[Laser] PWM of LED in QST

[n5gui at cox.net]

Comments within your message below:

---- Tom Becker <GTBecker at RighTime.com> wrote: 
>   Equally obvious to me is the notion that an AM signal must exhibit 
> variable power, lest there be no information conveyed by the mode.  

Indeed you are correct, an unmodulated carrier conveys no information, other than its existence.  However, variation in power that is necessary for one to send a 300 Hz tone, a 500 Hz tone, or a 700 Hz tone need not exhibit average power fluctuation if low pass filtered below 10 Hz.

Remember that I am using "AM" in the sense equivalent to radio emission Double Side Band Full Carrier.  It seems to me that you are thinking in terms of Single Side Band and Double Side Band Suppressed Carrier.  SSB and DSBSC are "efficient" in that when no information ( in this case audio frequencies of varying power ) is being transmitted, there is no power sent out.  DSBFC sends out full power when no information is being transmitted.  When information is sent, a scaled copy of the audio power information is sent out in the upper side band, and another in the lower side band.  The power in the carrier is correspondingly reduced.



> If it is average LED power that is the medium (opposed to frequency  
> or phase) then the power must change - and will be visible for the 
> low-frequency signal components.  The LED power is, in this instance, 
> changed by duty cycle.

Yes the power is changed.  But the average is not.  There are no signal components at low enough frequency for the eye to see the flucuations.  The audio is filtered well above 10 Hz.


> 
> Since this scheme uses duty cycle - the LED on:off ratio - as the medium 
> to convey the instantaneous 10kHz-sampled input signal voltage, a signal 
> that spends some time below zero, represented by a lower duty cycle, 
> will appear momentarily dimmer - due to the retinal integration - during 
> that period, and it will appear brighter for excursions above zero when 
> the LED On period is greater, a higher duty cycle.  The eye's 
> integration performs the same function


The eye does not follow the 120 Hz flucuations of light fixtures.  Why should it detect flucuations in the output of an LED when it is driven by 10 KHz pulses modulated by audio signals that are higher in frequency than the lights in the room?


>                                                                    as, I assume without seeing the 
> schematic, the receiver's integration of the incoming variable-width 
> light pulses, perhaps a simple a low-pass RC, to reconstruct the 
> transmitted audio before amplification to a speaker.  This is the same 
> effect as a motor, whose rotating mass integrates variable-width PWM 
> pulses to produce some degree of instantaneous torque.
> 
> Can you build the circuit portion you are concerned with, to verify 
> either your prediction or his reporting?  It does not sound complex and 
> could be breadboarded easily, I suspect.

I have not bothered to breadboard the circuit.  I have spent more than enough time with the PWM light output of my Ramsey LBC6K transmitter and my own home built LED AM transmitter ( electret mike into LM324 quad op amp audio amp / bandpass filter followed by a voltage to current converter which drives the LED ).  They don't flicker when properly operated.

I note that the author of the QST article did not report the quality of the audio received from his prototype.  I doubt that he had the capability to receive the light and convert it to audio.  I do not fault him since he does not claim any expertise, or for that matter any experience at all  with light communication. 

I do not doubt that the circuit that he built does flicker at the drive level the stated in the article.  I predict that if he cut the drive level in half, it would probably stop flickering.  And I would say that he should be able to hear a definite improvement in the audio quality.


James
 n5gui