# Improving the quality of an optical audio transmitter

I wired up a very simple device in which I use an audio signal transformer (EI14) to modulate the intensity of a laser beam (a cheap 650nm, 5mW diode) according to the audio output of my phone. The laser is then received by a photo-resistor (G5528 A205) wired up to the microphone socket of my laptop. With this setup I can transmit audio between the two devices.

Not surprisingly, the results are not astonishing. The audio quality is less than crystal clear, but that was expected. Here's an example.

My main concern that this system may be fundamentally limited is that it's not the data that I'm sending, from which the audio could be recreated; what I'm sending is just some real-time 'amplitude'. I'm actually surprised how well this worked since my intuition is that music consists of a range of Fourier frequencies and each point in the duration of the song comprises of some combination of those frequencies at corresponding amplitudes. At the same time, with a monochrome beam I'm limited to just varying the brightness of my beam and hence only able to send one amplitude at any given time, i.e. only the information about one of the Fourier frequencies (presumably the dominant one?) can be transmitted. Is this reasoning fundamentally correct and does that mean that a monochrome beam can't faithfully transmit a complex sound in real time?

My other questions are: what improvements to my very simple setup are possible to make the audio quality better? What are the limiting factors? Is it even possible to get a reasonable quality in such a system?

• only the information about one of the Fourier frequencies can be transmitted... that is incorrect ... what you are sending, in effect, is the speaker cone deflection amplitude. ... it is being sent to a speaker with a mass-less cone. the signal contains multiple frequencies at all times (except if you send a pure sine wave, of course) – jsotola Jan 29 '18 at 22:13
• use optical filter over the receiver that excludes all light frequencies except the laser .... run it in total darkness .... convert analog signal to digital – jsotola Jan 29 '18 at 22:17
• @jsotola Oh I see! I was wondering what it actually is that is being sent; I was dubious about my theory because how on earth would the decision about which frequency is being sent be made. I did run it in the dark but it did not seem to make much difference. I will think about the filter though! Much appreciated. Could you expand on analogue to digital conversion? At which step would that conversion take place? – Piotr Jan 29 '18 at 22:19
• @Piotr There's no such thing as "which frequency should be sent". An audio signal is just amplitude vs. time. Frequencies are something you can think about when analyzing audio signals but they aren't actually the signal itself. If you think of a speaker - there isn't a separate speaker for every frequency, there's one speaker that moves in and out. (Higher-end speakers may have two, that are optimized for different frequencies, but again, that requires some kind of signal processing to make that happen) – user253751 Jan 30 '18 at 0:00

I'll just point out one contributor to your audio distortion:

Figure 1. Extract from the PGM 5506 datasheet.

You didn't supply a link to your LDR datasheet but it's probably similar to the one above. As you can see the response times are in tens of milliseconds so at best they can only respond reasonably well to about 20 Hz. Anything above that will be "slew-rate limited" meaning that a step change in light intensity will give a response much like a resistor-capacitor (RC) charge / discharge curve.

Figure 2. RC charge / discharge curves for a squarewave input signal. Source: EEEGuide1. The LDR will give a similar response although the lower trace may be inverted depending on how you've wired it to the receiver input.

You might need to advance to a photo-transistor or photo-diode solution.

One other thing: you should have a DC decoupling capacitor between your receiver and the mic input of your laptop. This will remove any DC component and may prolong the life of your laptop.

• Thank you for a great answer. I'm very fresh when it comes to electronics; if I switch to a photo transistor, such as this one, which has 3 orders of magnitude faster response and should cover audio frequencies just fine, do I need to worry about the amplitude of my received signal? Will the sound card in my laptop do the job of amplifying the input or do I need to do this myself? – Piotr Jan 30 '18 at 20:15
• You have plenty of work to do. You have all sorts of problems to resolve including preventing saturation of the phototransitor at high laser power, etc. I haven't built anything similar but you should find some help on the web. – Transistor Jan 30 '18 at 20:21

I think the problem may have more to do with non-linear behavior of the laser diode than anything else. I believe you would need to apply some audio compression to the signal that is modulating the laser diode, in order to get a more linear output from it.

• Do you mean that if I knew in what way exactly my signal is non-linear, I could correct for that by skewing my modulating signal? That does sound interesting! – Piotr Jan 29 '18 at 22:20
• More or less, yeah. An LED has a characteristic non-linear V/I curve when its forward biased. Its light output is a linear relationship to current, but not voltage. You're putting a linear audio signal (voltage) to it but the voltage produces a decidedly non-linear current in the diode - it basically saturates very quickly, effectively clipping the peaks of your signal. Do a little research on audio compression circuits. – Norm Jan 29 '18 at 22:38

All the distortion is likely in your modulation of nonlinear current with a transformer instead of a linear Voltage controlled current source