Hi all,

I am trying to figure out how to make a circuit that will prevent a higher voltage AC signal from going into a differential amp input. This signal is NOT of interest but will occur in 100ms intervals followed by 400ms of the much smaller AC signal of interest.

I will try to give some background without cluttering the question. The AC signal of interest is from a receiving EMAT coil (inductor coil). It is approximately 20-50nV pp AC (at 4 kHz frequency). I am obviously using extremely high gain from the differential amp(s) to get this signal to microcontroller level. The only issue is that for 100ms previous to this tiny signal, there will be a much larger AC signal (generated by the transmitting EMAT coil) that will surely drive the amp into saturation and is NOT of interest.

The timing is controlled by a PIC microcontroller so I realize I need to make some switch that will block the larger AC signal from the differential amp's inputs.

The first thing that comes to mind is having the microcontroller hooked up to the gates of some transistors to allow the AC signal to reach the amp's inputs. I just don't know the proper configuration for manipulating an AC signal with transistors. I am generally worried about component selection here since it is already a difficult enough task to cleanly amplify the 50nV signal. I wouldn't want transistors (or any switch component) to introduce much noise/distortion/resistance to the signal when on.

Also to give more background, here is the differential amp circuit I am currently using. The only difference with my circuit is that I am using resistors to set the gain to ~500,000 so I can easily see the signal with my scope.

Any ideas/suggestions are appreciated!

 

What is the amplitude and frequency content of the unwanted signal?

A couple of cascaded JFET switches might be a good solution, but beware of the switching signal feeding through -that might take filtering to remove it.

 

Thanks for the input. The unwanted signal will be approximately the same as the wanted i.e. 4000Hz. I am not sure of the amplitude yet but it is likely to be much much bigger maybe 1000 or 1 million times bigger.

Also pardon my novice understanding, looking at the figure in my post would I insert the JFET switches before both points V1 and V2?

 

What is the maximum voltage you can except?
The wanted signal is about 50nV I am afraid that any electronic component placed in front will produce more then 50nV noise.
If you are able to accept 500mV two high speed clamping diodes opposite parallel and extreme low noise resistors in serial are probably possible. (those resistors are expensive)
Contact TI, NPX or an other component (diode) producer.
But measure noise before you implement it.

Picbuster

 

Yeah I definitely was concerned with noise issues. I've already brainstormed some options that don't include blocking the input signal and perhaps the best I've thought of is using the TI chip LMH6629 which is an ultra low noise op amp with a power down function. I could use the microcontroller to power it down during the unwanted signal and turn it back on for the wanted part. I just posted on a TI forum to ask if the power down function was meant to by repetitively cycled/used like that. I just wasn't sure what would be the better option but I suppose if I can use the amp like that, it would likely be best considering noise is my biggest enemy for this circuit.

 

PCB layout and its Power, Reference and analogue net is your biggest concern.
What model of ADC are you going to use?
LT is giving an excellent design support to ease the noise pain.
I am very please with their vast response and very useful tips.

Picbuster

 

Hi all,

I am trying to figure out how to make a circuit that will prevent a higher voltage AC signal from going into a differential amp input. This signal is NOT of interest but will occur in 100ms intervals followed by 400ms of the much smaller AC signal of interest.

I will try to give some background without cluttering the question. The AC signal of interest is from a receiving EMAT coil (inductor coil). It is approximately 20-50nV pp AC (at 4 kHz frequency). I am obviously using extremely high gain from the differential amp(s) to get this signal to microcontroller level. The only issue is that for 100ms previous to this tiny signal, there will be a much larger AC signal (generated by the transmitting EMAT coil) that will surely drive the amp into saturation and is NOT of interest.

The timing is controlled by a PIC microcontroller so I realize I need to make some switch that will block the larger AC signal from the differential amp's inputs.

The first thing that comes to mind is having the microcontroller hooked up to the gates of some transistors to allow the AC signal to reach the amp's inputs. I just don't know the proper configuration for manipulating an AC signal with transistors. I am generally worried about component selection here since it is already a difficult enough task to cleanly amplify the 50nV signal. I wouldn't want transistors (or any switch component) to introduce much noise/distortion/resistance to the signal when on.

Also to give more background, here is the differential amp circuit I am currently using. The only difference with my circuit is that I am using resistors to set the gain to ~500,000 so I can easily see the signal with my scope.
View attachment 136574
Any ideas/suggestions are appreciated!

Have you looked at chopper-stabilized amplifiers?

http://www.analog.com/media/en/training-seminars/tutorials/MT-055.pdf