• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 2006 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ADS1191: ECG lead I using just 2 electrodes with RLDOUT Bias

bl na
Prodigy 70 points
Part Number: ADS1191

I've read on the forum about using RLDOUT/RLDINV for input bias and Ryan Andrews suggested the following:

If we label the capacitor and resisters from left to right as C1, R1, R2, C2. They are C1=0.1uf, R1=5Mohm, R2=51K and C2=47pf. So the first C1R1 high pass filter cutoff frequency is about 0.3Hz. 

Same circuit for the INxN input as well.

What is the correct register settings for this RLD bias circuit to work properly?

Currently I have MISC2 register, RLDREF=1 so the input voltage would be biased to the middle (AVDD+AVSS)/2. Other settings related to RLD stayed at defaults.

The issue I'm facing now is that it will either get a few seconds of ECG signal about 2mv which is expected then just fall to 0mV flat, or simple stay at 0mv flat.

But if I remove R1, then ECG signal about 2mv would be captured continuously, of course it would DC drift without the bias pull up.

What could be the problem in the circuit or register programming?

  • 0
    TI__Guru 57336 points
    Hello Bl,

    Thanks for your post.

    Did you enable the RLD buffer? By default, it is powered-down. The buffer can be enabled by setting RLD_SENS[5] = 1.

    Best Regards,
  • 0 in reply to Ryan Andrews
    Prodigy 70 points
    Hi Ryan,

    Thank you for your prompt reply.

    I thought I had tried enable RLD buffer before and did not see any difference. But after getting your reply and setting that bit, now it works. I am enable to capture the ECG continuously now.

    The DC offset changed to a bit higher than 0v vs a bit lower than 0v without enabling RLD buffer. So I think when RLD buffer was disabled the ECG signal was leaked through RLDOUT since it is not in output mode.

    Another issue I'm facing is related to noise. If the PCB was held in hand by 2 fingers touching the edge of the board (not touching any component or trace) I can get a clean ECG signal. But the ECG signal would be very noisy and almost unusable if the small pcb was not held by hand.

    The board is powered by a 3V battery pack and the electrodes were applied on chest and connected to the board through a pair of 20" ECG cables. eventually the connection between electrodes and pcb will be very short so that would not be a problem. I also tried to shield the 20"ECG cable, but that did not have much effect on noise.

    Can you provide some suggestions on how to reduce the noise when pcb is not held in hand?

    Thank you!

    Jim
  • 0 in reply to bl na
    TI__Guru 57336 points
    Hi Jim,

    I'm glad we were able to fix at least one issue!

    It's very difficult to tell how the noise could be coupling into your system differently when you touch the PCB vs. when you don't. Are you placing the board down on a table or something conductive?

    Though the board is powered by a battery, power line noise (50 Hz or 60 Hz) can couple onto the board or the body anyway. This is where the RLD can become crucial. If you enable the RLD_SENS bits for the channel on which you are using to measure ECG, the RLD amplifier will sense the output of that channel and generate an inverted common-mode signal to drive the body (see Figure 47). This inverted common-mode signal is centered around the same RLDREF voltage to set the DC bias and helps cancel the common-mode noise on the body. I understand you are using AC coupled inputs instead with a pull-up resistor to RLDOUT, but this method may still work just as well. I'm curious to see the results if you could try this.

    Keeping the cables short is also important. Are you driving the shield of the ECG cables? Typically, the shield is driven to the inverse of RLD such that it carries the same common-mode signals as the ECG inputs do. This reduces the parasitic capacitance between the signal and the shield.

    Best Regards,
  • 0 in reply to Ryan Andrews
    Prodigy 70 points

    Hi Ryan,

    The PCB was just hanging in the air not touching anything conductive.

    Thank you for your suggestions. Need to confirm the changes: Set the RLD_SENS bit with existing circuit, 1P or 1N, what is the difference? Do I also need to add the following changes as shown in Figure 47? Not shorting RLDINV with RLDOUT, add Cext =  1.5nF and Rext = 1Mohm

    Right now there is no connection for the shield of the ECG cables, I will try connecting to RLDINV.

    Best regards,

    Jim

  • 0 in reply to bl na
    TI__Guru 57336 points

    Hi Jim,

    50 Hz and 60 Hz noise can still couple onto a circuit board through electromagnetic interference just by being exposed to the lighting in the room.

    The selection of RLD_SENS 1P or 1N depends on the electrode connected to the respective input pin. If you want the RA electrode to contribute to the RLD common-mode derivation, then you must select the PGA output which is buffering RA. The choice is really an application preference, not a device requirement. The R-C feedback components for the RLD amplifier are to limit the closed-loop bandwidth and help maintain amplifier stability.

    The RLDINV node will not have sufficient drive capability to connect to the cable shielding. I would recommend buffering RLDINV separately or inverting the RLDOUT signal to drive the shield.

    Best Regards,