# Understanding Audio Measurements

many bad measurements are caused by poor power supply or power supply related issues.

Does it matter if measured unit is 110V or 230V (or 110/230V switchable unit, 50/60Hz hum etc).

Thanks again

Does it matter if measured unit is 110V or 230V (or 110/230V switchable unit, 50/60Hz hum etc).

50Hz hum at the output (at the same level) might be slightly less audible...

230V may radiate EMF more strongly than 120V affecting whatever gets affected (guessing a bit here), or even more weakly, as the current flow would be reduced vs 120V (all else being equal).

Does it matter, one over the other? Probably not.

Last edited:
trl
Based on the above "answer", I have an EE question;

I have a wire energized with a sine wave voltage, and some current flow. Mabe later it is only energized, but no flow.

Is there one, or two fields around it... I'm thinking what I might call "electrostatic" induced by the voltage and the other "magnetic" induced by the current, or are they the same thing, or what?

Last edited:
@RayDunzl
There are two orthogonal fields: a concentric magnetic field and a radial electric field.

EDIT: no current (flow), no mag field.

Last edited:
^^^ You could be right! See Maxwell's equations (didn't everyone have to memorize them? And the periodic table...)

With voltage there is a electric field, which may be static or not. ESLs depend upon that voltage field. If there is current flow then there is a magnetic field at right angles to the current flow (see right-hand rule). The magnitudes depend upon the voltage, current flow, and distance from the conductor (wire). All basic EM problems seen in any undergrad course, i.e. something I knew something about in 1984 or so and have mercifully forgotten by now.

Is it possible to measure slew rate of an amplifier ? And how it is important ?

Is it possible to measure slew rate of an amplifier ? And how it is important ?

It is possible. Usually inadequate slew rate is all tied up with inadequate bandwidth. Slew rate being an issue is pretty much a problem of a bygone era unless of course someone is making designs of a bygone era.

Here is the formula for calculating slew rate:

Slew rate equals or exceeds 2pi the frequency times the peak voltage. A higher voltage at a given frequency requires higher slew rate. A higher frequency at a given peak voltage requires higher slew rate. So maximum voltage at maximum frequency is the slew rate limits.

Spectral amps had extremely high slew rates for power amps. They had a genuine megahertz or two of flat bandwidth, and that combined with their peak voltages gave them something like 500 volts/microsecond or later 1000 volts/microsecond. Way more than needed for clean audio reproduction.

So just as an example suppose you send a 20 khz signal at 30 volts peak and the amp can do that. 6.28*20,000*30 equals 3.77 volts/microsecond. Which would be about 56 watts into an 8 ohm load.

There are standards for measuring slew rate of amplifiers and can easily be done.
You just need an oscilloscope with a bandwidth that is wider than that of the measured device.
Some manufacturers actually have slew-rate numbers in their spec sheet.
Slew rate is not something that is in a standard measurement suite.
Slew rate can also be asymmetrical (rise time may be faster or slower than fall-time)

With modern electronics slew-rate should not need to be a limiting factor for reproduction of audio.
In practice it depends on design, used components and feedback.

When one applies a square-wave (with steep edges) the feedback circuit can/will be saturated.
In this case the output goes to the intended positive (or negative) output voltage as fast as it can right up to the point where feedback takes over again. Weird over and undershoot can occur which is easily spotted in squarewave response plots.
It says something about how fast the circuit is and how well feedback works.

upper Bandwidth and slew rate as well as 2nd harmonic distortion levels near the frequency limit are related to this.

How high slew rate should be thus depends on the maximum output voltage the amplifier must reach (includes current under load) and maximum frequency it must be able to reach.
For electrostatic HP amps and high power amplifiers higher numbers are needed than for line level stuff as slew-rate is expressed in Volt per time-unit

(EDIT: seems I was crossposting with Blumlein 88)

There are standards for measuring slew rate of amplifiers and can easily be done.
You just need an oscilloscope with a bandwidth that is wider than that of the measured device.
Some manufacturers actually have slew-rate numbers in their spec sheet.
Slew rate is not something that is in a standard measurement suite.
Slew rate can also be asymmetrical (rise time may be faster or slower than fall-time)

With modern electronics slew-rate should not need to be a limiting factor for reproduction of audio.
In practice it depends on design, used components and feedback.

When one applies a square-wave (with steep edges) the feedback circuit can/will be saturated.
In this case the output goes to the intended positive (or negative) output voltage as fast as it can right up to the point where feedback takes over again. Weird over and undershoot can occur which is easily spotted in squarewave response plots.
It says something about how fast the circuit is and how well feedback works.

upper Bandwidth and slew rate as well as 2nd harmonic distortion levels near the frequency limit are related to this.

How high slew rate should be thus depends on the maximum output voltage the amplifier must reach (includes current under load) and maximum frequency it must be able to reach.
For electrostatic HP amps and high power amplifiers higher numbers are needed than for line level stuff as slew-rate is expressed in Volt per time-unit

(EDIT: seems I was crossposting with Blumlein 88)
Demian Martin, designer of the early Spectral amps told me that slew rate wasn't the real reason for the extreme bandwidth. It was the ability to let go of a signal, the fall time, they were trying to make much better. The wide bandwidth, and the high slew rates, were simply side effects of getting extremely fast fall times to let go of signal transients. Those amps were spectacularly clean, transparent, and musical. Or so they seemed when driving my Quad ESL63 speakers which incidentally were the speakers Mr. Martin was designing for.

