What do these symbols mean in this diagram for creating an impedance rig in REW?

[IPunchCholla]

I was thinking about building a rig for testing my speakers impedance (for funsies) since I have a couple sound cards just sitting around, but I am unfamiliar with what these ovals in the diagram below mean and can not find their equivalent in searching. Any help is appreciated!

 

[MakeMineVinyl]

Those are the shield of the cable.

 

[LTig]

I think it's the ground connection of the RCA output.

 

[DVDdoug]

Most soundcards can't directly-drive a 4 8-Ohm speaker. This kind of test can work with headphones. All laptops can drive headphones but I'm not sure that all line-outputs of regular soundcards can drive headphones.

And most laptops only have headphone-out and mic-in, so that's no good either. (A headphone signal into line-in is OK.)

You can do a similar test with an amplifier driving the speaker but you should use a multimeter (instead of line-in on the soundcard) because a speaker-signal into a line-input can fry it.

 

[AnalogSteph]

Looks like the circuit is pretty much the equivalent of Fig. 6 here. Rsense adds to the speaker impedance so that the output doesn't break too much of a sweat, with total impedance ending up in lower headphone terrain.

All laptops can drive headphones but I'm not sure that all line-outputs of regular soundcards can drive headphones.

Output driving ability may in fact vary. If you don't have anything special floating around, try a Creative card, those generally have decent headphone drivers built-in (NJM4556A or more recently MAX97220).

 

[pseudoid]

Forget schematics and wiring diagrams: This is neither!
Darker black 'rings' (where you have the red-arrows) normally indicate a shield/ground (rather than a wire normally used as a return wire).
But if they were true shields meant for grounding (one or both ends); they would have to be terminated and also NOT used for the return of a +/- signal pair.
It appears to be just a sketch and take it as that....

 

[LTig]

Most soundcards can't directly-drive a 4 8-Ohm speaker. This kind of test can work with headphones. All laptops can drive headphones but I'm not sure that all line-outputs of regular soundcards can drive headphones.

And most laptops only have headphone-out and mic-in, so that's no good either. (A headphone signal into line-in is OK.)

You can do a similar test with an amplifier driving the speaker but you should use a multimeter (instead of line-in on the soundcard) because a speaker-signal into a line-input can fry it.

In this case the load is the speaker + Rsense. Rsense can (and should) be much higher than the nominal impedance of the speaker. The higher Rsense is the closer is the relation between measured voltage and speaker impedance.

 

[MAB]

I was thinking about building a rig for testing my speakers impedance (for funsies)

Did you ever build this? It's a good project and it works well.

Slightly counterintuitively, the best value is of Rsense is determined by the ability of the soundcard's output drive capability. It's the sense-current sourced to the speaker and the L-R soundcard inputs.

If using the raw output of your soundcard, you need a resistor in the ~1 kOhm region. That means very low measurement current even at the upper limit of the soundcard's output voltage capability.

I use either a headphone amp or a power amp hooked up t the soundcard output. With a headphone amp I use 30 Ohm Rsense since I have a headphone amp that can handle a load of 30 Ohs plus the driver's resistance. For the power amp, I use somewhere around 8 Ohms.

Here is a comparison of a driver measured with an amplifier and a 10 Ohm resistor, and a 2 kOhm resistor, and DATS:

The data captured with REW is unfiltered. The 2 kOhm Rsense resistor has more noise than the 10 Ohm, but you need to zoom in to see it. But both are more detailed than the over-smoothed DATS.
Here is the same data but 1/12 octave smoothing:

A headphone amp gives great results, really not any worse. You can do 4M sample sweeps and averages, plus several noise filters, which give results even better than the above. I can see changes in the impedance of a woofer in a cabinet more easily than I can measure changes in the radiated sound. I can even see the effect of the heater fan in the measurements. Most drivers, DATS is closer matched.

The fixture is simple. Two halves of a TRS cable. Right output and left input wired together, shield wires also wired together. Rsense driving the woofer and the right input in parallel.
Here is an example, with the 2kOhm resistor.

