Amp01 Guard Drive Circuit

Hi Reddy,

Is there any reason why you are looking at AMP01? This is a really old part, you may be better off using a more recent device, like AD8422 or AD8421. If you can provide some of your requirements, I can make a better recommendation.

For the time being, I will say that you need to connect all the solder links. The dashed lines need to be connected too, but the resistor values depend on the gain you need.

Regards,

Gustavo

Hi Gustavo,

Thanks for replying,

I am working on Data Acquisition of  EEG signals. These signals will be in the frequency range of

1Hz to 3kHz and  in the Voltage range of 0.1uV to 1uV.

Please guide me to Acquire these signals by eliminating Noise.

Suggest me to choose proper Guarding Techniques and the instrumentation amplifiers for this application.

In the Amp01,  the gain of the amplifier is adjusted by varying the Resistor values Rg and Rs. But in the Guard Drive circuit, how  exactly the gain is going to vary with Resistors R3, R4 and R5. 

Thanks,

Thejesh

Hello Thejesh,

I still believe that for an EEG application, you should consider a newer amplifier like AD8421 or AD8422. You can still implement a common-mode shield as with AMP01. Also, these inamps only require a single resistor to set the gain, so it would be a little simpler to select the value of Rg. You can derive the shield driver input as shown in the AMP01 data sheet or do it as shown in the AD620 data sheet (page 16), which can be found here

http://www.analog.com/static/imported-files/data_sheets/AD620.pdf

The only thing you need to do is to divide by 2 the desired value of Rg and connect the input of the shield driver to the center connection. Note that the output of the shield driver has a 100Ohm resistor; you'll need these so that the opamp can drive the capacitive load of the cable and to protect the opamp output. A more modern choice for the shield driver could be a FET-input opamp such as the AD4062-2.

An alternative is to buffer the inputs with FET input buffers, like the AD8244. In addition of providing high input impedance, they are specified to deliver good matching and high CMRR. If you have a shield for each input wire, you can use the output of each buffer to drive it. If you only have one shield per conductor pair, you can either use one of the two or take the average by connecting both buffer outputs through equal resistors to the shield. Keep in mind that in this case, the assumption is that the differential voltage between the two inputs is small and it won't be requiring a lot of current flow between the two resistors.

I hope this helps,

Gustavo

Hello Gustavo,

Thank you for your suggestions, I will try the configurations you have suggested and seek your guidance.

I have two more queries:

1) There is a DC offset in the signal (the Dc offset is 600mV )

2) The DC offset keeps on changing. (the Dc offset shifts in the course of Data Acquisition for constant input)

Thank you again for your help.

Regards,

Thejesh

Hello Thejesh,

Is this with the amplifier inputs shorted or do you have something else connected? Have you accounted for electrode offsets? This is a common issue with biopotential measurements. If the offset comes from the electrodes, it would also get amplified.

Regards,

Gustavo

Hello Gustavo,

Thanks again for your support.

Yes I checked, the offset is coming from the electrodes, when it is connected to body. Is there a way to control this offset ?

As you suggested, I have procured these two Eval Boards for testing AD8422/21,

1)EVAL-INAMP-62RZ

2)EVAL-INAMP-82RZ

INAMP-62RZ as the guard Drive which I am going to test. I don't know which connectors to use for interfacing EEG Electrodes to Amplifier Inputs(+IN, -IN). Can you give us some idea on this.

I am attaching the Pictures for electrodes and for  EVAL Boards.

Regards,

Thejesh

Hello Thejesh,

You can only use EVAL-INAMP-82RZ to evaluate AD8422 or AD8421. The EVAL-INAMP-62RZ board will work for AD620 inamp family only. The eval board for AD8244 (quad buffer) is not released yet (this buffer was released only a few weeks ago!). The signal inputs are clearly indicated on the board as IN+ and IN-, as described in the user guide:

http://www.analog.com/static/imported-files/user_guides/UG-261.pdf

You'll also need a common bias to ensure that the signals are inside the common-mode range of the amplifier. This is usually done through a right-leg drive circuit (RLD) or simply by biasing each terminal to a bias voltage referenced to the amplifier ground. You can find implementation examples in the AD8232 data sheet:

http://www.analog.com/en/specialty-amplifiers/instrumentation-amplifiers/ad8232/products/product.html

The offset variation is usually slow. Traditionally, the technique to eliminate electrode offsets is by placing a high-pass filter following the instrumentation amplifier. The corner frequency of the high-pass depends on your design requirements, but typically is placed <0.1Hz.

Regards,

Gustavo