• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 46 subscribers
  • Views 1831 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.

ADS1292: Using the ADS1292 without RLD electrode

Chief Consultant
Prodigy 160 points
Part Number: ADS1292

Dear TI,

This question is asked a number of times already on this forum. The advice is to go for an AC coupled design and DC bias the input terminals to the mid supply voltage. The DC bias can be done by:

  • a voltage divider e.g. 2 resistors of 10M between Vss and Vdd
  • using the output RLD amplifier as a unity gain buffer and configure the input of the RLD amplifier to the mid supply and use a large resistor to pull the inputs of the ADS1292 to the mid supply

I am wondering why AC coupling is advised with respect to reducing common mode problems, because the to be expected common mode problems are from capacitive coupling of the mains (50 / 60Hz) grid  and the high pass filter that forms the AC coupling will just let these frequencies pass. They hopefully will be sufficiently rejected when the CM components at the input terminals are subtracted from each other (made differential).

The whole idea of 24-bit dynamic range from what I understand is to be able to cope with large electrode offsets (+-300mV)  and the CM mode components on the terminals and still be able to sufficiently quantise the relative small (2mV) ECG signal.

I understand of course that if you would use DC coupling  some form of high pass filtering would be required to remove base-line drift of the ECG signal but this could be done digital that does not necessarily has to be done by AC coupling.

Can you shed some light on this?  I am sure it will be valued by many users that want to build an application were the use of an RLD electrode is not possible.

TIA for your support

  • 0
    TI__Guru 57306 points

    Hi Chief,

    Thanks for your post!

    I think this question comes from confusion between two DC voltages: the DC common-mode voltage on each input signal and the DC differential voltage that is converted by the ADC.

    The RLD electrode is used to establish a known DC common-mode voltage on the body. This ensures that the electrical signals measured by each electrode remain within the acceptable common-mode voltage range of the amplifiers at the channel inputs. Without the RLD electrode, the electrode inputs must be AC-coupled and biased after the series capacitors to establish a known DC common-mode.

    The gain stage (PGA) in the ADS1292 is comprised of two amplifiers in an INA configuration and the common-mode voltage passes through. It is the ADC itself which cancels the common-mode voltage when it converts the difference between PGAxP and PGAxN. Assuming that the impedances in the signal path for each electrode input are equal, this common-mode voltage will be cancelled completely by the ADC. However, differences in the impedances of each input path cause "common-mode to differential" conversion, which can produce a DC offset at the ADC output. AC signals which are common to both electrodes can also become a differential noise in the same manner.

    As you mentioned in your second bullet, the RLD amplifier can still be used as a bias voltage for AC-coupled inputs. If you were to enable the appropriate RLD_SENS bits in the register settings, you can also sense the AC common-mode voltage at the PGA outputs and derive a cancellation signal at the RLD output. It's difficult to quantify how effective this is at cancelling 50 Hz / 60 Hz interface compared to driving the body directly, but it's definitely an option.


    Best regards,

  • 0 in reply to Ryan Andrews
    Prodigy 160 points

    Hello Ryan,

    Thanks for your answer. I am not sure if I completely understand. Maybe I can be a little bit more specific. Lets suppose a battery powered device because this would make the most sense not using RLD feedback because of very high isolation to CM ground and suppose a two lead system which is typical for the ADS1292.

    The input network you propose would look something like:

    You state that that the electrodes must be AC coupled, this is done by C1,C3 and C5. My question is if I remove them what would go wrong?

    IMHO  patient leakage current is not a problem, high-pass filtering can be done digitally and 50/60 Hz common mode rejection will be better, so why do I need AC coupling?

    Kind regards

  • 0 in reply to Chief Consultant
    Prodigy 160 points

    Hello Ryan,

    I am still very curious why AC coupling is required, and probably others too who are interested in an application without RLD feedback. So your response is highly appreciated.

    Kind regards

  • 0 in reply to Chief Consultant
    TI__Guru 57306 points
    Hello,

    Please excuse the delay, I was out of the office for a few days.

    If you were to short the series 1-uF capacitors in the circuit above, the individual inputs would still be biased to mid-supply via the 10-Meg pull-up resistors to 1.65 V. There would be no concerns with violating the PGA input common-mode range. In that case, you would not need the series caps. In fact, not having them would actually be better for CMRR. You may see a gain error from the mismatch in the resistors, which form a resistor divider, but that's it.

    I'm curious if there was ever another reason behind that original recommendation. We have seen it done that way before. I'll ask around the team if there are any other reasons.

    Best regards,
  • 0 in reply to Ryan Andrews
    Prodigy 160 points

    Hello Ryan,

    Thanks for your response. Looking forward to hear if there is any argument against DC coupling.

    Kind regards,

  • 0 in reply to Chief Consultant
    TI__Guru 57306 points

    Hi Chief,

    As I explained in the other post, there may not be any issue with the purely DC coupled approach with pull-up resistors to RLDOUT. Any offsets introduced by the pull-up resistors may not be an issue for lower gain settings. Most likely, PGA gain of 3 to 6 V/V will still have sufficient headroom and still give you the benefit of input-referring the ADC noise low enough.

    I'm not convinced yet whether it's possible to achieve a CMR benefit from the RLD loop in this manner. 

    Regards,

  • 0 in reply to Ryan Andrews
    Prodigy 160 points

    Hi Ryan,

    " I'm not convinced yet whether it's possible to achieve a CMR benefit from the RLD loop in this manner", I agree, that is the consequence of not using an RLD electrode.

    High CMR is achievable only by very good isolation to the CM ground. In my opinion, the best way to do that is having a battery operated device.It would be nice if TI made a application note using the ADS1292 without using a driving electrode because for ambulatory devices an extra electrode often is not an option and it addresses a considerable share of the market.

    Kind regards,

  • 0 in reply to Chief Consultant
    TI__Guru 57306 points

    Hi Chief,

    Thank you very much for your suggestions. We appreciate getting feedback from customers who work more closely with the end application on a regular basis. I agree that CMR for 2-electrode portable ECG applications has been a gap in our technical collateral for some time.

    I will keep this topic near the top of the list for ADS129x related collateral. Please keep an eye on the growing FAQ list that Alex Smith and I have been writing (linked below). We'll likely start with a post there before expanding further on the subject.

    Best regards,

  • 0 in reply to Ryan Andrews
    Prodigy 160 points

    HI Ryan,

    This is highly appreciated. If I can contribute in some way please let me know.

    Kind regards,