ADS1299 BIAS, REF, Offset, Ground

Thomas Schmitt

Other Parts Discussed in Thread: ADS1299

RE – Bias (RLD), Offset voltage, Reference Electrode and Ground.

TI and forum.

I am coordinating the hardware development for a neuro group. We are about to conduct some tests on the ADS 1299 using the demo board.

We have some questions regarding Bias (RLD), Offset voltage, Reference Electrode and Ground.

First single rail supply. For a single rail supply to work there must be a floating ground produced (offset voltage) usually at ½ ref voltage. There has to be a path to ground from the electrodes either directly or through the skin. This is shown in Fig 21 on the data sheet. The signal from those circuits is shown in Fig 22.

That produces a basic EEG front end. Offset voltage is there, ground is there and the electrodes have a path to ground (for single ended input several channels may use “grounded” electrode” as a reference electrode).

What is missing is RLD. RLD could be added as a separate electrode somewhere (the other earlobe.)

It appears to me that there is an option, even preferred, to tie the reference electrode to the BIASOUT from the ADS 1299 for single rail supply. This is in section 4.6.2 of the Demo Document.

It may be that that section applies to dual rail supplies also but was mentioned by Brian Pisani, TI, on a post regarding single rail supplies.

Question 1.

Does the BIAS provide the offset voltage, the RLD, and a path to ground for the electrodes?

It seems to be doing so.

Question 2.

Can this be done with single rail supplies and dual rail supplies?

For dual rail there is no offset voltage and the non-inverting input to the Bias op amp is ground.

Again it seems so.

BIAS is doing several jobs!

Great chip. Have enjoyed our work to date and looking forward to finish up the bench work and getting a board out. Forgive the belabored preamble.

Tom Schmitt

over 10 years ago

0 Brian Pisani over 10 years ago

TI__Mastermind 38745 points

Hey Tom,

I'm glad you think the ADS1299 is a good chip. I tend to agree.

You have the right idea. In a dual supply configuration, the board ground is conveniently present to serve as a bias for the patient or for a reference voltage with respect to which you may measure other electrode voltages. However in a single supply configuration, a mid-supply voltage must be manufactured to bias the patient and perhaps to provide a reference. This is provided on the ADS1299 as the bias amplifier.

You may or may not connect the bias amplifier output to the reference electrode input. In the scenario where you have the reference electrode simply as the negative input to all the channels (since the ADS1299 only makes differential measurements), you have connected it ostensibly just like any other electrode. In that case, the bias output can be connected to the patient in the manner of a right leg drive rather than a reference. You could also just have them share an electrode in which case you would measure all other electrodes with respect to the bias amplifier output.

In addition to its functionality as a general purpose mid-supply voltage, the amplifier can be configured to send a negative feedback signal for common-mode changes in the input on to the body. This is achieved by using the bias_sens functionality and can be understood by studying Figure 52 in the datasheet where the BIASP and BIASN switches are closed to allow for certain inputs to provide feedback. This functionality can be used in any supply configuration since its intention is purely to improve CMR. You may not want to connect the reference electrode directly to the bias amplifier when you are using this functionality since the output of the bias amplifier will not be DC. Now for your questions...

1. If you think about it, the offset voltage and the RLD are sort of the same thing. The RLD electrode is used to bias the body such that the other leads have a common-mode voltage near the analog mid-supply. I'm not sure what you mean by "path to ground" since we hope very little if any current flows on the patient in the normal scenario.

2. I would not connect the output of the bias amp to ground directly. That reference voltage is generated internally and it's possible that it's not exactly the same voltage as the board ground. Thus connecting the output to ground could force the amplifier to source more current than it is intended to. However, as I mentioned before, the output of the bias amplifier may be connected to the bias electrode in the bias_sens configuration to aid in common-mode rejection.

Regards,
Brian Pisani