Connect with us

Building an EKG analog front end circuit

Discussion in 'General Electronics Discussion' started by Md Omair Ahmad, Oct 24, 2016.

  1. Md Omair Ahmad

    Md Omair Ahmad

    3
    0
    Oct 24, 2016
    I am trying to reproduce an EKG analog front end circuit. I have attached the schematics of the circuit.
    In my design I have used instrumentation amplifier IC - INA128P and for quad op-amp IC - TL084CN.
    The problem I am facing is that, when I am trying to acquire the EKG signal on the oscilloscope i am getting pretty distorted signal and it is no way near ideal ekg signal. please help me out with your suggestions.
     

    Attached Files:

  2. Harald Kapp

    Harald Kapp Moderator Moderator

    9,218
    1,856
    Nov 17, 2011
    Welcome to EP.

    The main issue is: your supply voltaeg (+33 V) is too low for the amplifiers used.

    The INA128 requires at least +-2.25 V, the TL084 is specified for aminimum supply voltage of +-5 V.
    Use the original amplifiers as shown in the schematic (the MCP6004 can be used down to 1.8 V) or use a higher supply voltage, at least +-5 V.

    Plus there is another issue that may come to light only once the opamps work properly using the correct supply voltage: when using the TL084 and higher supply voltage, a re-scaling of the resistors is likely to be required, too. The input current of the TL084 is ~1000 times higher than the input current of the MCP6004. This will lead to significant voltage drops across the1 MΩ resistors in the feedback networks.
    Scaling these resistors will require chnaging the capacitors in turn to keep the frequency characteristic the same.

    All in all, better use the original opamps to avoid a major redesign of the circuit.
     
  3. Md Omair Ahmad

    Md Omair Ahmad

    3
    0
    Oct 24, 2016
    thank you harald kapp for some important advice.
    i am applying the power supply of 8.0V in this circuit. and i agree with your point that TL084 is drawing more current than MCP6004. currently i dont have MCP6004 with me so I am going with whats available. I am attaching the snapshots of the oscilloscope output when no limb is in contact of the circuit and other when limb is in contact with the circuit.
    your help is highly appreciated.

    after contact.jpg no limb contact.jpg
     
  4. Harald Kapp

    Harald Kapp Moderator Moderator

    9,218
    1,856
    Nov 17, 2011
    In that case you haev an open high impedance input which will pick up any stray noise from the environment, noticeably the 50 Hz or 60 Hz mains frequency.

    Some tips:
    • Check the circuit with short-circuited input leads.
    • Also use shielded cable to connect the electrodes to the input resistors, connect shield to ground )´(0 V) or virtual ground.
    • Keep all connections as short as possible.
    • Put the whole circuit into a shielded case (any metal box will do) to keep environmental nosie away from the sensitive circuit.
     
  5. davenn

    davenn Moderator

    13,260
    1,748
    Sep 5, 2009
    I assume you meant supply voltage of 3.3V ?
     
  6. Harald Kapp

    Harald Kapp Moderator Moderator

    9,218
    1,856
    Nov 17, 2011
    A good assumption - and a typo on my side.
     
  7. Audioguru

    Audioguru

    2,570
    580
    Sep 24, 2016
    Most ECG circuits have the patient connected to an inverted common-mode AC and DC signal that cancels most hum and muscle movement pickup like this:
     

    Attached Files:

  8. Md Omair Ahmad

    Md Omair Ahmad

    3
    0
    Oct 24, 2016
    Thanks again herald kapp and audioguru for your advice. I think my quad op-amp is drawing a huge amount of current so, instead of redesigning the whole circuit I guess I should change the quad op-amp with mcp6004.
    One more thing i would like to share, I am implementing this circuit on a bread board so that I can tweak my circuit anytime. Can breadboard be a problem for some of the noises?
    I am just a beginner and I desperately needs your expert advice. Please keep me posted,
     
  9. Audioguru

    Audioguru

    2,570
    580
    Sep 24, 2016
    No huge current.
    The input current of the TL084 opamp is very low because it has Jfet inputs. Its error with the 1M resistors in the circuit is only 0.4mV which is nothing. The MCP6004 has even lower input current that makes a difference if the resistors are 500M or more.

    Your problems are the solderless breadboard's rows of contacts and messy wires all over the place are antennas that are picking up mains hum and all kinds of other interference.

    A problem with using the TL084 is that its inputs do not work properly when they get within a few volts from ground in the circuit but the rail-to-rail MCP6004 inputs work fine all the way down to ground. If you added a negative supply voltage to the TL084 like all the circuits shown in its datasheet then the input voltages will never get near it and it will work like the MCP6004.

    The wires from the patient to the circuit are also antennas that pickup mains hum and other interference. These wires should be shielded audio cables. The shield blocks the interference.
     
    Last edited: Nov 3, 2016
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-