• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 46 subscribers
  • Views 2031 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.

ADS1198 Reference Voltage with no RLD

Alan Amin
Prodigy 70 points
Other Parts Discussed in Thread: ADS1198

Hello,

Based on ADS1198 ECG datasheet, in a single supply (Unipolar) application, all the non-inverting inputs

can be tied-up to mid-supply voltage referred to as common voltage or Vref. If RLD circuit is not used for

ac noise elimination, should this Vref be applied to patient’s body (e.g. RA) as a reference voltage for the

input signals? Should also a patient’s protection resistor be added in series to this reference electrode

alike the ones added to the non-inverting inputs for purpose of keeping the front-end differential inputs

balanced as much possible?

Appreciate clarification and/or any related application notes.

Regards,


Alan Amin

  • 0
    TI__Guru 57336 points

    Hello Alan,

    Thanks for your question.

    First of all, I just want to clarify the difference between "single-supply" and "single-ended" measurements - it sounds like you are talking about both in your post.

    Single (unipolar) supply refers to the analog voltage supply used by the ADC (AVDD and AVSS), where AVSS is tied to the system ground. This does not mean, however, that you must connect your input signals in a single-ended configuration, where INxN is tied to a fixed common-mode voltage.

    The standard ECG leads (I, II, III) are typically measured differentially. For example, LA and RA would be connected to one channel (INxP and INxN, respectively) to measure Lead I. This can be done using either a unipolar or a bipolar supply, as long as the input signals are properly biased to remain within the input common-mode range (typically mid-supply is used as the bias). The RLD can be used to drive the patient to this bias point. If your system does not use a dedicated electrode to bias the patient, you might consider using an AC-coupled input scheme followed by pull-up/down resistors to form a resistor divider that sets the signal DC bias to mid-supply.

    On the other hand, if you wish to measure your inputs in a single-ended configuration, then the INxN inputs could be tied to a common-mode voltage. This should be the same common-mode voltage as your INxP inputs so that the ADC output only contains the differential measurement. Also, note that this "Vref" you mention should not be the same as the reference voltage used by the ADC for conversions (VREFP - VREFN).

    Hope this helps!

    Best Regards,

  • 0 in reply to Ryan Andrews
    Prodigy 70 points

    Hello Ryan,

    Thank you very much for your detailed explanation. Since our application is a bit unusual, we would appreciate your further expertise inputs on our setup for ECG measurement.

    We are planning to measure the differential ECG voltages of several internal/invasive electrodes with respect to a single external electrode attached to the body skin. Utilizing an ADS1198, eight internal electrodes are routed to eight INxPs. All INxNs are tied together and routed to the single external electrode. While all channels measuring differentially, channels 6, 7 and 8 are selected as RLD Sense-Amps for common-mode DC biasing/AC noise average calculations. After filtering, the RLDOUT is routed to a second external electrode attached to the skin. As a result, the biased DC voltage and the low frequency noise will be canceled out by this common-mode Op-Amp configuration. Then, the onboard ADC will process the total of eight channels’ ECG differential signals further.  The questions here are:

    1. Is the forgoing a proper setup for differential ECG measurement?

    2. Assuming step 1 is correct, for the purpose of eliminating the RLD electrode, the RLDOUT is disconnected from the second electrode and is connected to the first external electrode on the patient that connects to the INxNs. Thus, a single electrode is serving as both RLDOUT and INxNs. Would this setup work properly?

    Appreciate your expertise feedback.

    Best regards,

    Alan A

  • 0 in reply to Alan Amin
    TI__Guru 57336 points

    Hi Alan,

    Alan Amin said:
    1. Is the forgoing a proper setup for differential ECG measurement?

    Yes, this configuration is fine. It is very similar to the standard chest leads in non-invasive ECG measurements (i.e. V1 through V6), where each electrode is measured against a common WCT voltage.

    Alan Amin said:
    2. Assuming step 1 is correct, for the purpose of eliminating the RLD electrode, the RLDOUT is disconnected from the second electrode and is connected to the first external electrode on the patient that connects to the INxNs. Thus, a single electrode is serving as both RLDOUT and INxNs. Would this setup work properly?

    This answer I am not completely sure of, neither from a device perspective nor an ECG-theory perspective. I will share my initial thoughts below, please feel free to tell me if they make sense to you or not.

    I believe that the drive capability of the RLD amplifier is not enough to drive 8 INxN inputs directly (typical short-circuit current = +/-250 uA). If you buffered the RLD output, or if you pulled-up each INxN pin to RLDOUT with ~10M, then the RLD amplifier should be ok.

