# Freq Response of Reactive Load

#### p_haslam

Hey folks

Apologies, I'm still in the asking question phase of life!

I've built a reactive load box, here is the circuit, with a tap off to go to an audio pcb board (between the 47k and the 1k resistors)

In LT spice it gives me the impedance response I want and a flat frequency response. However, when I test it with ARTA and white noise I get this:

Really (obviously) noticeable when I plug in the guitar amp and listen to the output (I have a speaker emulator on the headphone output)

I can't work it out. Is it layout? here is the prototype:

#### p_haslam

I'm assuming I'd expect a flat frequency repsonse for the load and then I'm doing a speak cabn circuit sparately....maybe a wrong assumption?

#### rayma

Between which two nodes are you testing the frequency response?
If measuring across the 8 ohm resistor, you should move it so one end is grounded.

#### 6A3sUMMER

I agree with rayma.

Looks like the load might be 8 Ohms.
But 8 Ohms is Floating at both ends.

Your signal source connections are well defined, but . . .
Define where you are measuring the output.
From what point to what point?
Some software might not be able to calculate a floating measurement.
Some measurement hardware is not able to measure a floating circuit load.

What is the purpose of your network?
(the intended function of your network).

15mH, 220uF is parallel resonant at 87.6Hz.
Above and Below 87.6 Hz, the impedance drops rapidly.
15mH is 4.7 Ohms at 50 Hz, and 220uF is 4.8 Ohms at 150Hz.
Above and below the 87.6 Hz resonance, the 150 Ohm resistor is totally swamped out by Lx and Cx.
Woofer resonance simulator?

R4 and R5 appear to do almost nothing versus the low impedance of the rest of the circuit.
The signal source is a Zero Ohm Voltage source, Right?

Is the 8 ohm simulating the DCR of a woofer?

And how about the 1mH inductor, does it simulate the woofer voice coil inductance?

Last edited:

#### rayma

If we assume that you want to consider the output voltage to be across the 8 ohm resistor,
some ballpark estimates for the frequency response:

You get no attenuation of the voltage across the 8 ohm resistor at most frequencies below 8kHz.

Around 90Hz there's a tuned circuit that will drop the 8 ohm load voltage
by around -26dB within a certain bandwidth that is centered on 90Hz.

Above 8kHz, the 8 ohm load voltage will begin to drop off at 20dB/decade,
down to an asymptotic level of -17dB.

Layout or lead dress has nothing to do with any of this.

Last edited:

#### tikiroo

In LT spice it gives me the impedance response I want and a flat frequency response. However, when I test it with ARTA and white noise I get this:
You should use pink noise not white noise. White noise gives a rising response with no equalisation.

#### 6A3sUMMER

Is your signal source is a Resistive signal source?
A voltage source that drives a series 1 Ohm resistor, for example, to simulate an amplifier 8 Ohm secondary tap that has a damping factor of 8 (referred to 8 Ohm load).

You stated the resistive divider is to drive an audio board.
That makes sense.
Are you measuring the signal frequency response there?

We need to know more about your measurement points, signal source output impedance, etc.

#### Mister Audio

You should use pink noise not white noise.
What is the purpose of your network?
(the intended function of your network).
???
As you can see, it's all questions at this stage.

#### wg_ski

Really (obviously) noticeable when I plug in the guitar amp and listen to the output (I have a speaker emulator on the headphone output)
A headphone output is not designed to drive a speaker (Well, a tiny “speaker” much higher in impedance than 8 ohms). Frequency response error can be far worse that what the “emulator” circuit is supposed to do when connected in place of the actual speaker.

#### p_haslam

Hey all,

Thanks so much. Sorry, mad day.

Apologies for noob status, haha.

I did have problems with the 8r resistor floating so I grounded it and seemed better (as in the 50 cycle hum went!)

The purpose is to act as a reactive load for an amp, simulating a speaker. So I can plug the speaker output from my guitar amp into it and five it hard with destroying the OPT.

I tried measuring the response at the resistor divider and at the 8r resistor, seemed the same.

At first I used ARTA to generate the signal and record. Then moved to a separate sig gen putting out ‘noise’ which I assume is white noise. As I wondered if the computer output (via a scarlet solo usb interface) wasn’t accurate. That seemed to lessen the HF response. Loser to flat but not completely. Perhaps I need to screenshot it.

