What is clipping?

[michael hurd]

Here is a link to an explaination of exactly what clipping is, and why it is bad. A clipped signal contains a DC portion, where the cone / diaphragm does not move or change direction.

http://www.prosoundweb.com/studyhall/studyjump.php?pdf=clipping

You will need Adobe Acrobat to view the link.

[lwhaples]

Michael, Thanks for posting the link.

Larry

[djk]

I am tired of the urban myths about DC and HF harmonics.

That guy means well, but he doesn't have a clue.

High frequency drivers are damaged during clipping by the increase in average power (heat), nothing more (mechanical damage can also occur with 6dB/oct crossovers). The same damage will occur from an un-clipped signal from a larger amplifier.

Low frequency drivers are ruined during clipping by mechanical damage caused by asymetrical signals generated by the time constants in the feedback loop and power supply. These signals are usually below 8hz, but they are not DC. Crown has a circuit in its largest MacroTech amplifier called a loudspeaker off-set eliminator that is designed to prevent this. It behaves like kind of a DC servo and subsonic (infra-sonic) filter. Asymetrical input signals can damage woofers from off-set even if the amp is NOT CLIPPING.

I had input on the JBL Tech Note: Danger, Low Power. At the time I thought it was correct. 30 years later the ony part of that note that I consider correct is the part pertaining to tape lifters during RWD and FF.

Rane has a good note about clipping, and ESP has a good note about asymetrical signals producing woofer off-set (leading to mechanical failure).

[dragonfyr]

Low frequency drivers are ruined during clipping by mechanical damage caused by asymetrical signals generated by the time constants in the feedback loop and power supply.

And from an earlier post:

I agree, using a resistor and plotting the impedance vs frequency will tell you everything you need to know.

'j' numbers convey nothing of use for the average person involved with audio. Knowing what they are and how to use them has little or nothing to do with whether of not you can build a good crossover for a speaker system.

Ironic these two statements coming from the same source!

So which is it?? Are they important and his earlier statement nonsense? Or are they completely superfluous and this latest comment nonsense? I guess that either he is not able to comprehend the significance of some information as he is "average", or he simply thinks everyone else is too stupid to understand! In any case, who cares!

Michael, I guess that you have no valid reason to be looking into this and attempting to gain insight! []

Or you can simply rest contented knowing that you are not "average"!

But I certainly appreciate your continuing efforts to share the information with us! Thanks!

[michael hurd]

I have copied and pasted the article by Rane, it is also in adobe acrobat, so you do need ther free reader to view it.

rane clipping.pdf

[Duke Spinner]

Geez, DraGon...............[:#]

you irritable today ...????

DJK is a Great source of info ....

[8-|]

we're lucky he stops by now and then .....

[dragonfyr]

djk wrote:
I agree, using a resistor and plotting the impedance vs frequency will tell you everything you need to know (about impedance).

'j' numbers convey nothing of use for the average person involved with audio. Knowing what they are and how to use them has little or nothing to do with whether of not you can build a good crossover for a speaker system.

---

Drop by the Measuring Impedance thread and find out just how insightful and knowledgeable someone who is "such a great source of info" can be!

Heyser could have stopped after he took his first AC circuits class! After all, that is more then ANY crossover designer needs to know! Of course some have the audacity to aspire to something more than mediocrity!

Pardon me if I prefer to try to learn from someone who has just a little broader range of knowledge and just a little higher goal!

[Duke Spinner]

well ....I'm not Always right ....

no-one can be right all the time ....[:^)]

[sunburnwilly]

5 yard penalty ?

[Speedball]

Thanks Michael,

I heard what clipping sounds like on my RF7's. At mid volume (maybe 60db) on a few ocassions there seemed to be a buzzing/vibrating sound coming from the speakers. If the volume is turned down the buzzing/vibrating will cease.

......strange...[8-)] Time for a big amp.

[PrestonTom]

Michael,

Thanks for the post. However, I am still confused about where the DC might be coming from.

Let me digress. When I was new to DSP, I frequently would make a waveform (in floating point arithmetic) that would exceed the number of bits (when converting to integer arithematic) during the DAC conversion. It's a rookie error, and everyone in the lab was guilty at one time or another.

We would always catch the error when we hooked up the spectrum analyzer. If we were generating a sine wave for instance and went beyond the 16 bits, the signal was definitely clipped. The tops and bottoms (an AC waveform) were both flattened out. This I believe is the same thing that happens with an amp that cannot drive the transducer "far enough" (perhaps a lack of sufficient / transient levels of current - if the driver has a low impedance). I assume the effect is comparable if the driver excursion is impeded in a mechanical fashion.

The reason I mentioned the specrum analyzer is that the effects of a clipped waveform were easily seen. There was harmonic distortion at the odd harmonics. With a Fourier analysis this would also be expected.

My question: if the clipping is for both the "tops" and the "bottoms" of the waveform (and clipping from the mechanical limitations may not be the same) and the clipping in both directions is the same magnitude, then where is the DC coming from? Additionally, the harmonics introduced were usually 10's of dB down so I can't imagine that any extra energy going to the tweeter would be appreciable. Certainly the signal sounded like crap, but I don't see where the damage would be coming from.

Is it the case that the clipping is asymmetric (more flattening in one of the directions). Then certainly a DC component would be added.

Curious,
-Tom

[djk]

Maybe we should dig up PWK out of his box and let you read him the riot act.

PWK didn't even own a calibrated mic when he did his first designs. He said he could either buy a mic or a generator, so he bought the generator.

The Cornwall is an interesting speaker. Designed without benefit of T/S parameters, FFT analyzers, or the anechoic chamber built around 1980; it sounds pretty good. The Chorus suffers by comparison (the Chorus II is better). PWK had all the goodies when doing the Chorus design, what happened?

Just because you have a PHD in electical engineering and/or $1M+ in lab gear, don't think it will be duck soup to produce a good sounding speaker design.

Unless your amp has a problem it is not going to put out DC, even driven into clipping.

Many McIntosh amplifiers are transformer coupled, they cannot pass DC. Yet they can and do exhibit dynamic off-set problems when driven with asymetrical music signals and/or are driven into clipping.

Ported speakers have long been known to have dynamic off-set problems, much more so than sealed designs. A ported design can exhibit dynamic off-set even when driven with a sine wave. Yamaha and MB Quart have patents on port designs that try to reduce this. Many speaker manufacturers play with spider design to try and tame this.

At the risk of repeating myself; amplifiers clipping do not produce DC, and DC is not the problem.

[Spkrdctr]

What happened to our other more detailed thread on clipping? Was is lost in the change over to the new format? Anyone know?