MEASURING amplifier clipping

[Head_Unit]

OK, I'm tired of threads on various sites with people asking about amp rating and speaker rating and do they need an external amp and people positing all kinds of distance formulas and rules of thumb and the (often-exaggerated) speaker sensitivity to essentially GUESS if the amplifier is clipping.

Let's improve on that!

Most folks do not have a storage oscilloscope on hand (though if someone knows a good phone or PC/Mac/tablet app like that please post!).

Here's my idea of a procedure to measure if the amp is close to clipping or not (for digital sources at least):
(1) Note the maximum volume setting you play at.
(2) Get a good AC voltmeter. Connect it across the front right or left speaker.
(3) You need a say -20 dB test tone, like several hundred Hertz.*
(4) Turn the volume DOWN a chunk, play the track, then move quickly up to your maximum volume and read the AC voltage.
(5) Compare to this chart (the watts being the 8Ω rating of the amplifier):
30W: 1.5V 50W: 2.0V 60W: 2.2V 80W: 2.5V 100W: 2.8V 120W: 3.1V 150W: 3.5V 200W: 4.0V
The chart is somewhat approximate but power increases exponentially in decibels, just use the closest number. Actual voltages with a maximum 0 dB digital signal would be 10X higher.
- If you're below the voltage listed for your amp, you should not be clipping with music. (Even though every AVR I ever saw tested drooped its power with all channels driven, someone here posted data that even on big peaks the surround channels were below the front LCR power. And music is not continuous tones).
- Above the voltage, you might be clipping.

Any comments/corrections please fire away!

*This is the hard part. The Stereophile 1st CD track 27 might work https://www.discogs.com/release/3692778-Various-Stereophile-Test-CD2
If someone could generate and upload suitable test tone(s) you would be the thread HERO!!
300 or 400 Hz at -20 dB is probably good since that will go into the woofers but without huge excursion, and also I dimly recall most voltmeters should be OK at those frequencies even if they are not good across the whole audio band.

 

[Scytales]

I find the idea relevant to have an hint of possible clipping.

I would add a scale of the actual difference in dB between each levels (in V RMS) in your charts :

2.5 dB less<=0.8 dB less<=1.1 dB less<=1 dB less<==0.9 dB less<==1 dB less<===1.2 dB less<=
30W: 1.5V __50W: 2.0V___60W: 2.2V ___80W: 2.5V__100W: 2.8V__120W: 3.1V __150W: 3.5V __200W: 4.0V

The reasons is the following. In certain circumstances, a PCM digital source can put out an instantaneous analogue voltage which corresponds to a digital level above 0 dBFS. Moreover, the 0 dB SA-CD (normally aligned on the 0 dBFS in PCM) is defined as a level -3.10 dB under the actual possible peak level of a DSD signal. For those reasons, it's a good idea to keep a 3 dB headroom (or more) in the analogue domain to cope with possible peak levels above 0 dBFS (or 0 dB SA-CD).

 

[Sokel]

If someone could generate and upload suitable test tone(s) you would be the thread HERO!!
300 or 400 Hz at -20 dB is probably good since that will go into the woofers but without huge excursion, and also I dimly recall most voltmeters should be OK at those frequencies even if they are not good across the whole audio band.

Here you go:

WeTransfer | Send Large Files Fast - Up To 2GB Free

The simple, quick and secure way to send your files around the world without an account. Share your files, photos, and videos today for free.

we.tl

WeTransfer | Send Large Files Fast - Up To 2GB Free

The simple, quick and secure way to send your files around the world without an account. Share your files, photos, and videos today for free.

we.tl

(generated by Multitone Analyzer,thanks to @pkane )

Edit: Added and a 16-bit one for the friends that wants to play it as a CDP would.

 

[Hayabusa]

Indeed weird that not more amps have a proper clipping indicator...
Cost: one op-amp and a LED

 

[KSTR]

Any comments/corrections please fire away!

Work on your math ;-)

Why not simply use P=U²/R?

EDIT: And the whole procedure relies on a lot of assumptions. To see if you're clipping your amp, get a cheap analog oscilloscope off of ebay/fleamarket etc. Or, even simpler, just listen ;-)

 

[Head_Unit]

the following.

In certain circumstances, a PCM digital source can put out an instantaneous analogue voltage which corresponds to a digital level above 0 dBFS.
@ Yes, I saw Thomas Lund of TC Sound give a talk about that from his AES paper. Unfortunately it didn't seem like many hardware makers allowed any headroom at least in automotive. Fortunately IIRC those "legal overs" are rare as they are confined to the higher frequencies. Moreover, the 0 dB SA-CD (normally aligned on the 0 dBFS in PCM) is defined as a level -3.10 dB under the actual possible peak level of a DSD signal.
@ Eh? I was not aware of that. Ah to be sure I understand the maximum level out of DSD is actually then "+3.10 dB"? But if say the CD layer of a hybrid disc puts out 2V with a 0 dB PCM tone, the output would be 1V with a "0 dB" DSD tone?

