If you ever want to try this I have a few 560 LED panels, 850nm 20 degree, capable of putting out about 60W CW or more pulsed. You'd also need some microlens sheets to reduce vertical divergence to maybe 10 degrees or less - I may have some left.
I'm dubiuos that such an AM signal could be resolved above all the interference though.
Thanks for the offer Tony.
I'm dubiuos that such an AM signal could be resolved above all the interference though.
The actual interference present on the RF carrier impressed on the infra-red should be very slight - the receiving set would only be sensitive to that lying (say) between 41 and 48 MHz. This is a very narrow-band signal w.r.t the infra-red carrier.
However, as you say, this is AM, so any other disturbance, almost from DC up, will have an effect on reception. The set's agc will help here, but nevertheless (for example) there will be disturbances caused by rising warm air. If this 'rippling' of the image of the emitting array, as seen at the receiver, simply shifts its apparent position, then the field of view of the receiving lens (or lens set) should take care of this. However, when the emitter appears so small it appears point like, an effect like 'twinkling' of the stars sets in, resulting in signal 'flutter'.
I have worked out that for an emitting array 2 metres across, this would occur by around a distance of 14Km (8 miles). This is at or beyond the optical horizon so may not get a chance to kick in. Again, my simple calculation and underlying assumptions on request.
Steve
Car failed NCT (MOT) and been sick. But I will look at this.
Good luck with this Michael - you'll need it! I too will be having a go over Christmas. The fastest LED's I have seen have a rise time of 'only' 20nS so they will struggle to run at 45MHz. Much faster photo-diodes are available though. The capacitance of the devices could be a problem.
Steve
I've driven 1500pF gate SMPSU fets to over 20MHz & 20W. Driver had 50 Ohm to 7.5 Ohm transformer.
This is also why I will ignore high power devices but use arrays of regular ones. Up to 100MHz carrier should be possible. I'll measure the capacitance and do a model first.
very high speed detectors are no problem.
Still very ill!
Dannk Doo Trebor
NCT (MOT) retest late tomorrow (20th), hopefully well enough to drive by then.
Hi Trevor,
I too am still beavering away on this... it's proving difficult... teaching myself right now about 'transimpedance amplifiers' (tips welcome). The photodiode I have been using is a Hamamatsu S5973, which is very fast. I shall try to attach its data sheet here, but I am not sure if .pdf's are 'postable'. I am using Channel E2 in colour from my WC-01 as a source.
Steve
A long time since I've played with this sort of thing but nowadays a simple approach might be to connect it directly to the inverting input of a suitably wideband opamp.
This is exactly what I was doing last night. AFAICS though, the op-amp's feedback resistor - and thus the gain - has to be fairly low at these frequencies to cope with the stray capacity.
They certainly weren't around when I last used photodiodes. Transimpedance is simply a posh way of saying convert current to voltage with gain. The opposite of transconductance or gm.
So does that mean an emitter follower is a tranconductance amplifier? Sorry if I am making a boo-boo but I clearly have a lot to learn here.
Hi.
I've reduced the operating voltage of the laser and an modulating the laser with 1v p-p signal, there are chokes and other filtering in the supply. Ill order a photo diode from Farnell.
Even though the capacitance of this photodiode is low, the operating load resistance also seems to need to be low to work successfully at this frequency, with a consequent penalty in voltage output. Thus the 'transimpedance amplifier' - an op-amp with the two input terminals representing a short circuit.
How are you modulating your laser in a linear fashion? If a laser diode is what you are using, these are delicate things and modulating circuits also need good soft-start protection (normally built in to their attached powering boards).
Steve
As an aside, if we can develop this technology, requiring a really linear method of modulating lasers or (fast) LEDs at high frequencies, I can't see any reason why the whole 'radio' band shouldn't be available on the light carrier from a given station simultaneously. 1500 metres anyone? Book your wavelength now! 
Steve
Mustn't keep replying to my own posts... but I've just seen the first inkling of a result!
The photodiode was connected straight across the input of the LM6171 op-amp with no other biasing.
Currently it's very insensitive - the photodiode had to be millimetres away from the LED and the 'picture' is a mere shadow, swamped by intereference from the RF leaking from the boards and cables. But it's a start... I suppose!
Steve
(Edit) I've increased the value of the op-amp feedback resistor and have now got a real picture, albeit at a range of three inches, and crawling with interference and hum. Placing a sheet of paper in the way cuts off the picture, demonstrating this is not due to stray RF pick-up.
