Analog Front End for Capacitive Micromachined Ultrasound Transuder(CMUT)- AFE5807 or other device

if your output is current, 5807 may not be suitable. to answer your question, could you provide more information of your transducer output characteristics? such as output impedance, current noise, capacitance, and etc. then the opamp team can look for sutiable devices. 

Output impedance = 315KOhms,  Variable plate Capacitance =0.56pf to 0.95pf (Under maximum plate deflection), Noise current < 5pa/Sqrt(Hz). Kindly suggest me a suitable pre amplifier for the CMUT signal processing application.

Hi Xiaochen Xu,

Do you have any inputs for me? Typically the Capacitance Micromachined Ultrasound Transducer has high Output impedance and the capacitance variation produced with incident sound pressure is much smaller(<500fF) and the minimum change in capaciatnce that need to be preferably detected can be on the order of 0.5fF (femtofarad). Hence the application need a high precision analog front end pre amplifier with very low input referred nose to acquire a very low magnitude current output from the capacitive transducer. Kindly let me know a suitable Transimpedance amplifier or precision opamp to suit the requirement.

Haridas,

Can you provide a few more details?

What is the frequency range of interest?

Is the impedance source impedance (315k ohms) only due the the capacitive reactance of the transducer?

Is the transducer biased with a polarizing voltage or electret-type polarization to create a voltage output?

I will provide a general comment that relates to condenser type microphones, a similar type of transducer. Most use a single discrete FET to buffer the input signal. This approach allows selection of a transistor with very low capacitance and noise to match the application. The single transistor improves the noise performance compared to a differential input stage of an op amp. Since signal levels are very low, the extended linear input and output signal range of an op amp offer little advantage.

Regards, Bruce

Hi Bruce,

Thank you for the clarifications.

Frequency range : 5MHz to 50MHz

Yes it is the capacitive reactance of the transducer alone computed from (1/jwC)

Yes it is  polarized bias , +10V DC to the top capacitance plate and bottom plate is at substrate level, In the transmit mode, +10Vpp will over ride the DC level.

 Equivalent circuit of the CMUT transducer is modeled as a simple current source in parallel with a resistor and capacitor.  Most of the published work uses a transimpedance stage as a preamplifier and hence I was looking for a similar solution to suit my requirement. Thanks for the idea to use  discrete FET to buffer the input signal. I will surely look into this method and study the suitability in detail to meet the design requirements like noise performance, gain and bandwidth.

One of the IEEE publication and a PHD theses from the same lab (as in below link)  I referred has used a commercially available TIA for signal acquisition and most of the CMOS implementations traditionally use TransImpedance AAmplifier as a pre amp.

   http://smartech.gatech.edu/jspui/bitstream/1853/31806/1/guldiken_rasim_o_200812_phd.pdf 

Thanks again for your insightful recommendation which I will surely follow up and kindly let me know if there is any opamp configuration which is comparable to discrete transistor performance.

Hi,

A small correction in my previous reply. I mentioned the frequency range as 5MHz to 50MHz. There may be a shift in frequency for the individual transducer depending upon the medium and otherwise it is narrow band operation. 

Frequency band : Transducer 1 is at 7 MHz and Transducer 2 (smaller dimension) is at 50MHz.

Haridas,

This is an interesting application with many complex trade-offs.

The 10V poloarizing voltage is rather low. Output will be proportional to this voltage so there is much to gain if it can be increased. I understand that this may be a processing limitation.

There are two modes of operation that may be possible. You wish to operate in transimpedance mode where the voltage across the transducer is held constant and variation in capacitance causes a signal current to flow through the transimpedance resistor. The other mode of operation holds the charge on the capacitor constant and varying capacitance is allowed to create a voltage change on the output. I have never resolved whether there is a fundamental advantage in one vs. another. Professional condenser microphones use the second approach (at least all that I am familiar with).

I am concerned that the wide bandwidth that you require may place difficult demands on the op amp. With limited gain-bandwidth in the op amp, the summing junction is not truly held at a constant voltage and some portion of the signal is lost. Very wide bandwidth op amps may have other undesirable trade-offs.

I think there would be much to be learned by a thorough SPICE analysis of application--both approaches. Our free SPICE program, TINA-TI, would be a good start if you have not currently using a SPICE program.

If you wish to communicate further on this issue I think it best to take this off the forum. If you want to provide an email I will contact you directly.

Regards, Bruce.

Hi Bruce,

Thank you so much for your observations and valuable suggestions. As you recommended, I will do a spice simualtion to study the various options for a low noise front end pre amp. We will take the topic from the forum. Kindly send me your email id. I hope I can still seek your guidance on the topic

Thank you

Haridas