ADS1292: removing common mode noise (60 Hz)

Krithik Chandrashekar

Part Number: ADS1292
Other Parts Discussed in Thread: ADS1298

We are currently observing that AC is getting coupled into the system. And would like to know what the best way to remove this without burying the signal of interest.

We currently have the electrodes placed on IN1P and IN1N. On the muscle of interest they are placed adjacent to each other
There are passives placed before the IN1P and 1N2P follows exactly the design from the Development Kit for the ADS1292.
The device is battery powered.

Do we require additional electrodes placed? I'm not sure placing the notch filter is the best approach, does this have to be done through software.

Thanks!

over 6 years ago

0 Brian Pisani over 6 years ago

TI__Mastermind 38745 points

Hello Krithik,

It is very common in biopotential applications for mains interference to couple into the signal chain. Typically this happens because the body acts as a large antenna for interference for signals from overhead lights, fans, motors, etc.

There are two ways this is dealt with. The first is using the integrated Right Leg Drive amplifier on the ADS1292. You can find out more information about it in the datasheet as well as this app note. The app note was written specifically for the ADS1298, but the right leg drive works the same on the ADS1292 so you should be able to carry the concepts over.

The other way it is handled is through a digital notch filter like you mentioned. Most systems employ both methods.

Brian

0 Krithik Chandrashekar over 6 years ago in reply to Brian Pisani

Intellectual 855 points

HI Brian,

Appreciate the prompt response! Is a RLD common (no pun intended) for EMG (as opposed to ECG) measurements?

0 Brian Pisani over 6 years ago in reply to Krithik Chandrashekar

TI__Mastermind 38745 points

Hello Krithik,

In any dc coupled biopotential application, you need some kind of drive signal to maintain the common-mode of the signals you are measuring. This is required for EMG if it is being dc coupled to the inputs of the ADC. Alternatively, you could ac coupled the inputs, but you lose the ability to use the closed loop structure of the RLD circuit to reject mains interference.

Brian

0 Krithik Chandrashekar over 6 years ago in reply to Brian Pisani

Intellectual 855 points

HI Brian,

We have now enabled the right leg drive circuit and see little to no improvement, this is essentially the description of the setup:

Schematic follows the reference design of the EVB. There's a 1MOhm and 1.5nF in parallel to the RLDOUT and RLDINV pins.

[Hardware]

1. Electrode pair on IN2P and IN2N, goes to the muscle of interest
2. Electrode from RLDOUT, going to a bony portion of the body as a reference

[Software]

1. DR1 bit is set in CONFIG1
2. RLD2N and RLD2P bits in RLD_SENS are are set
3. RLDREF in RESP2 is set

Thanks,
Krithik

0 Brian Pisani over 6 years ago in reply to Krithik Chandrashekar

TI__Mastermind 38745 points

Krithik,

Have you set PDB_RLD in the RLD_SENS register? This is required to turn the RLD amp on.

Brian

0 Krithik Chandrashekar over 6 years ago in reply to Brian Pisani

Intellectual 855 points

Yes, we tried that initially and toggled it off to see of there was any change. No change was seen between the two measurements. The documentation in the datasheet wasn't clear on what "RLD buffer" exactly was.

0 Brian Pisani over 6 years ago in reply to Krithik Chandrashekar

TI__Mastermind 38745 points

Hello Krithik,

Can you please attach your schematic here?

Brian

0 Krithik Chandrashekar over 6 years ago in reply to Brian Pisani

Intellectual 855 points

ads1292_schematic.pdf

0 Brian Pisani over 6 years ago in reply to Krithik Chandrashekar

TI__Mastermind 38745 points

Hello Krithik,

I believe there are two problems:

The cutoff frequency of the high pass filter at the inputs is verh high. I calculate it is 1/(2*pi*2.2nF*(10MOhm||10MOhm)) = 14 Hz. This will attenuate a lot of your ECG signal.
The component mismatch on the high pass filter component is degrading CMRR significantly.

Actually the high pass filter is not necessary since you are using the right-leg drive amplifier with a RL electrode. The RL electrode will bias the patient properly. I recommend you remove R3, R4, R5, R6, R11, R12, R13, and R14 and replace C2, C3, C12, and C13 with 0 Ohm resistors to short out the pads.

Regards,

Brian

0 Krithik Chandrashekar over 6 years ago in reply to Brian Pisani

Intellectual 855 points

This was on spot. We still see a spike at 60 Hz (no harmonics) but definitely more usable emg content compared to before. Thanks for your help through this, it was very appreciated!

Data converters

Data converters forum

ADS1292: removing common mode noise (60 Hz)