with a precision scale for $99.88.

http://www.parts-express.com/pe/showdet ... er=390-803

[FYL]

Driving the speaker with 1Vrms/1kOhm = 1mA. That's not enough drive to indicate what the speaker impedance will be at playing volume.

Agreed, but the difference is negligible.

[Andy Le Blanc]

That's why I posted, it "looked" like a good Idea.

Don't mind plotting a lot of points, is there a known criteria to test this with?

What would be enough V to drive a test of this sort? It would be great to
have a bench test like this with very simple tools.

Is it an issue? you turn up your signal generator until you measure 1v over
the 1k at your given freq. then measure across the speaker as 1ma=1ohm,
as the impedance at that freq.

Practical?

[fperron_kt88]

Works great, albeit with lower values res (1 Ohms 10w). you can sweep the sine freq and plot the voltage drop on substracting channels on a scope. Instant resonance freq.

Very usefull to understand how a cab shifts the params of a speaker. Usefull when verifying thiele small params.

[Andy Le Blanc]

Not everyone has the budget for tools, you have to be practical.

So... a 1ohm and a DMM or two.

Can you nail a test procedure... pictorial and simple?

[vibratoking]

I like this thread...I'm about to learn someting

You may also consider buying a dedicated tool such as Dayton's WT3, currently available in a bundle with a precision scale for $99.88.

Thanks for the info on that. Interesting...reviews show that the results can vary widely based on the test setup. My guess is that this is one of those measurements that is easy with an anechoic chamber and not that easy anywhere else. Anybody have an experience with this?

Agreed, but the difference is negligible.

Really? I don't have any experience either way, but it seems like it should not be neglible if you make the measurement in a cab so that the effect of the cab is part of the measurment. I would also think that the internal mechanical effects of the speaker would be very different for low vs high current, but this is not true? I would love to be educated.

Don't mind plotting a lot of points, is there a known criteria to test this with?

Not that I'm aware of. If you have a network analyzer, you can pretty much select whatever number of points you want at the cost of a small amount of time.

[fperron_kt88]

Just connect a 1 ohm res and measure Vac with a dmm. Sweep freq for minimum voltage. This is your resonance freq. Install the speaker in a different cab and measure again. You can figure Vas if you know the volume of two closed cab that fit your speaker.

Simple with simple tools.

[fperron_kt88]

Memories are coming back: you want to measure the -3 dB points around resonance so that the Q can also be computed with VAS.

[vibratoking]

Simple with simple tools.

Well kinda simple. You either need a digital sig gen to know what the frequency really is at min voltage or you have to have a scope or some other device to measure that frequency.

Memories are coming back: you want to measure the -3 dB points around resonance so that the Q can also be computed with VAS.

This is why I would want to sweep and generate a full plot. Then I could find the 3dB points by interpolation, if I had enough points to start with.

[FYL]

Can you nail a test procedure... pictorial and simple?

Sure. Easiest is the constant current method you've outilined, the constant voltage method is better - you just need to add an amplifier between the generator and the load.

http://www.t-linespeakers.org/tech/cons ... index.html

Worth reading:
http://sound.westhost.com/tsp.htm

Useful free software and resources:
http://www.speakerworkshop.com/

[FYL]

Interesting...reviews show that the results can vary widely based on the test setup. My guess is that this is one of those measurements that is easy with an anechoic chamber and not that easy anywhere else.

Measurements are AOK if you do them with the speaker mounted free-field. Results will vary with other positions.

And you don't need anechoic chambers anymore - except for some fancy measurements, MLS and similar techniques on computers allow to get near perfect amplitude measurements in standard rooms.

it seems like it should not be neglible if you make the measurement in a cab so that the effect of the cab is part of the measurment.

Of course : tests should be done on bare speakers in free-air. Adding any kind of load will greatly modify what you measure.

I would also think that the internal mechanical effects of the speaker would be very different for low vs high current, but this is not true?

There are really negligible when it comes to basic Z and T/S params. All manufacturers use small signal measurements, typically with 2.83V (1W) drive.

Large signal analysis becomes important when you want to measure power handling and real-world frequency response. Then params such as dynamic compression (for modern drivers) and mechanical excursion (for vintage ones, with their itsy bitsy Xmax) become really important...

[Andy Le Blanc]

Excellent..... More than one method.....

DIY One heck of a lot better than guessing.

[Andy Le Blanc]

can these methods be modified to test other inductors?

For example, guitar pick-ups, chokes, opt?