    Now the remaining question is: do you still need to drive the body with the external electrode (also connected to RLDOUT)? Can you tell me if your inputs to the ADS1198 are DC or AC coupled? Often, I have seen systems with AC coupled inputs that use a pull-up resistor (~10M) to RLDOUT. If your INxP inputs were AC coupled and also pulled-up to RLDOUT, would you even need to drive the body with the external electrode?

    Best Regards,

  • 0 in reply to Ryan Andrews
    Prodigy 70 points

    Hello Ryan,

    Thanks for sharing your candid thoughts with us.

    First of all thanks for pointing out the required buffer at the RLDOUT port. Second, to answer your question about the inputs coupling, all are DC and each has a 100K in series for patient protection. 

    Just to explain our analogy for the step 2 in the previous post, where the RLDOUT is connected to the single electrode through a buffer to INxN inputs, the internal electrodes that are connected to INxPs would be biased to RLDOUT level through the body impedance. In other words, RLDOUT feedback signal is fed directly into INxN inputs and through the body impedance into INxPs. If this analogy is correct, then each front-end amplifier is biased in a common-mode configuration! Please let us know of your thoughts on this analogy.

    On the other hand, you brought up the subject of AC bias, which might be a better option in our setup. In which, you mentioned a series DC blocking capacitor in each INxP input path. There are some questions:

    1. What is a typical Cap value and type?

    2. Is it correct that no RLD electrode would be required in this configuration?

    3. What would be the status of INxN inputs in this case?

    Appreciate your thoughts and clarifications.

     

    Best regards,

    Alan A

  • 0 in reply to Alan Amin
    TI__Guru 57336 points

    Hi Alan,

    I've been discussing the pros and cons of these two approaches with a designer on our team. Here are our thoughts:

     

    1. With all inputs DC coupled, it sounds like you have a choice between

      1. Driving the body with RLDOUT through one external electrode, and connecting INxN inputs to a separate, external electrode (your original idea). INxP internal electrodes would be biased through the body's impedance.
      2. driving the INxN inputs directly with RLDOUT while also using a single, external electrode to bias the body. INxP internal electrodes would be biased through the body's impedance.

    My concern with "b." is that the negative input path would be significantly different than the positive input path. Therefore, the RLDOUT will be "less common" to the positive and negative inputs, creating a differential signal error.

     

    The RLD amplifier is rather noisy (7uVrms). This is not a concern when the positive and negative input paths are very similar, as in option "a." In option b., however, this would be a concern, so a low noise external RLD buffer should be used. Try to find a buffer with the same noise or less than the channel PGAs. I think this is likely the best option.

     

    2. If you chose to AC couple your inputs, you would not require a dedicated RLD electrode. Instead, INxN inputs would be tied to RLDOUT (preferably buffered with a low-noise amplifier) and INxP would be pulled-up to RLDOUT following the DC-blocking cap. The size of the cap will form a high-pass filter with the RLD pull-up, so choose both components to set the corner frequency you want in your application.

     

    I have not had a chance to analyze the difference between option 1 b. and option 2 for best common-mode noise rejection, but please let me know your thoughts.

     

    Best Regards,

  • 0 in reply to Ryan Andrews
    Prodigy 70 points
    Hello Ryan,

    Thank you so much for the detailed explanation and the extra mile that you went to provide us the precise answers for various approaches.

    As far as your concern about section b in part 1 of your text; do you think placing an RC network (51 K and 47 nF in parallel) in the feedback path to the INxN equivalent to body’s impedance, will balance the input paths close enough to avoid the differential error? Thereby improving the CMR! If you have any thoughts on this subject, please let us know.

    Best regards,

    Alan A
  • 0 in reply to Alan Amin
    TI__Guru 57336 points
    Hi Alan - I'm happy to help. I'm not sure how much improvement you might see, but it may be worth a try (you can always uninstall the components and short across :) ). The external buffer on RLDOUT was suggested to reduce the RLD amplifier noise, which was important because the noise was no longer common to both inputs. With the external buffer, that noise is less of a concern.

    Best of luck!
  • 0 in reply to Ryan Andrews
    Prodigy 115 points

    I'm trying to use the ADS1198 in AC coupling approach, because I need best resolution (with gain of 6 I have 12.2uV of LSB and I want 6.1uV).
    Someone could show me a schematic showing this setting to take a 12-lead ECG, (RA, LL, LA, and V1 to V6)
    Thank you very much