1

I am measuring EMG signals using ADS1299, and I have problems removing/eliminating the 60Hz noise in my measurements. I measured the EMG signal in differential mode, and I also incorporated a driven right-leg circuit to eliminate the 60Hz noise, but there are still some (a lot of) 60Hz noise in the received EMG signal.

I know that the 60Hz noise is mainly picked up from the power line, and I wrapped the wires and the power cables with aluminum foil sheets, but it didn't help much. I've also learned that human body acts like antenna and will receive/be coupled with that 60Hz power line noise, but I thought that because I am measuring the EMG signal in differential mode (use two electrodes, a "+" and a "-" to get signals), the 60Hz noise from human body should have been eliminated already. I also checked the contact between the electrodes and the skin, and I am pretty sure there wasn't any loose contact. I tried both dry electrodes and wet electrodes, but I have the 60Hz noise in both cases.

I don't understand why I still have the 60Hz noise in my EMG signal. I cannot simply apply a notch filter to filter out the 60Hz thing, since 60Hz is within the EMG signal frequency range, and I don't want to lose any original EMG signal.

I was wondering where exactly has the 60Hz power line noise been picked up? Does the 60Hz noise mainly come from the electrodes, human body, ADS1299 itself, wires, or nearby devices?

Thanks!

Natalie
  • 13
  • 1
  • 4
  • Can you test it? For example short the electrodes together with no human body, if you still get noise it's not the human. – user253751 Jul 10 '18 at 22:58
  • `I wrapped the wires and the power cables with aluminum foil sheets` .... did you ground the foil? – jsotola Jul 10 '18 at 23:00
  • 1
    What is the CMRR of the cables ? CMRR of INA or diff.amp? . Resistor tolerance error? RLD amplitude of 60 Hz ? Stray unbalanced leakage to ground ? to nearest 60Hz E-field? Your RLD is not working well enough. Something is unbalanced? – Tony Stewart EE75 Jul 11 '18 at 00:42
  • If you want see where it is pickup from clip scope probe to foil wrapped around hand, with probe ground disconnected then move other hand towards source. You ought to see 100Vpp now redo your CMRR on cables and amp with CM gain . I expect you are looking at 0 to 500uV signals and 10 to 50 uV noise and want better results but just guessing. – Tony Stewart EE75 Jul 11 '18 at 00:50
  • @immibis I shorted the electrodes together, but I still have the strong 60 Hz noise. – Natalie Jul 12 '18 at 14:28
  • @jsotola Yes, I collected EMG data when 1) no aluminum foil applied 2) aluminum foil applied and connected to Driven Right-leg electrode 3) aluminum foil applied and connected to the board ground (I am using ADS1299 development kit to check the signal). When the aluminum foil was connected to the board ground, the 60 Hz noise has been improved a tiny bit (but only a tiny bit). In general, I have a lot of 60 Hz noise in all three cases. – Natalie Jul 12 '18 at 14:32
  • @TonyEErocketscientist I am using ADS1299 development kit to check my signal. When you are talking about the CMRR of the cables, are you talking about the wires connecting the development kit and my electrodes, or are you talking about the cables connecting the computer and the development kit? – Natalie Jul 12 '18 at 14:38
  • @TonyEErocketscientist Could you please clarify your second reply a little bit more? I am still confused about how to test to see where the noise has been picked up. – Natalie Jul 12 '18 at 14:39
  • common mode noise must be rejected by balanced impedance on each shielded pair . The residual CM noise is then inverted to cancel on RLD. If the inputs are not balanced R and shielded pairs then you will see induction noise. This is why they use INA’s and shielded pair. Sometimes the shields are driven with CM signal (non-inverted) to eliminate the the current. Or “Active guarding” – Tony Stewart EE75 Jul 12 '18 at 14:49

0 Answers0