I am trying to design a PCB for amplifying the signal coming from a VS-900m which is an acoustic sensor. This signal is fed to an amplifier AD8551 to amplify it with 20, 40 and 60db.
The question is do I have to integrate an impedance matching circuit in that case? Knowing that have no idea about the input and the ouput impedance of the sensor and the amplifier since they are not mentioned in the datasheet.
I've already read that it depends also on the length of the PCB track and they shouldn't be larger than 10% of the wavelength. So, the question is how to know if I need an impedance matching circuit in my case ?
Impedance matching is usually only required to ensure maximum power transfer which may also have other implications.
The in/out impedance of the op-amp will be what you make it to be with your design.
The impedance of the VS-900m may not be explicitly stated but can probably be inferred from other specifications.
Usually, most sensors have a high output impedance and require an interface circuit with an equally high or higher impedance.
Its response is characterized by two peaks at 190 kHz and 350 kHz with accompanying anti-resonances at 200 kHz and 400 kHz.
How long is the wavelength of a 400kHz signal in copper? How long is your signal path?
If your VS-900m is on the same PCB as the AD8551 then don't worry about it unless your PCB is something like 70m along one side. If it is some distance away connected by wires then there's various things you need to take into account, including impedance matching, which I won't go into unless you need to know.
Yes both sensors and AD8551 are on the same PCB. So, no need to take into account the output impedance of the sensor and the input impedance of the amplifier ?
The frequency response is 100-900kHz, let's take 500kHz as the center.
The capacitance of the transducer is 540pF and assuming the impedance is totally capacitive we have impedance = 1/ (2πfC) = 1/( 2π*500kHz * 540pF) = 589Ω
That is fairly low. I would design the opamp to have an input impedance of at least 10 times that, 10K is common though.
My point was that you were very unlikely to be at 10% of the wavelength unless you have a record breakingly large PCB, it would have to be about 70m in one dimension.
I was sure, then I went out and started to think about it. I'd still say 'don't worry about it' but add 'too much'. @jim-p 's is a really good answer, go with that.
Some further thoughts from me:
You get maximum power transfer when the source and load impedance are the same. Think of the gears on a bike, they match the 'source impedance' of the power supply (your legs) to the 'load impedance' of the wheels, which varies according to whether you are going up or down hill etc. You use the gears to adjust how hard it is to pedal to get the best match to suit your strength and the terrain you are riding over.
If you have an amplifier with a higher impedance than the source then the result will be a higher input voltage to the amplifier, so you need less gain. Unfortunately that will be accompanied by more noise than there would be if the impedance matched. If you have a lower input impedance amplifier than the source then the input voltage will be lower so you need more gain. While the noise at the input will be lower the extra gain will amplify the noise as well as the signal, net result more noise. Impedance matching is what gets you the sweet spot of lowest noise to highest signal. The next question is to ask if it matters in your particular situation, my instinct is that it probably doesn't matter, hence 'don't worry about it'. All that said Jim's done the maths, go with what he said, the only point I'd disagree is match the source impedance, don't go higher, but you can build the amplifier, change the resistors and experiment and find out for yourself if it makes any difference.
The issue is there's no THT amplifier with that reference. I already bought it to test it but it is an SMD component that can't be experimented on breadboard!
Not sure why you chose that amplifier but the GPB is not high enough for your application. I would select one one with a minimum GPB of at least 10MHz.
Thank you for pointing this out! I just realized that for 400KhZ, to have a gain of 1000 without signal distortion, I need at least an amplifier with a GBP of 425Mhz!
Well I assumed you wanted to amplify the entire bandwith between 100kHz and 900kHz. In any case you would not amplify to 1000 all in one stage, so 425MHz is not necessary.
How do you know that you need a gain of 1000?