The circuit is based on the Aitken article with me trying to get close to a Celestion creamback (and missing!)

Pat

#### rayma

The two resistor divider response will be flat.
The response across the 8 ohm power resistor will not be flat at all.

1 user

#### p_haslam

Sorry, values were based on wanted to be able to handle 200w into 8 ohm load, ie so I could test big amp rigs with space to spare. The divider allows me to send an input into an opamp setup to trim the level to line level

#### p_haslam

The two resistor divider response will be flat.
The response across the 8 ohm power resistor will not be flat at all.
Trying to understand why, because of power dissipation? Inductance of the resistor?

#### rayma

No, the circuit does that due to its topology and choice of component values.

#### p_haslam

Right, yes of course. Misunderstood.

#### rayma

The two resistor divider operates completely independently of the speaker impedance circuit,
simply because both are directly driven by the input voltage source.

So the two resistor divider simply reduces the input voltage (by about -33dB),
without changing it otherwise at all.

1 user

#### 6A3sUMMER

Purchase a 400Watt 8 Ohm non inductive resistor (or some combination of series resistors, 2 + 2 + 2 + 2 at 100 Watts each).
That will load most of your guitar amplifiers, except some that were used at Woodstock 1.

Then, use the same resistor divider to get a sample to send to your scope, sound card, headphone amplifier, etc.

Many guitar amplifiers are not going to have a flat frequency response from 20Hz to 20,000Hz. Why should they?
They are guitar amplifiers, they are Not Hi Fi amplifiers.

That is why there is an Instruments &j Amps section of threads on diyAudio.
(Guitar Amps).

We are at thread # 17.
'Nuf said.

#### Mister Audio

It looks to me like all you need to do is change the 'Hot/Active' input connection of R4 to the Hot side of C1.
Then your low level output can be taken across R5 (one side being ground). You will then hear the 'sound' of your filter.

#### 6A3sUMMER

You really need to draw out an accurate and correct schematic.
Post # 1, plus all the other posts descriptions of what to connect and how, do not describe a correct electrical model of a dynamic speaker:
(voice coil, magnet and magnet structure, cone and suspension, cabinet or not, and cabinet description such as open baffle, ported, closed box, horn, etc.

Load simulators that simulate a speaker can be done right, or can be done wrong.

Many amplifiers are tested with a simple non-inductive resistor of the proper resistance and conservative power rating (do not burn the bench or your fingers).
Any amplifier that can not do a power output test with such a resistor (without oscillating), such an amplifier probably very badly needs to be re-designed.

A speaker simulator load is most often used to see if an amplifier works properly with more complex load that is not purely resistive across the audio band. If the amplifier oscillates when connected to a good speaker simulator load, the amplifier probably very badly needs to be re-designed.

Amplifiers outputs are not always understood . . .

I do not know what your amplifier model is, but it very well might not be a true Voltage source.
Any voltage source that is loaded with a low enough impedance will reduce its output voltage (or go up in smoke).

Many solid state amplifiers "act" like a voltage source, until they run out of current, or until the negative feedback loop causes oscillation.
Many tube amplifiers "act" like a voltage source, until they run out of current, or until the negative feedback loop causes oscillation.
Many other tube amplifiers act like a voltage source in series with a resistor inside the amp at the output connectors). The load resistor is separate and outside the amplifier.
Many other tube amplifiers act like a current source with a resistor in parallel inside the amp at the output connectors). The load resistor is separate and outside the amplifier.

Thevenin's theorem, and Norton's theorem come to mind.
Look them up.

An amplifier with moderate output impedance that drives a speaker load simulator . . .
If you tap off of that simulator (filter as some of you want to call it; it is a Load),
It will not necessarily sound like the actual speaker's sound.

All Generalizations Have Exceptions.

Just my opinions

Last edited:
1 users

#### Mister Audio

If you look at the circuit, you will find R3 = 8 ohms, so the source amplifier
is protected from the reactance of the filter.
This is probably a totally legitimate circuit to simulate a particular 'sound'.
(it is just a case of R4 being properly connected)

1 user