EEVblog Electronics Community Forum
General => General Technical Chat => Topic started by: daqq on January 29, 2017, 05:55:26 pm
-
Hi guys,
The NVidia busts Moon conspiracy ( https://www.eevblog.com/forum/chat/nvidia-debunks-apollo-conspiracy/msg1122656/#msg1122656 (https://www.eevblog.com/forum/chat/nvidia-debunks-apollo-conspiracy/msg1122656/#msg1122656) ) got me thinking - to get a laser signal all the way to Moon and receive it you need a lot of big hardware - a powerful laser, a respectable light sensor and a good light gathering apparatus (telescope).
But that's for a single, powerful laser pulse as I understand it. What about using a modulated laser pulse, modulate it in a similar way GPS signals are transmitted - their power received on Earth is below the thermal noise for the receiver, but using correlation you can know when you are or are not receiving them - it's an amazing trick to get out of the noise and also determine the phase of the signal.
How far above noise could one get this way? With a little over a second in c between the Earth and the Moon, it should be possible to send a fairly large block of data and do an auto-correlation of the received values.
Still, a fair amount of apparatus would be needed of course, but this might lighten the requirements for such devices.
-
They have a few retro-reflectors on the moon and they regularly use them for a kind of bounce back. From a rather powerful laser pulse they get back 0 - 2 photons in the right wavelength window on a sizable telescope. With those short pulses (< 1 ns) there is not much of correlation one can do. It is more of having a short acceptance window (e.g. < 1 µs) at the right time after you send the pulse and than measuring a lot of pulses to get enough statistics.
-
Hi Kleinstein,
Yes, I know how they do it currently - short, powerful pulse, small window.
My question was whether a lower power modulated beam stretched into a longer time might yield the same or better results. If you'd send a 1 second long pulse, modulated with a PRN sequence, you should be able to receive some of it back and get a higher correlation with the delay applied.
-
Spreading out the power over a longer time will reduce the SNR. The problem is the background light, that gives a significant amount of shot noise. Otherwise one might get the same SNR, but no better. The idea if using short pulses is to get as little of the wrong light as possible. A longer pulse would only work if it could be used to go for a smaller wavelength window.
In theory there might be a choice in a kind of interferometer, but this would require a stable distance on the 100 nm scale over longer time. This does not work because of the atmosphere.
-
A well chosen pseudo random noise sequence should always improve the SNR as it is lengthened. Otherwise you would have to believe that the interfering light is somehow correlated with your sequence. Which would be more interesting than any other result.
-
It's not quite that simple, because you can only send 'positive' light pulses - you can't transmit dark. Codes do exist (Golay codes, simplex codes) but the coding gain is limited to about 10dB or so at the most, compared with just sending more pulses and averaging their returns.
-
It's not quite that simple, because you can only send 'positive' light pulses - you can't transmit dark. Codes do exist (Golay codes, simplex codes) but the coding gain is limited to about 10dB or so at the most, compared with just sending more pulses and averaging their returns.
Interesting, thanks. I don't suppose that modulating the light not from 0% to 100% but rather from 50% to 100% to gain an effective 75% "DC" value could offset the lack of "negative" pulses?
Still, 10dB is quite a lot when you think about it - that's 10 times less power, or a 10 times less smaller area telescope.
-
The size of the telescope has a big influence, at least in the smaller range: A larger mirrors allows for better focusing. So you gain from better focused light send, better focus on receiving (thus better ration of send light to background) and the larger area.
Even with just pulses one uses a kind of correlation: in a first step one gets returned light only in a moderately short window (e.g. a few µs). Later, with more knowledge about the distance one can make the window small (e.g. 100 ns range) and reject even more photons received at the wrong time. One also needs the very fast modulation / fast raise and fall times to get good time resolution.
It is only with the high peak power of a pulsed laser that we can compete with the intensity of background light. It really helps if more than 10% of the light received in the timing window comes from the laser and not just background.