This forum is run by the BATC (British Amateur Television Club), it is service made freely available to all interested parties, please do not abuse this privilege.
To save you having to wait months for the next issue, here's some snippets that might be of interest:
I have been using a SR of 150 ks/s from the ugly mode output of the RPi at 17.5 MHz, and the signal is then cleaned up by a wide B/W crystal filter, and downconverted to a final centre frequency of 150 khz. After a gain block, the signal powers a 1.5W red LED.
The receiver head uses a photo-diode in voltage mode (ie, feeding a very input high impedence buffer). Frequency compensation follows (not nrcessary in current mode), and the signal is then mixed up to 146.5 MHz and fed into the MiniTiuone receiver. The estimated range when using A4 size Fresnel lenses is over 30 km - I know, I should confirm that with a real life QSO...
The circuit you linked to has the photo-diode working in current mode, ie, a near zero ohm input impedance (by virtue of the op-amp feedback). I haven't had chance to compare the sesitivity in this mode against the voltage mode version used to date, but hopefully will have done so by the time the next edition of the mag comes out.
Everything is fairly basic so far, and there's lots of room for experimentation/improvement. Also, since the results already look encouraging, increasing the symbol rate would be worthwhile - I have just been going for maximum range to date.
Thanks for your reply and the info.I would just mention that there is an ON Semiconductor chip called LA72912V which is an FM Modulator/De-Modulator chip, you may well be aware of this; it looks promising.It depends how you are doing the modulation. I had considered using a laser unit for transmitting but, looking into it, there appear to be some difficulties with this, at least with direct modulation.So I shall stick to LED's, much simpler.
To avoid clogging up this forum perhaps we could communicate by email; firstname.lastname@example.org. Look forward to hearing from you.
I'm hoping there might be others interested in optical datv - so let's stay here!. I had been meaning to start a thread anyway after the CAT18 talk.
Modulation is effectively AM, since for RF use you are wanting to use the minimum bandwidth possible and I can't see any reason to move away from that in the case of optical DATV.
i have been having more thoughts about this; the circuit in the link keeps the video signal at baseband analogue and, as you rightly point out, it is subject to all sorts of optical and electromagnetic interference . I have a couple of fluorescent lamps in my workshop and they create havoc. So I am now looking into the possibility of using pulse width modulation. In theory it sounds simple, a high frequency ramp generator, a bespoke comparator chip followed by a simple circuit to splice the waveform, and a driver for the LED. I suspect also that a low pass filter would be necessary after the video amplifier. One possible difficulty would be obtaining the high frequency ramp with sufficient linearity. How fast do you think I would have to sample the video waveform? Any thoughts on the this would be appreciated.
The issue of receiced optical noise is really only a problem for audio baseband work, say up to 5 or 10 khz. With a datv carrier frequency at 150 khz, it can be high pass filtered out of the rx passband, thus becoming a non-issue.
Digital modulation removes any requirement for the transmission system to be linear, as it is being turned fully on or off by square waves. However the digital method does require the system to handle higher frequencies. I hope that G4HJW will be able to test the digital method shortly to see how they compare. I haven't used any optical devices yet, so this is somewhat theoretical, but MiniTiouner can receive the RF test signal on the bench.
The test system is effectively digital QPSK. 4 digital carriers are generated inside the Raspberry Pi Zero. These are 90 degrees out of phase with each other. QPSK works by modulating 2 bits at a time (a symbol) so each pair of bits selects one of the 4 carrier phases and outputs a few cycles of this on a digital port. This happens at the symbol rate, so the phase of the digital output can potentially change for each new symbol.
The test system in the photos was using 4 cycles of 1MHz carrier per symbol, which gives a symbol rate of 250kS. The spectrum of the RPi output is effectively 'ugly mode'. This doesn't matter at the transmitter because it is being driven with square waves, but the received spectrum is likely to be something similar, rolling off as the frequency increases.
At the receiver, the carrier phases could be recovered and the 2 bits representing each phase be applied to a standard QPSK modulator to be received by MiniTioune. I think it would be preferable to treat the received signal as analogue (a 1MHz carrier) and mix it up to a frequency that MiniTiouner can receive and let MiniTioune decide what the 2 bits are.
On the bench, I've applied the low pass filtered RPi output and a locally generated signal 90 degrees out of phase, to a standard QPSK modulator acting as an image cancelling mixer. MiniTioune has received this OK.
The current test system (not the one in the photos) is using a carrier frequency of 651kHz. The harmonic content of the RPi output is so great that just by connecting a piece of wire to the RPi output pin, MiniTioune can receive a signal around 437MHz on the 671st harmonic!