• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 46 subscribers
  • Views 2004 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

EMG Amplification Poor Signal to Noise performance in full wave rectifier

Chad Knutsen
Prodigy 140 points
Other Parts Discussed in Thread: TINA-TI, INA121

Hello,

I'm an engineering student but this is more of a personal project.  My grandma suffered a stroke recently and lost the use of her right arm.  She has very limited muscle contraction but enough that I think a myoelectric controlled assistive prosthesis would help in her recovery.  The circuit works as expected except at the full wave rectifier.  When I do the noise analysis in TINA-TI, the signal to noise ratio looks good at all stages except the final rectification stage.  The signal is going to a microcontroller so it needs to be a DC signal.  I've been working with a couple professors at school but haven't received any feedback on this issue.  I've setup the noise analysis as follows, frequency 1-500Hz, 500 data points, S/N Amplitude 1.5m.  I chose 1.5m for the S/N because I am figuring on amplifying a 1.5mV biopotential.   I guess I should ask if I'm doing the noise analysis correctly before asking about my circuit design.  Would that be correct for the S/N Amplitude?  At the output of the LPF I get a SNR of 33, but then after the rectification, at 150Hz the SNR is 0.731. I've attached the circuit that I used for the transient analysis that includes common mode, right leg drive, EMG potential, and DC offset, and I also attached the single source circuit I used for the noise analysis.  Please take a look and let me know if there is a better way of doing this. 

Click the picture to see the circuit

EMG Amplification Circuit

Thank you, Chad K

  

  • 0
    Guru 21975 points

    Chad,

    Your schematic is difficult to read due to poor resolution. From what I can see in the signal path, noise performance should be determined by the first stage, INA121, assuming you have enough gain in this stage. I can't decipher your S/N figures but concentrate on acquiring a clean signal in the first stage. With adequate gain in this stage, its S/N will dominate.

    There may be problems with doing a noise analysis through the FWR, I'm not sure. It, by definition, is not a linear process. The op amp must slew through a dead band around zero. This may present some analysis issues.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 140 points

    Here are the TSC files.  What does the S/N Amplitude mean in the noise analysis setup?  If I set it to 1.5m, does this mean my voltage source inputs a 1.5mV signal?  I notice that when I set the S/N Amplitude to 1.5m, my SNR at 150Hz is much lower that when I set the S/N Amplitude to say, 1.EMG AMP Circuit.TSC

  • 0 in reply to Chad Knutsen
    Guru 21975 points

    Chad,

    The S/N function plots the ratio (in dB) of the signal (the magnitude you specify in the setup of the simulation) relative to the integrated noise from the start frequency of the plot to the given signal frequency at which you are reading the plot. It's a bit obtuse. Anyway, when you supply a larger signal amplitude for the simulation the S/N will obviously improve.

    Your noise is dominated by the series input resistors which create about 110nV/rt-Hz. After that hardly anything else matters in your circuit. Furthermore, the macromodels for the amplifiers you are using do not model noise accurately.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 140 points

    The raw EMG signal amplitude before amplification will be about 1.5mV, and after the INA121 Gain of 30, it will be about 45mV, then 4.5V in final amplification.  I'm a mechanical engineering student so I don't have a really solid understanding but I'm very interested in learning as I go.  The final stage seems to rectify first and then amplify by 100, is that right?  So when I chose the S/N Amplitude, would I chose the expected input of 45mV.  Or would I chose 4.5V?   If I'm just overcomplicating this, it would be just as helpful if you just said whether or not this circuit is going to be too noisy to amplify the 1.5mV EMG signal. 

    Second, and last question.  What type of diodes would I use in this circuit?

    I really appreciate your help on this.
    Chad K

  • 0 in reply to Chad Knutsen
    Guru 21975 points

    Chad,

    You should not have any problems with noise. The principles of S/N mean that once you have developed significant gain in the circuit (the INA121), the noise of subsequent stages do not contribute any significant noise.

    The final rectifier stage performs both the gain and rectification, essentially in one step. I recommend signal diodes such as 1N4148. Do not use power rectifier diodes such as 1N4001.

    One final warning:  Be very cautious. Avoid using any power-line operated instruments or power supplies. I recommend that everything be battery powered. We don't want to fry Grandma. Commercial medical equipment uses galvanic isolation devices for safety when operating from AC power.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 140 points

    Thank you for the warning.  I did find that issue come up a lot in my research but there wasn't any focus on battery powered circuits or USB powered circuits, just AC.  As of right now, everything is battery powered.  Would I need to use isolation if I added a USB connection?

     

  • 0 in reply to Chad Knutsen
    Guru 21975 points

    Chad,

    If you run from a laptop under battery power, you can power from its USB port. I do not recommend running off the charger.

    Bruce