[BC] Modulation measurement

[Cowboy]

On Friday 20 February 2009 06:55 am, Goran Tomas wrote:
>  Can anyone help me in determining what is the required peak response
>  time (attack) for valid modulation measurements according to the FCC
>  rules?

 Depends what you really want to measure, and do you *really* want
 to know ? ( since you can be liable if you *do* know )

 Used to be that whatever hit 75 kHz deviation was defined as 100%
 modulation, UNTIL someone came along and made arguments about
 that whole 1ms peak duration thing, and what is "frequent occurrence"
 anyway ?

 Now, 75 kHz is still 100%, but peaks less than 1ms in duration don't
 count ( presumably they can be 1 MHz deviation and cause no harmful
 interference under the rules ) and you can have less than 10 of them
 per minute ( presumably 9.9999 per minute, every minute ) and THEY don't
 count.

 Used to be, the FCC would measure with a calibrated scope across
 a discriminator. Response time, zero. Ballistics, none.
 Peak flashers were built to respond to near zero attack, and stretch
 the pulse to light a flasher. The length of time the flasher was illuminated
 being irrelevant. A peak in excess of 75 kHz deviation had occurred.
 "Frequent recurrence" was in the eye of the beholder. ( the inspector )
 The meter movement never counted for anything anyway, so the
 ballistics there were irrelevant. Merely an indicator, but not ever a
 calibrated indicator. Just a relative approximation.

 Today, nothing so simple and accurate meets the rules.
 Today, you still need a zero attack time, but now you need to analyze
 what caused that attack to see if it counts under the rules, or not.
 ( based on it's individual peak duration, and how many peaks of potentially
 varying duration occur within any rolling minute )
 It appears nothing short of a zero attack time, very fast DSP can possibly
 analyze the waveform fast enough to know whether it's modulation
 at all, or just something that doesn't count. Even then, it needs to store
 that information for the duration of the rolling minute, to count ( or not )
 against anything that might happen in the immediate ( minute ) future, or
 has occurred in the immediate ( minute ) past.
 Further, to what precision to we resolve a minute ? Which minute ?
 What's the duration of the peak ? Does a peak HAVE a duration ?

 SO, the answer to your question, is that you'll be MUCH better off
 by NOT asking the question !
 The technology to insure compliance with the rule simply does not exist,
 and in theory, can not exist in the real, physical world.
 Proving compliance is not possible, just the same as proving 
 out of compliance.
 ( although compliance with the old rule will probably be assumed
 to comply with the new, at some expense in "loudness" )

 What wonderful things this "mod-minder" thing did for us !
 Just because we can, does not mean we should.
 Apparently, Mr. Small didn't think this one through,
 as clearly ( IMHO ) we should never have started down this path.
 It's placed us squarely between a duck.

-- 
Cowboy