Comments on Noticeable dip in amplitude response at 50 Hz when shielding circuit with driven right leg network.
Post
Noticeable dip in amplitude response at 50 Hz when shielding circuit with driven right leg network.
I have built an ECG amplifier that uses an instrumentation amplifier and a driven right leg circuit to attenuate 50 Hz common mode voltage stemming from the body. A portion of the circuit is seen below:
I measured some very nice and low noise ECG signals without enclosing my circuit in a shielded can. For example, I recorded an ECG signal with an amplitude spectrum that can be seen below:
I acquired a closed box, whose insides were covered in copper foil, to act as a Faraday cage and placed the circuit inside. I connected the circuit ground to the copper foil inside the box and I still get very nice ECG recordings, but I also get a noticeable dip at 50 Hz in my amplitude spectrum compared to when I recorded without shielding my circuit. Here is my recorded ECG signal along with the amplitude spectrum.
Although the shape of the two amplitude spectra look a little different, I don't think it's a coincidence that there is a dip at 50 Hz. I know that the driven right leg circuit attenuates the common-mode voltage on the body (which is 50 Hz) but it shouldn't attenuate the 50 Hz component of my ECG signal, which is what seems to happen with the shielded circuit.
I haven't been able to figure out why this is, and I am hoping that someone here has an explanation for this.
Answers to questions in comments
Did you measure an actual patient? If so, why isn't there a peak at the heart rate?
I measured my own ECG. Just because a signal is periodic with frequency $f_0$ does not mean the amplitude spectrum will have a peak/component at $f_0$. For example $\cos(2\pi \cdot 2t) + \cos(2\pi \cdot 3t)$ has a period of 1 s, but the amplitude spectrum contains no peak/component at $f=1 \ \text{Hz}$. The same is the case here, where my heart rate is clearly around 60 beats/minute, but no obvious peak is noticeable at 60 Hz.
Is this just electrodes hooked to a dummy picking up ambiet noise?
No, this is a real ECG measured with real electrodes.
How long duration?
20 seconds.
What does the spectra look like with no signal in?
I can answer that tomorrow when I am at the lab again.
1 comment thread