"StudioT"

The range of frequencies I want to transmit are 200hz to 3.2Khz
The frequency spectrum will have a lower sideband at 36.8Khz and an upper sideband at 43.2Khz. The technique will cause the amplitude of these sidebands to be lowered, additionally the frequency response of the transducer varies greatly within this frequency band.

Are you saying that the AGC will somehow compensate for this variation in frequency response?
I have attached the functional blocks diagram.

https://www.dropbox.com/s/rsxlso0g0ejait5/functional blocks.JPG

 

You should really post your diagram here.

You will need some buffer amps, for instance you cannot feed a mike directly into the modulator and the transducer will need a driver amp.
This driver amp is where the AGC is placed as a feedback control. This amp must have sufficient gain to bring the lowest (highest frequency) signal up to minimum and the AGC cuts back the gain to equalise the rest.

You can forma cowan modulator from a single transistor or a ring modulator from 2. There are probably enough components lying around in your department to do this. In superheterodyne terms you would call this a mixer.

The block diagram should look something like this. Perhaps sirch can offer some refinements.

 

What sort of range were you expecting with this system? The higher the frequency, the faster the sound attenuates in air (from what I could tell online). Which is why low sounds travel further. I wouldn't hope for more than a few meters or so, but who knows. Humidity, temperature, and pressure also play a role.

I would reconsider on-off keying and sending digital data. It may be possible to set up a simple serial bus over the air with that, and you could send any data you want. Real-time voice might be too much, especially for a school project. Start with something simpler, and work from that. Try sending text first, for example.

 

"StudioT"

Wow this really made it much clearer. I will most certainly look into AGC's and ring modulators tomorrow. So the AGC boosts the two side bands to compensate for the non linearity of the frequency response of the transducer?

I think I have the first few blocks covered, I think a pierce crystal oscillator circuit may be better as it doesnt require an op amp, this carrier frequency can be modulated with the signal through a simple transistor and transformer circuit. The output of this circuit should be m(t) x c(t) .. then it will be ready to go into the amplifier.

Regards

 

"Austin Clark"

Thank you for your reply, I am looking to operate at a frequency of approximately 40Khz ideally no more than this as you said the signal becomes attenuated at higher frequencies.

In terms of range, the aim should be as much as possible without errors and problems occurring. (2 metres is a good target)

And yes I have reconsidered on-off keying as I had great doubts as to whether the channel could handle the bit rate of real time voice. Also the techniques involved are more complicated..

Regards

 

So the AGC boosts .....

No the amp boosts the sidebands and the AGC cut back to standardise the output.

So you might present a gain of 1000 to the sidebands and 1 to the centre frequency.

 

In terms of range, the aim should be as much as possible without errors and problems occurring. (2 metres is a good target)

Error correction (a la internet) should be a part of the plan.

2 meters? That's pitiful. I can hear the festival in my town from 4 miles away and easily tell you what song the band is playing - if I know the song so that my brain can do the error correction.

 

I think those suggesting a heterodyne solution are on the right track, the naval underwater phones for communications with submarines use this principle. Works pretty well in air too.

EDIT Oops, should have read the later posts, seems the favoured solution

 

"StudioT"

Could a notch filter be used instead of an AGC to suppress the carrier frequency and allow the sidebands to pass - I have been looking at an Active Twin T notch filter.
Also could a simple common emitter amplifier circuit be used to drive the ultrasonic transducer?

Regards

 

1)You could suppress the carrier, but I would not bother, especially on your budget. It is causing you no harm. That is not the purpose of the AGC.

You still have not understood the relationship between the transducer response curves (both transmit and receive) and your input signal.

Let us suppose your voice signal has two components, of equal amplitude.

(I know it is a very stange voice signal but I am trying to emphasise a point)

Let one component be at 3khz and the other at 300hz.

What will be the output from the transducer of this combined signal?

Is this a faithful reproduction of the input? Remember you want good quality audio.

If the penny still hasn't dropped then definitely ask your tutor about this.

2)Transistors are cheap (less than a penny) and multistage circuits are easier to stabilise and get the right input and output conditions, gain etc.

Your next step is to define the (signal) voltage and current levels between each block.
This will define what each has to do and the details will then follow.

 

If this is not what you mean I dont know what you are talking about.

All i know is the variation in frequencies in the transducer are not reproduced linearly and the maximum power is at the centre frequency and it is weaker at either sides. Causing the side-bands to be attenuated by different amounts.

Did you get my private message with the schematic? I wont be able to define voltage and currents at each block at a precise level until I know the microphone we will be given, the operating voltages of the transducer etc.

 

I said I would post some references.

Ultrasonic transmission and reception : Practical Electronics May 1972 p374

Ultrasonic transmission and reception Elektor July 1977

Ultrasonic transmission and reception : Elektor August 1978

Ultrasonic Audio Sender : Electronic Today International July 1992 p50