I think it should be biased like this:

 

@eetech00 and @Audioguru again

You are correct about the bias reference voltage and the decoupling capacitors. Please note that these were implemented correctly in the actual circuit. Unfortunately I had to used different components for the schematic and just rename them, as the ones I need are not available in the library (i.e. no simulation was done, hence why I did not cross-check the design). Below is the corrected circuit which I implemented in practice, just not including the decoupling capacitors.

Will reducing the values of R1 and R2 will eliminate this problem? My concern is that if I reduce the value of these resistors, the voltage dividers R1-R3 and R2-R4 will attenuate the input signal by a lot. I can reduce the values of R3 and R4 to compensate for this, but that would defy the purpose of these resistors, which are these to protect the person for getting electrocuted in case something goes wrong.

 

Your R6 and its opamp should be removed because the instrumentation amplifier needs REF to be fed from a low impedance like the output of the vs/2 opamp. Look at my previous post.
Here is the INA333 on its datasheet with a dual-polarity supply and REF is at ground:

 

@Audioguru again , that REFERENCE output is for the reference electrode and for the reference of the IA. All EMG circuits are wired this way (example the circuit below from Analog Devices, where C is the reference electrode).

In my circuit, the reference of the IA is supplier from U1:B, which has no resistors at the output.
I used a resistor at the output of U1:A as a protection to an electrical shock. This is used in most EMG circuits.

 

Before you did not say and did not show the patient connected to R6.

 

I already tried this without R6. The only difference i am getting is less (much less) noise at the output. The waveform below is what I am getting with R6 connected. Still getting the saturation when touching an earthed device.

 

Your Ref input to the IA is done differently from the post #24 circuit. Why?

 

Hello

Here is a simulation of your circuit

eT

 

@Alec_t that was just a circuit which I got from the internet to show Audioguru that the reference is referring to the reference electrode. This is a different circuit from what I'm using (different components etc). To be honest, I tried to understand that circuit, but could not understand the use of the integrator op-amp being used for the REF pin.

@eetech00 , thanks for the simulation. I will try i out on LTspice. It seems to operate as in practice, but unfortunately i cannot simulate the saturation which I am getting.

 

@Alec_t that was just a circuit which I got from the internet to show Audioguru that the reference is referring to the reference electrode. This is a different circuit from what I'm using (different components etc). To be honest, I tried to understand that circuit, but could not understand the use of the integrator op-amp being used for the REF pin.

@eetech00 , thanks for the simulation. I will try i out on LTspice. It seems to operate as in practice, but unfortunately i cannot simulate the saturation which I am getting.

Did you try lowering the values of the bias resistors? Just to check the effect?

What are you using to breadboard?

It seems like the ground signal is reaching the -IN of the INA333....

eT

 

I tried to understand that circuit, but could not understand the use of the integrator op-amp being used for the REF pin.

It's my understanding that it averages the output voltage and uses that as the Ref, so that changes are centred around the average. That may avoid the saturation effect you are seeing.

 

I already tried this without R6. The only difference i am getting is less (much less) noise at the output. The waveform below is what I am getting with R6 connected. Still getting the saturation when touching an earthed device.

View attachment 200340

Hello again,

I'm curious as why so much gain (500) is needed?
There is the potential for more than 100mv to be measured at the input (resulting with 50v (!) at the output).
Wouldn't a gain of, say...10, be better, and add second stage for more gain if necessary? It seems the large gain may be contributing to the noise problem..

eT

 

@eetech00 , yes I reduced the values of the bias resistors from 47k - 100k and try again.

Yes, this circuit in implemented on a breadboard.
I used 200ohms to get a gain of 500 as the raw EMG signal is in the low mV range. The amplification of the raw EMG signal worked perfectly (with no saturation), but eventually I will use a potentiometer for variable gain adjustment.

 

An ECG circuit takes the addition of each sensor which has a common-mode signal on each sensor (50Hz or 60Hz electricity interference picked up by the patient's body), inverts and filters it and feeds it to the patient to cancel it.

 

Hi again,

I did some testing on TINA and did the following observations:

1) When using the component values used in the circuit shown in post #22 and slightly varying the value of capacitors C1 or C2, this is causing the output to swing to GND (eventually saturating if the ratio between C1 and C2 is further increased) - Please see 1st snapshot below.

2) If I change the values of C1 and C2 to 2.2nF and the values of R2 and R3 to 18Mohms (please see 2nd snapshot), this will have little effect on the output when varying the ratio between C1 and C2 - as can be seen in the snapshot.

Could it be that this is what is causing the output to saturate to 0V when touching earthed equipment?

 

Could it be that this is what is causing the output to saturate to 0V when touching earthed equipment?

I think what you are seeing in the sims is the effect of the time constant R2C1 (or R3C2). In the first sim the time constant is about 650ms, whereas in the second sim it is only about 18ms (or 40ms).
I don't see how either sim represents 'touching earthed equipment'.

 

Hi again,

I did some testing on TINA and did the following observations:

1) When using the component values used in the circuit shown in post #22 and slightly varying the value of capacitors C1 or C2, this is causing the output to swing to GND (eventually saturating if the ratio between C1 and C2 is further increased) - Please see 1st snapshot below.

2) If I change the values of C1 and C2 to 2.2nF and the values of R2 and R3 to 18Mohms (please see 2nd snapshot), this will have little effect on the output when varying the ratio between C1 and C2 - as can be seen in the snapshot.

Could it be that this is what is causing the output to saturate to 0V when touching earthed equipment?


View attachment 200962
View attachment 200963

Your simulation is a little flawed.

1. Use a separate opamp for the right leg driver. Include a 220k resistor in series with its output (per your safety requirement). This helps keep it isolated as much as possible from the INA Ref pin and Vs/2 Bias voltage.

2. The body voltage generators (E1 and E2) should be 180 degrees out of phase so noise is cancelled by the INA but allows signal to pass.

3. The signal path from E1 and E2 should include "human body resistance". I used 1M ohm resistor in the path of each signal source to represent the body resistance.

While I realize the input signal will be small, I still don't understand why there is so much gain needed for the INA stage.
I also don't think the bias resistors have to be so large (18MEG).

Anyway, good luck.

eT

 

I would beware of rather small CMMR just 90 dB according to datasheet and very low BW of 350 Hz. However, probably it is enough for Your aims. As well the load max capacitance 500 pF may cause many troubles.