Impedance mismatch issue in receivers

[nauman]

Hi all

I have an acoustic piezoelectric sensor whose output impedance should be ideally 50 ohm and a preamplifier circuit specifically designed for this sensor having input impedance of 50 Ohm. The piezoelectric sensor output voltage range is from uVrms to few mVrm. These low voltage signals are then amplified by preamplifier.

My query is that if there is a mismatch b/w output impedance of sensor and input impedance of preamplifier, what will be the side effects? In case of Transmission chain, it is obvious that such mismatch (e.g. b/w Power Amplifier and Transducer) causes power losses, however, what will happen in case of receiver chain when very low voltage signals are present ?

 

[berkeman]

If the preamp is specifically designed for it, why would there be an impedance mismatch? Can you post links to the datasheets for the transducer and preamp?

 

[nauman]

If the preamp is specifically designed for it, why would there be an impedance mismatch? Can you post links to the datasheets for the transducer and preamp?

Thanks for reply. Both transducer and preamplifier are own developed and are not COTs. Controlling the output impedance of acoustic transducer is quite challenging and main impedance deviation is in output impedance of transducer (>67 ohm instead of around 50 ohm).

 

[berkeman]

Do you have gain/phase plots of the transfer function from an acoustic driver through the (liquid?) medium to the output of the transducer? What is the frequency range of operation? Can you post a PDF or JPEG image of that gain/phase plot?

 

[nauman]

Do you have gain/phase plots of the transfer function from an acoustic driver through the (liquid?) medium to the output of the transducer? What is the frequency range of operation? Can you post a PDF or JPEG image of that gain/phase plot?

Frequency range of operation is from 90 KHz to 110 KHz.
As transducer is being used only in receiving mode (i.e. as a hydrophone), there is receive sensitivity vs frequency graph available indicating how much voltage will be generated at transducer output for standard input of 1upa sound pressure in water for a typical frequency range. If you are interested in this graph, i can post it.

 

[berkeman]

Great, that would be helpful. And have you also measured that same transfer function to the output of the preamp? That might start to show you if the impedance mismatch makes much of a difference.

 

[Tom.G]

Since your signal frequency is so low (110kHz), the signal wavelength in coax cable will be roughly 6880ft.

Using a 'rule-of-thumb' that a cable less than 1/10 wavelength long has minimal impact on a signal, I wouldn't worry about it until the cable length approaches 1/10 of a mile.

I would worry more about the capacitance of the cable. For example, using RG-58 coax at 30pF/foot, you can expect the following signal level losses due to the cable capacitance loading the transducer: (the 67Ω with the cable capacitance acts like an RC low-pass filter)

Length Loss
7ft . . . .1%
50ft. .. 10%

If you need a fast risetime waveform (square wave) at the receiver rather than a Sine wave, you need either a short cable or some active electronics (a Line Driver) at the transducer to keep those sharp edges. That's another case of the 'factor-of-ten' rule-of-thumb, if you need a square wave, for many uses you need a bandwidth of around 10x the pulse repetition rate.

Cheers,
Tom

 

[nauman]

Great, that would be helpful. And have you also measured that same transfer function to the output of the preamp? That might start to show you if the impedance mismatch makes much of a difference.

 

[berkeman]

Frequency range of operation is from 90 KHz to 110 KHz.

It looks like your plot is centered on 100MHz, not 100kHz. Am I misreading something?

 

[nauman]

It looks like your plot is centered on 100MHz, not 100kHz. Am I misreading something?

Sorry, frequency axis label is misleading (i.e. should be Hz instead of KHz), sensor is certainly centered at 100KHz

 

[rude man]

Since your signal frequency is so low (110kHz), the signal wavelength in coax cable will be roughly 6880ft.

Using a 'rule-of-thumb' that a cable less than 1/10 wavelength long has minimal impact on a signal, I wouldn't worry about it until the cable length approaches 1/10 of a mile.

I would worry more about the capacitance of the cable. For example, using RG-58 coax at 30pF/foot, you can expect the following signal level losses due to the cable capacitance loading the transducer: (the 67Ω with the cable capacitance acts like an RC low-pass filter)

Length Loss
7ft . . . .1%
50ft. .. 10%

If you need a fast risetime waveform (square wave) at the receiver rather than a Sine wave, you need either a short cable or some active electronics (a Line Driver) at the transducer to keep those sharp edges. That's another case of the 'factor-of-ten' rule-of-thumb, if you need a square wave, for many uses you need a bandwidth of around 10x the pulse repetition rate.

Cheers,
Tom

If he can make his transducer, cable chas. and load impedances roughly the same there is low or negligible capacitive load effect regardless of cable length; he just gets a gain of ~1/2 with pure delay.

Ideal cable assumed of course).

 

[rude man]

Or if you just pick a cable with $Z_0$ matching the transducer, then the receiver can be any impedance $Z_L$ , again with no attenuation or distortion, but with gain = $Z_L/(Z_L+Z_0)$.