AD8421 LtSPICE

Hello,

I'm trying to perform a very simple LtSPICE simulation using the AD8421. I'm running into a weird issue that I can't figure out. The gain of the system is basically non existent. I've read that analog devices cannot confirm the accuracy of their products in LtSPICE. What does this mean? Does the circuit setup below look correct for a gain of 10? I've tried both the 8421 online model (shown below) and the model built into the "opamps"component section of LtSPICE. 

The voltage was measured at the output. The gain of the system is negative. For anyone not familiar with LtSPICE input syntax, the input voltage here is 0.1V P-P at a DC offset of zero and frequency of 200. I was thinking since the positive terminal amplifies in reference to the grounded negative terminal, the P-P voltage should increase by 10 at the output, and thus be 1V P-P.

NOTE I did not use the circuit on the left to take out an extra troubleshooting step. This circuit is per the datasheet of the AD8421 for properly creating a reference voltage. 

  • 0
    •  Analog Employees 
    on Jul 24, 2018 3:59 PM

    Hi Michael,

    This is one of the common issue that the customer encounter in using instrumentation amplifier.

    The problem arises because you do not have enough input common mode voltage especially if you are running in a single supply voltage. Please see this FAQ for more detailed discussion. It is recommended to set the common mode voltage center of your supply voltage to get the optimum voltage swing.

    For your simulation, you can either use dual supply probably +/- 8V or set the common mode voltage to 3.4V and see if you get your desired output range.

    You can also check this very helpful tool which validates your input conditions for functionality test of the inamp.

    http://www.analog.com/designtools/en/diamond/#difL=-0.05&difR=0.05&difSl=-0.05&gain=10&l=3.7&pr=AD8421&r=3.7&sl=3.7&tab=1&ty=1&vn=0&vp=7.4&vr=3.7

    I hope this helps.

    Best regards,

    Emman  

  • Again, thanks for the help. This revised circuit amplifies the gain as expected. 

    (a) This has given rise to another question. I'm looking to create a simple EMG circuit using an instrumentation amplifier and a bandpass filter. The inputs to the IN amp, positive and negative, would be a pair of bipolar electrodes. Wouldn't this common mode voltage be a risk to the patient? I remember reading that ISO standards limit the input voltage to something like 10 uA for a biasing circuit. 

    For example, see the attached circuit. The dual supply is nice, since it seems to eliminate the need for a buffer (I think?). However, the patient must be placed at "ground" which is actually 3.7V above his normal ground, causing a current to flow through him. Depending on the impedance path and a bunch of other factors, this voltage could easily cause the current to exceed 10 uA. 

    Some additional questions:

    (b) Does the "ground" voltage still need to be buffered to eliminate AC noise? 

    (c) What battery solution would be recommended for this type of device? I've been looking into rechargeable coin cell batteries for low power applications, such as a Panasonic ML920. 

    Thanks,

    Mike

  • 0
    •  Analog Employees 
    on Jul 25, 2018 8:10 PM

    Hi Michael,

    The common mode voltage is usually implemented by  having a third electrode and we usually put a series resistor at the output of the amplifier driving this electrodes to limit the current going to the patient of not more than 10uA at fault conditions. You can easily compute for the value of the series resistor by taking the supply voltage divided by 10uA.

    A coin cell battery is common for portable devices which could last from 5 to 10 years or more depending on the total current consumed by the system.

    You can also read this article which could help you in designing and finalizing your ECG/EMG circuit.

    ECG Front-End Design is Simplified with MicroConverterRegistered | Analog Devices 

    There are also a lot of solutions and discussions on EMG-related topic on this forum  which could help you and you can just search it by putting keyword of EMG on the search box. Here are some of the threads:

    http://ez.analog.com/message/48403#48403

    http://ez.analog.com/message/47968#47968

    http://ez.analog.com/message/62546#62546

    http://ez.analog.com/message/56319#56319

     

    I hope this helps.

     

    Best regards,

    Emman