Third harmonic distortion

"Sy*phil*is Allison"

And SINAD can not be applied outside of communications? You have limited abilities.

... John

ses

MS

d,

D,

If you don't mind a freakin' breadbox, the 8903B is a nice box to have too. While the DSA goes for about a penny on the dollar, they still get money for the 8903B. Mine was $200 a few years ago. I see there are no serious sellers on ebay for the box at the moment, just overpriced "buy it now" peddlers.

Incidentally, there is a hack for the DSAs to compute group delay. It was in one of the app notes, not in the manual. Unfortunately, I don't have it handy. It was done with about 5 simple math functions.

e."

THD

t uses

e RMS

ield,

Es,

HD

up

INAD,

Not a problem. I just thought you were indicating what I said wasn't correct.

I was trying to think of a case where THD+N would not be SINAD. Can't the harmonics mix to form intermod? I've only seen intermod with a twin tone test, so if an amp produces intermod with just one tone, I suspect it is buried in the noise.

Tell that to Agilent. They put the term SINAD in their distortion analyzer datasheets.

"John S*it for Brains S*****ad"

Just to prove it can be done.

Wot a retarded f****it.

.... Phil

Intermodulation products of harmonics are also harmonics.

Cheers

Phil Hobbs

That was actually a quote from Wickipedia that I posted. I have since come to disagree with the Wicki assertion because I cannot get the equations to agree. Mathematically, I can get an agreement between the equations this way:

SINAD+1=1/(THD+N)

To me, it appears that there is no direct relation as asserted in the Wiki article. As a result, I consider my HP 331A instrument to be a THD+N instrument and not SINAD after all.

Have I made sense?

Amazing! I was just thinking the same about you.

... John

Thanks for the steer. The next thing on my breadbox list is probably an HP 3577 200 MHz network analyzer, but I also need a proper optical table, some office furniture, and a nice big whiteboard.

Cheers

Phil Hobbs

"Phil Allison" "John S*****ad"

Tell that to Agilent.

They put the term SINAD in their distortion analyzer datasheets.

Now for the missing CONTEXT:

-------------------------------------

From the HP8903A Operating Manual, page 1-3:

" 1-9. Transceiver Testing

The Audio Analyser has several measurements and features specifically designed for TRANSCEIVER testing.

It has SINAD and signal to noise measurements for receiver testing .... "

Go away you stupid jerk.

.... Phil

When a single tone is used and the amplifier generates harmonics, any intermodulation products are at frequencies which are integer multiples of the fundamental, i.e. they fall on directly generated harmonics, so how are you going to determine which is harmonics of the fundamental or intermodulation products of the harmonics :-).

Of course, if you expect e.g. only odd harmonics (symmetric clipping) and you still observe even harmonics, then you might expect intermodulation between harmonics. However, even in this case, a more believable explanation would be that one stage produces symmetric clipping and an other asymmetric clipping.

ate."

s THD

-it uses

rue RMS

field,

EEs,

.)

THD

ds up

SINAD,

own

Yeah, you got me there. So it seems you can't get an artifact that isn't an harmonic with a single frequency test source.

Now there is thermal distortion. For those not familiar with the term, that is where the distortion is caused by heating of the output stage as the signal swings. That is the parameters (parametrics?) of the output stage device are changing as the signal changes.

E.pdf

. =A0"

Bitch slapping Phil once again.

RMS

SINAD,

You can, e.g. a parametric amplifier, where there's an instability that leads to a pump frequency f_p spontaneously giving rise to a signal frequency f_s and an idler frequency f_i = f_p-f_s, but it isn't that common. You can get limit cycles in peak detectors and delta-sigmas, which also aren't harmonics. (Peak detector limit functions tend to cluster around subharmonics.)

That'll cause beaucoup IMD, for sure, including intermodulation of the noise with the sine wave signal. I'd expect the noise sidebands to roll off at modulation frequencies comparable to 1/(2*pi*tau_thermal), where tau_thermal is the dominant thermal time constant in the output stage.

Thermal stuff is less likely than low frequency circuits to show inverse-polynomial-ish rolloff behaviour--there's a useful rule of thumb in thermal diffusion, namely that for every factor of e the amplitude rolls off, you get another radian of phase shift. That keeps on growing, unlike a rational function rolloff, and so it sets a hard limit to the bandwidth of temperature control loops, i.e. you can't make a thermal control loop arbitrarily fast by adding higher and higher derivatives to the control law.

Cheers

Phil Hobbs

-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics

uses

RMS

SINAD,

communicate."

THD

uses

RMS

field,

EEs,

up

SINAD,

Complete word salad. What's the difference in the output spectrum of the two cases?

Cheers

Phil Hobbs

For any N, sums of and differences between multiples of N are also multiples of N.

I wonder if you wanted Phill Hobbs to say overtly what the "also harmonics" are harmonics of. It does appear to me that the "also harmonics" are harmonics of the same signal that the original harmonics are of.

Arrrgh...

SINAD-1=1/(THD+N)

( snip absurd, irrelevavant s**te )

F*ck off.