The main limitation is the amp's output voltage overloading the inputs.
I use a divider on the inputs to keep the voltage below the input's overload. This also allows measurement of a driver's large signal parameters too, which do change from the normal small-signal T/S parameters. Although not as much as many people suggest. For large signal, you need a 10 or 20dB reduction on the soundcard input. And very tolerant household members, and even neighbors!

 

[Flaesh]

divider

What should a divider be and what determines it; how to correctly create a divider for some arbitrary significant signal?

 

[DVDdoug]

Actually Rsense together with the speaker make a Voltage Divider.

With a 2K resistor the signal will be reduced by enough to be safe.

 

[MAB]

What should a divider be and what determines it; how to correctly create a divider for some arbitrary significant signal?

For me, I have a Topping A30 Pro headphone amp with only one channel working that I use for speaker testing, it is stable down to below 30 Ohms. So I use a 30 Ohm Rsense. With a 4 Ohm driver that's a 34 Ohm load the amp sees, this keeps me in the stable region for the A30 that I have. Obviously no way the soundcard output could do this. 30 Ohm may be too low for some devices. 50-100Ohm Rsense would work with most any headphone amp or soundcard headphone-out.

In fact, the A30 Pro I use can source over 13 V into 34 Ohms, so I am really limited by the soundcard input which starts getting unhappy above 2V.

So the answer is complicated.

If you have a soundcard with a headphone out and an input, using a 100 Ohm Rsense is going to give great results. If you have a stout soundcard or external amp you can go even lower Rsense, but you have to be careful to not overload your soundcard inputs.

 

[IPunchCholla]

Did you ever build this? It's a good project and it works well.

View attachment 520366
Slightly counterintuitively, the best value is of Rsense is determined by the ability of the soundcard's output drive capability. It's the sense-current sourced to the speaker and the L-R soundcard inputs.

If using the raw output of your soundcard, you need a resistor in the ~1 kOhm region. That means very low measurement current even at the upper limit of the soundcard's output voltage capability.

I use either a headphone amp or a power amp hooked up t the soundcard output. With a headphone amp I use 30 Ohm Rsense since I have a headphone amp that can handle a load of 30 Ohs plus the driver's resistance. For the power amp, I use somewhere around 8 Ohms.

Here is a comparison of a driver measured with an amplifier and a 10 Ohm resistor, and a 2 kOhm resistor, and DATS:
View attachment 520376

The data captured with REW is unfiltered. The 2 kOhm Rsense resistor has more noise than the 10 Ohm, but you need to zoom in to see it. But both are more detailed than the over-smoothed DATS.
Here is the same data but 1/12 octave smoothing:
View attachment 520377
A headphone amp gives great results, really not any worse. You can do 4M sample sweeps and averages, plus several noise filters, which give results even better than the above. I can see changes in the impedance of a woofer in a cabinet more easily than I can measure changes in the radiated sound. I can even see the effect of the heater fan in the measurements. Most drivers, DATS is closer matched.

The fixture is simple. Two halves of a TRS cable. Right output and left input wired together, shield wires also wired together. Rsense driving the woofer and the right input in parallel.
Here is an example, with the 2kOhm resistor.
View attachment 520379

The main limitation is the amp's output voltage overloading the inputs.
I use a divider on the inputs to keep the voltage below the input's overload. This also allows measurement of a driver's large signal parameters too, which do change from the normal small-signal T/S parameters. Although not as much as many people suggest. For large signal, you need a 10 or 20dB reduction on the soundcard input. And very tolerant household members, and even neighbors!

I did! It’s been a while but I do have the measurements around here somewhere. Being a newbie something about that diagram broke my brain. I think I modeled it in SPICE before building (something I learned for the project) Used an old external sound card the two weight method to get the measurement.

Rehoused a pair of three ways with new crossovers and they measured in line with the models.

I then bought a pair of monitors at a stupid good price that performed better than I was ever going to get the drivers, so the customer cabinet rehoused speakers largely are unused now.