 

[Head_Unit]

And the whole procedure relies on a lot of assumptions.

Work on your math ;-)
@ ah quite possibly, could you be more specific?
@ I did say the chart is approximate, the voltages are all rounded since [see below]
Why not simply use P=U²/R?
@ I'm not sure what you mean here...that IS actually how the "chart" is generated
To see if you're clipping your amp, get a cheap analog oscilloscope off of ebay/fleamarket etc.
@ The idea of this thread is geared towards ordinary folks who wonder if they needed a bigger/outboard amp and who are NOT going to get an oscilloscope from a flea market. With music those are only good to see heavy clipping; you'd need a rarer and more expensive storage scope to look for momentary clips. (Well, if the scope has movable grid lines you could clip heavily and move the marker lines to that point and then zoom in)
@ You're definitely correct about assumptions, the procedure I drafted just gets you close-for certainty you'd need a scope as you say.
@ I had looked for a storage scope app like last year but didn't find anything that grabbed my attention.
Or, even simpler, just listen ;-)
@ Sure but it is hard to be definitive if the AMP is clipping or the SPEAKER distorting.

 

[Scytales]

@ Eh? I was not aware of that. Ah to be sure I understand the maximum level out of DSD is actually then "+3.10 dB"? But if say the CD layer of a hybrid disc puts out 2V with a 0 dB PCM tone, the output would be 1V with a "0 dB" DSD tone?

The 0 dB SA-CD and peak output DSD signal are defined in the Sony+Philips document "Super Audio CD Audio Signal Properties", version 1.3 from March 2003, annex D, §§ D.2 and D.3, which are mandatory to the format.

The document also contains recommendations (annex E). One is that bi-format players (CD+SA-CD) should put out the same level for a 0 dBFS 16 bits PCM CD signal and a 0 dB SA-CD DSD signal (§ E.1). But this recommandation isn't normative in order to comply to the SA-CD specifications.

 

[Zapper]

Work on your math ;-)
@ ah quite possibly, could you be more specific?
@ I did say the chart is approximate, the voltages are all rounded since [see below]
Why not simply use P=U²/R?
@ I'm not sure what you mean here...that IS actually how the "chart" is generated
To see if you're clipping your amp, get a cheap analog oscilloscope off of ebay/fleamarket etc.
@ The idea of this thread is geared towards ordinary folks who wonder if they needed a bigger/outboard amp and who are NOT going to get an oscilloscope from a flea market. With music those are only good to see heavy clipping; you'd need a rarer and more expensive storage scope to look for momentary clips. (Well, if the scope has movable grid lines you could clip heavily and move the marker lines to that point and then zoom in)
@ You're definitely correct about assumptions, the procedure I drafted just gets you close-for certainty you'd need a scope as you say.
@ I had looked for a storage scope app like last year but didn't find anything that grabbed my attention.
Or, even simpler, just listen ;-)
@ Sure but it is hard to be definitive if the AMP is clipping or the SPEAKER distorting.

"Storage oscilliscope"? That's a blast from the past. In the analog CRT days there was a specialized CRT that could retain the glowing trace on the phosphor for a period of time. But digital scopes are all "storage oscilliscopes" in that they can save a waveform indefinitely after a single trigger. And nearly all scopes are digital now.

Even with a digital scope detecting clipping accurately is not simple. You could monitor the supply voltage and the outputs but you would also need to know the minimum (complex load and frequency dependent) voltage drop across the output transistors. That's for a Class A-B amp. For a Class-D you would have to probe the switching node before the LP filter and look for max or min PWM duty cycle. Or one could look across the inputs of the differential feedback amplifier (for any class of amp) and look for non-zero transient voltages, indicating a loss of feedback control of the output.

 

[Schollaudio]

I use a scope and look for the onset of the signal hitting the power supply rail voltages.

Also, some amps have limiters.

 

[mhardy6647]

Yes, I'd (also) say an oscilloscope is the best way to establish the clipping level of an amplifier into a given load, FWIW.

 

[Scytales]

Of course, but the whole point made by Head-Unit is the give anybody who don't have a scope a rough idea about possible clipping in his power amplifier.

 

[restorer-john]

Of course, but the whole point made by Head-Unit is the give anybody who don't have a scope a rough idea about possible clipping in his power amplifier.

A 'rough idea' is about as useful as a dog with no legs.

Clipping varies enormously depending on the frequency, duration of the stimulant waveform and the connected load characteristics.

The only way to accurately determine what an amplifier is capable of is to extensively test the unit, with appropriate test equipment.

