Hi, I’m designing a circuit with the primary function of acquiring myoelectric signals (EMG). The circuit includes an instrumentation amplifier, which amplifies the myoelectric signals acquired by the electrodes while eliminating external noise, there’s a filtering stage: a second-order Butterworth band-pass filter, targeting the frequency range relevant for this type of signal, between 20 Hz and 500 Hz. Following that, there’s a precision full-wave rectifier, and finally, a peak detector. The ultimate goal is to read this through an analog-to-digital converter (ADC) using an STM32.

I should also mention that since I only have a 24V power supply, I built a voltage divider circuit followed by a buffer, which allows me to have a dual supply of +12V and -12V.

Firstly, I’d like to ask the more experienced folks if they could provide feedback on this circuit, as I’m sure there are some that could be improved.

I'm trying to simulate the circuit in LTspice, but it’s reached a point where the simulation simply won’t start and becomes very slow. When I test each stage individually, LTspice manages to represent the desired outcome, but when I combine all the circuits, it just "blows up." I’m using components from the library (http://www.bordodynov.ltwiki.org/). Does anyone know why this issue is happening? I tried different solvers and integration methods, but nothing seems to work.

 

I recommend:

(1) Simulate the individual sections, analyze and improve. Pay attention to input and output impedances.
(2) Build the individual function blocks- test, verify
(3) Cascade the blocks, test, verify.

Get working with real stuff- the simulation only gets you so far, as you see.
It also becomes impractical as the complexity increases.
Simulation also conceals real-world problems that you need to know about- sooner the better.

 

"Ampop" is temporarily removed.
______

 

Check the grounding of your inputs and review your LTSpice connections. Adjust the gain settings.

 

I suggest, at least for the simulation, use two DC supplies, +12 volts and -12 volts. And then get rid of the peak detector, it is better to do that in software. All of the voltage gain required can come from the instrument amplifier. That should help the simulation.