An embarrassing update... I believe the results this morning were false! What seems to have happened is thiat the photo-diode amp was picking up rf leakage direct, but only passing it through when the photodiode was irradiated with light. Still checking up on this...
So there's still everything to play for. Let's see how Trevor gets on - and Michael too?
Steve
I have now improved the screening to prevent such 'false positives' in future. The limiting factor now seems to be the speed of the LED. Data on the speed of LEDs in general has been hard to come by, beyond the occasional mention of a rise-time of 20nS on some products. This is not fast enough.
However, I have found a resonant cavity LED intended for use with plastic optical fibre, which is fast enough. Five are now on their way to me from Japan.
http://www.hamamatsu.com/eu/en/product/ ... index.html
This LED is red. I am sticking to visible light rather than infra-red for this first experiment. A visible beam will be easier for setting up the optics and should give greater safety from accidental direct exposure into the eye.
Steve
That is a very posh LED.
It looks expensive too.
I have now improved the screening to prevent such 'false positives' in future. The limiting factor now seems to be the speed of the LED. Data on the speed of LEDs in general has been hard to come by, beyond the occasional mention of a rise-time of 20nS on some products. This is not fast enough.
However, I have found a resonant cavity LED intended for use with plastic optical fibre, which is fast enough. Five are now on their way to me from Japan.
Excellent work.
That's why I plan to try miniature low power IR LEDs in a an array, each driven separately. Certainly the emitters used for optical digital links should be fast enough.
Screening and separate power and no earths is certainly needed for tests.
I've finished writing for now and health is a little better. Yesterday I got about 1/2 way through clearing up workshop.
Certainly the emitters used for optical digital links should be fast enough.
Standard method of moving HD digital video around 'long' (i.e around 100m) distances in television studios, at 1.5Gb/sec.
Hi.
No success here at all up at 45 mhz but at 1mhz it makes a cracking pantry TX even with an led from a GU10 lamp! Works from the shed to the house beautifully at night. I used the photodiode a lens and an op amp to feed to a radios aerial socket, a lash up but perfect quality once aligned.
Yes, one of the benefits of this idea is that the whole radio spectrum is up for grabs! Then over again and again - for each 'transmitter' and of course also for point-to-point links.
My tests so far, and yours too it seems, indicate that the average LED tails off in its response beyond a megahertz or two.
As for these special LEDs. I shall certainly make available any surplus to others. The more brainpower and enthusiasm we have behind this project - the better. The LEDs are a special import and I won't have them before early February. I have no prior experience of such 'resonant cavity' LEDs. It seems they are a halfway house between standard LEDs and Laser diodes.
Steve
Transmission of audio by laser between the two cathedrals in Liverpool.
http://hopestreetproject.blogspot.co.uk ... oject.html
Till Eulenspiegel.
Hi Trevor,
Progress at this end remains difficult. To date (according to the scope) I've obtained an adequate - if reduced - response with the emitter on Channel E2 (vision: 48.25MHz) but nothing over a few MHz at the receptor end.
I show the two circuits I am using for the emitter and the receptor. The transistors are shown by their RS part numbers. I have been careful to minimise capacitance in the construction. Electronics design has never been my strong point so corrections or suggestions for improvements are welcome! It would be helpful if others engaged in this project also would post their circuits here.
Trevor, I would also be happy to pop one of those 'fast' LEDs in the post to you - free - if you would like it. The same offer applies to Michael too, if he feels this part would be useful.
Cheers,
Steve
Trevor, it's FOC only, on the understanding you'll be 'open-source' with your circuits like I have been. This is administratively simpler. PM me your address and it will be done. 
Steve
I wasn't able to send an attachment with the PM back to Trevor, so here is the pinout for the L10762 LED, bottom view. Pin 1 is positive, pin 2 is negative. Current must be limited to an average of about 20mA.
It's important that it's connected the right way round, since more than 3v in reverse bias could wreck it and I won't be sending out replacements! 
This diagram is also on the data sheet but is so tiny it's hard to see.
Steve
PS.
WARNING. THE LIGHT SOURCE ON THIS LED IS PINHOLE-SIZED AND VERY INTENSE. NEVER LOOK AT IT DIRECTLY OR YOU COULD RISK YOUR VISION.