 

[Scytales]

Of course, but the goal of Head_Unit's proposal is obviously not to determine what an amplifier is capable of, but to help people become aware of the fact that his amplifier might sometime enter into clipping by listening his music because said amplifier may be under-powered for the job .

As I understand it, it's a simple eye opening exercise, not an absolute metrological measurement.

 

[DonH56]

If you're savvy enough to look at clipping, why not build or buy an attenuator network (use an H pad so ground is isolated) to feed the line input of your sound card and use the ADC in it as a 'scope?

As for me, I bought a fairly inexpensive Digilent USB-based DSO/digitizer kit several years ago, and still haven't used it. Work so often interferes with Life...

Many, many years ago I built a simple op-amp based clipping indicator with an attenuator and potentiometer in a little project box to use in testing my own and friend's systems. Ended up generating a little money making them for others. Not terribly accurate since it was voltage based; I pulse-tested the amplifier output into a dummy load to set the threshold when the output waveform compressed. I used the same scheme on the input of the amplifier, setting the (much lower) threshold based on the input signal that made the output clip.

 

[BDWoody]

A 'rough idea' is about as useful as a dog with no legs.

"Honey, is it your turn to take the dog for his drag?"

 

[mhardy6647]

"Honey, is it your turn to take the dog for his drag?"

This quite literally made me LOL -- to the extent that Mrs. H asked suspiciously "What's so funny?"

... and here I was trying to think of a way to work this tasteless old joke (see below) into this thread when I saw @restorer-john's apt simile!
You motivated me.
Here's my situational re-write.

Q: What do you call a dog with no legs whose owner takes him out to the ocean?
A: Bob.

 

[Zapper]

If you're savvy enough to look at clipping, why not build or buy an attenuator network (use an H pad so ground is isolated) to feed the line input of your sound card and use the ADC in it as a 'scope?

As for me, I bought a fairly inexpensive Digilent USB-based DSO/digitizer kit several years ago, and still haven't used it. Work so often interferes with Life...

Many, many years ago I built a simple op-amp based clipping indicator with an attenuator and potentiometer in a little project box to use in testing my own and friend's systems. Ended up generating a little money making them for others. Not terribly accurate since it was voltage based; I pulse-tested the amplifier output into a dummy load to set the threshold when the output waveform compressed. I used the same scheme on the input of the amplifier, setting the (much lower) threshold based on the input signal that made the output clip.

The way to do a simple circuit to detect clipping without characterizing the amp: make a differential amplifier with adjustable attenuation on inverting input. Split the power amp input signal and apply to non-inverting input. Apply power amp output to attenuator on inverting input. Play the power amp into speaker and adjust attenuator for minimum ac output of diff amp. The attenuated power amp output should almost match the input giving a good null. Now the diff amp will give a large signal when the power amp clips and its output no longer tracks the input signal accurately.

 

[solderdude]

Indeed weird that not more amps have a proper clipping indicator...
Cost: one op-amp and a LED

There is a bit more involved than that. For instance at which levels the clipping indicator should light up. Certainly with amps using a non regulated supply rail voltage this can vary. Also there should be a minimum LED duration (so requires additional parts).

So using a continous tone and fixed load to determine clipping might show lower clipping levels than when using music and real speakers. Granted.. differing only a few dB at most but at loud levels could make a difference. Think: NAD power envelope and alike circuitery and power supply rails sagging much under continuous loads.

Clipping, often, is not like digital clipping and can sound quite different. Also, when an amp clips the distortion of the speakers themselves will also be high at those levels (with powerful amps).

@pma already has a thread including clipping measurements. The behavior will differ with different design amps.

It would be a good thing to test for that (showing scope images of bursts under various loads). That is a LOT of extra testing and not your pre-programmed test sequence with standard loads.

 

[Schollaudio]

There is a bit more involved than that. For instance at which levels the clipping indicator should light up. Certainly with amps using a non regulated supply rail voltage this can vary. Also there should be a minimum LED duration (so requires additional parts).

So using a continous tone and fixed load to determine clipping might show lower clipping levels than when using music and real speakers. Granted.. differing only a few dB at most but at loud levels could make a difference. Think: NAD power envelope and alike circuitery and power supply rails sagging much under continuous loads.

Clipping, often, is not like digital clipping and can sound quite different. Also, when an amp clips the distortion of the speakers themselves will also be high at those levels (with powerful amps).

@pma already has a thread including clipping measurements. The behavior will differ with different design amps.

It would be a good thing to test for that (showing scope images of bursts under various loads). That is a LOT of extra testing and not your pre-programmed test sequence with standard loads.

People should consider clipping under dynamic conditions vs steady tone. Music can have a dynamic range from 6 to over 20db. Music signals reading 10 watts on a analog power meter could produce peaks that clip a 200 watt amp. It's not a bad thing to have power amps with supply rails over +-60V.