[Laser] pulsed laser comms

[glennt at charter.net]

Hi Art!

Uhhh... well, I agree that the tx & rx require strict synchronization and that adding successive samples can add signal - though you may want to subtract estimated noise from each sample to aid in the process. I suspect that we're either arguing apples & oranges or we are in "violent agreement".

The main advantage is that pulsed lasers usually run a LOT more (peak) power output than a CW laser. Thus it's not so much a matter of processing sophistication by the receiver as it is  brute force by the transmitter.  The SNR is improved by (say) 40dB simply because the transmitter is sending 40dB more power.

If the comparison is between CW and pulsed lasers at the same peak power level, then, yeah, my guess is that superpositioning will not work as well as an FFT, because the FFT exploits signal coherence while superpositioning doesn't. FFT also does not require any particular knowledge of the pulse timing while superpositioning relies on it.

Using the astronomical technique, what you'd do is look at a CW laser (NOT a pulsed laser)  with some kind of integrating detector (perhaps a CCD), compensate for estimated noise and then see how many photons you collected. That's not really a pulsed laser system, it's an integrating detector - an entirely different animal. The "pulses" then are really samples from the detector integrating over some time. How well this can work depends not only on how much energy the detector collects, but also on how accurately you can estimate the detector noise.

The FFT process is somewhat similar in that you estimate that the noise is of a different frequency than your desired signal so you can subtract it off. FFT does nothing to remove noise that is about the same frequency as the desired signal.

73 de Glenn


---- Art <KY1K at verizon.net> wrote: 

> 
> Almost Glen..............
> 
> You do not look at the signal during a time when you think it is 
> present although there are gated receivers.
> 
> You do use time synchronization on both the receiver and the 
> transmitter. But, the sampling is quite constant. The kicker is that 
> you do not throw away the samples that should contain a weak pulse 
> buried in noise (as you do with fft). Instead, you average them, and 
> over a given number of samples, the desired signal begins to become 
> visible. In this manner, the noise averages out so long as you can 
> maintain time synchronization.
> 
> I don't know if it can work any better than FFT.
> 
> Yves says he tried it and it doesn't do nearly as well as FFT.
> 
> Regards,
> 
> Art
> 
> _______________________________________________
> Laser mailing list
> Laser at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/laser