Demian Martin, designer of the early Spectral amps told me that slew rate wasn't the real reason for the extreme bandwidth. It was the ability to let go of a signal, the fall time, they were trying to make much better. The wide bandwidth, and the high slew rates, were simply side effects of getting extremely fast fall times to let go of signal transients. Those amps were spectacularly clean, transparent, and musical. Or so they seemed when driving my Quad ESL63 speakers which incidentally were the speakers Mr. Martin was designing for.

"Ability to let go of a signal"? That is actually slew rate or power bandwidth, but if the amp can drive the speaker at full amplitude at 20 kHz, then the amp is "able to let go of a signal" fast enough.

EDIT: no current (flow), no mag field.[/QUOTE]
Based on the above "answer", I have an EE question;

I have a wire energized with a sine wave voltage, and some current flow. Mabe later it is only energized, but no flow.

Is there one, or two fields around it... I'm thinking what I might call "electrostatic" induced by the voltage and the other "magnetic" induced by the current, or are they the same thing, or what?

sine wave voltage imply time-dependent power supply so current is also time-dependent (alternating current or AC). Time-varying current will generate time-varying magnetic field which in turn will induce time-varying electric field that will oscillate orthogonal to the original time-varying magnetic field. In physics terms it is not meaningful to say "energized wire." If there is current (net flow of charge from point A to B) then there must be a potential difference (voltage) between points A and B.

electrostatic refers to electric fields associated with stationary electric charge. of course, motion is relative so someone running past a stationary electric charge will also "see" magnetic field. this is related to the theory of special relativity.

@amirm one question, let say you're measuring headphone amplifier with certain resistor value to simulate the load, where do you place the resistor, close to the source output port or the analyzer input port?

Setup A or Setup B ? Thanks!

Setup B is what I started with until some amps got so good that I also built a setup A (for Benchmark HPA4 for example). I still use my fixture for B since it is more flexible (multiple impedances) but if I suspect an amp is very good, I resort to A. Once I get some time, I will phase out B.

For speaker testing, it has been setup A (Kelvin-type connection) for a while.

Setup B is what I started with until some amps got so good that I also built a setup A (for Benchmark HPA4 for example). I still use my fixture for B since it is more flexible (multiple impedances) but if I suspect an amp is very good, I resort to A. Once I get some time, I will phase out B.

For speaker testing, it has been setup A (Kelvin-type connection) for a while.
Sorry if this is the wrong place to ask but I've been thinking about trying to measure the output impedance of the DX3 Pro V2 that I recently bought since Topping only specs it to be <10 Ohm, but I have a few questions about how I would go about to do that.
1. How do I set everything up to do the measurement, do I need anything other than a set of resistors and a multi-meter?
2. How good does the multi-meter need to be? I'm not trying to be super exact in my measurements so as long as I could get the measurement to be within +- 0.5 Ohm or there about I would be happy. Also I would like to measure at different frequencies between 20 Hz and 20 kHz to see if the OI changes with frequency and from what I understand a lot of cheap multi-meters are only accurate around a certain frequency (200 Hz?) so how expensive would it be to get one that can measure accurately close to 20 kHz?

Also I would like to measure at different frequencies between 20 Hz and 20 kHz to see if the OI changes with frequency and from what I understand a lot of cheap multi-meters are only accurate around a certain frequency (200 Hz?) so how expensive would it be to get one that can measure accurately close to 20 kHz?
This, would be the challenge. I don't know which meters have such wide bandwidth. Otherwise it is very simple. You play a low frequency tone. Set the output level to 1 volt or so to avoid clipping. Then load it down with a resistor. Then a simple formula will give you the output impedance.

This, would be the challenge. I don't know which meters have such wide bandwidth. Otherwise it is very simple. You play a low frequency tone. Set the output level to 1 volt or so to avoid clipping. Then load it down with a resistor. Then a simple formula will give you the output impedance.
So if I only measure it at a low frequency I could get away with using a cheap multi-meter? Also exactly how do I load the output with a resistor, do I just put the end of one of these in the output of the amp and put the multi-meter on the other end?

Setup B is what I started with until some amps got so good that I also built a setup A (for Benchmark HPA4 for example). I still use my fixture for B since it is more flexible (multiple impedances) but if I suspect an amp is very good, I resort to A. Once I get some time, I will phase out B.

For speaker testing, it has been setup A (Kelvin-type connection) for a while.

Noted Tx! May I know why setup A is preferred? Is it to make the test cables less relevant to the test result?

One more question, for measuring line output what would be the preferred load impedance? Is it 10kohm? Or can we just connect it to the analyzer HiZ input?

Thanks Amir!

So if I only measure it at a low frequency I could get away with using a cheap multi-meter? Also exactly how do I load the output with a resistor, do I just put the end of one of these in the output of the amp and put the multi-meter on the other end?

This probably help to explain:
http://www.zen22142.zen.co.uk/Theory/inzoz.htm

Scroll down to Output impedance.
Check the multimeter that you're going to use, what is the AC voltmeter frequency range, and use a tone within the range.
Check the watt rating of the resistor that you're going to use, make sure the power rating is high enough to handle the power dissipation during measurement. Basically no need to set the output voltage too high. Around 2 Volt rms, unloaded (without resistor) is usually good enough. 0.25 - 0.5 watt of 33 ohms resistor is good enough for the measurement.

A small comparison of different measuring methods performed on the same component. One can see that THD+N at 1kHz (or call it SINAD if you want) is the easiest and least demanding test for the component under test, which is true for any amplifier and as such may be generalized, and it tells the least about the component parameters.

Replies
108
Views
12K
Replies
7
Views
702
Replies
2
Views
2K
Replies
57
Views
15K
Replies
49
Views
19K