• State Resolved
  • Locked Locked
  • Replies 19 replies
  • Subscribers 54 subscribers
  • Views 12231 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Electrodes/ADS1298 Interface

Julian Grisales
Intellectual 300 points
Other Parts Discussed in Thread: ADS1299, OPA2241, ADS1298, ADS1294R, ADS1291, ADS1198, TLC2272, OPA336

Hello,

I maybe over-read some documentation and now I'm a little bit confused in how to implement the connections between the electrodes and the ADS1299, so I'm asking here for a little guidance.

I'm using a single pole 6V power source trough 4 AA batteries, this voltage go to a TPS71202 dual channel LDO regulator configured to throw  +5V and +3.3V voltages working as AVDD and DVDD respectively. I'm also using a REF5025 low noise external reference connected to AVDD and giving 2.5V, this will be used as reference voltage for the ADC in the ADS1299, this reference voltage is being divided by a basic voltage divider with two resistors and an OP-AMP buffer (OPA2241) creating the Common-Mode Voltage of 1.25V, to be used with the Bias Drive amplifier as a body reference. In short, I'm using the ADS1299 with unipolar supply and external reference of 2.5V, also the AVSS will be equal to AGND and DGND.

I'm developing a surface EMG acquisition device, I have look for a lot of ways of doing this interface between the electrodes and the chip, but always got confused. As first, I was following the datasheet in p.18. In single-ended input mode I assumed all the negative input electrodes will be connected to CM (Common-Mode), connecting them in-between over the body (what cares of sense for me), this confusion was developed by a schematic I found were they use a pre-stage with IN-AMPs where the two electrodes were differentiated before and then compared with CM at the same analog input of the ADS1298:

When I realized this configuration was for a unique electrode for channel then I decided to use the differential one, but all the information I have gathered for this is using (I suppose) a bipolar supply configuration, because the electrodes seems referenced to AGND by a capacitor, I couldn't be sure if is as simple as just replacing the AGND for CM. Also, I didn't quite understand why in the datasheet they talk about of a 180° phase angle between the two inputs of the channel.

After I did an in-depth search of related topics in this same forum, the most similar case was: http://e2e.ti.com/support/data_converters/precision_data_converters/f/73/t/200221.aspx . Here the user asks for a way to correctly connect a single channel for ADS1298. He got as answer the diagram showed later; there is no connection to AGND or CM between the two electrodes, but a connection to RLD_Out through a high impedance resistor. This got me a lot more confused. Then there is no necessity to connect a CM? This configuration just work for one channel or could be configured to multiple ones?

So, my questions for the community are:

  • How could I correctly interface the electrodes to the ADS1299 taking into account that I’m using a unipolar supply and an external ADC reference?
  • How should be configured the Bias Drive channels internally (of the ADS1299) to work with multiple inputs? I'm talking about MUXx[2:0] and BIAS_SENSx[x] registers.
  • Should I use the internal Bias Drive amplifier as a buffer for another amplifier or it’s enough if I put the filter there?
  • I’m only taking into account a Bias Drive Electrode, but in the user guide for the ADS1299EEG-FE there is a Reference Electrode, what’s the differences between those? Do I need both?
  • In other users configurations as in the same user guide mentioned earlier I have seen a “shield” connection with the Bias Drive Amplifier (or RLD also), where is this pin connected to? Is this necessary?
  • Lastly, at the ADS1298 FAQ section of this forum says that unused pins should be connected to AVDD, and I actually could see that implemented in the schematics of the ADS1298ECG-FE, but not in the ADS1299EEG-FE schematics, is this AVDD connection necessary for ADS1299?

I hope this could be helpful for other users, not just me. Also I have taken a lot of assumptions; I would appreciate if with the answer you correct me of undetected misconfigurations I already have supposed.

Thanks in every way!

  • 0
    TI__Guru**** 180831 points

    Hi Julian,

    Sorry for the delay here - I'll talk with the Delta Sigma guys and see if we can't get you comprehensive answers to your questions shortly.

  • 0 in reply to Tom Hendrick
    Intellectual 300 points

    Hi Tom,

    Thank you very much, i'll be waiting here.

    :)

    Best  regards,

    Julián.

  • 0 in reply to Tom Hendrick
    Prodigy 10 points

    Hi,Dear engineer, I have used the ADS1294R for a long time, The respiratory wave is likely modulated by something. Please explain me what is the small wave on the respiratory wave? The heart beat or the motion artifacts? Another question is how you define the bioimpedance resolution ratio is 20milliohm? You compare this respiratory wave with what stardart respiratory signal?  It's a  very confuse problem for my research. If you  send a email to sony17305@163.com to collect me. I'll very grateful. Thank you for your applying. 

    ppreciate it.

  • 0 in reply to Ning Song
    TI__Mastermind 43915 points

    Hi Ning,

    I would recommend giving the following a read if you have no already.

    http://www.ti.com/lit/an/sbaa181/sbaa181.pdf

    Digital filtering is commonly used to remove the motion artifacts. They may couple through as they reside in the low bandwidths shared by ECG.

    Conversion of the data word to impedance (ohms) is explained on page 42 in the following user's guide -

    http://www.ti.com/lit/ug/sbau181a/sbau181a.pdf

    Regards,

    Tony Calabria

  • 0
    TI__Mastermind 43915 points

    Hi Julian,

    I will do my best to answer some of your questions that you have here -

    The ADS1291 figure I put together above is a way in which you can do single lead ECG without the need for a separate RLD electrode. As you research the application, you will see that medical grade ECG applications, commonly have a separate RLD electrode to set the bias point for the body. Without this RLD electrode, the body will float independent of the circuit which may cause the input to be outside of the chip operating limits. One way to get around a separate RLD electrode is to use a method like the one explained above using an AC coupled path to kill the DC and then set the bias point though high impedance resistors.This can be done on multiple channels.

    Unipolar supply is fine. You use the RLD to set the bias point and when configured for a a unipolar supply configuration, this will set the RLD to 1/2*AVDD to help maximize the usable code range. 

    Bias Drive configured to use the midpoint from multiple inputs explained here - http://www.ti.com/lit/an/sbaa188/sbaa188.pdf- This can be done and is explained in the datasheet for the device.

    You should be able to use the internal drive amplifier without any issues.

    The reference electrode is just a net name and is used to route the bias back through SRB1 as a possibility for the negative input pins.

    Shield is used to shield the RLD signal providing a 180 degree out of phase signal on top of the RLD signal. It is up to you if you need it depending on your system. It is important in very high performance ECG systems.

    Tie unused analog input pins to AVDD in both the ADS1298 and ADS1299. Digital input pins should never float so make sure the static lines have a definitive state. Digital output pins what you do not need you can terminate to GND through a weak resistor.

    Regards,

    Tony Calabria

  • 0 in reply to Tony Calabria
    Intellectual 300 points

    Hi Tony,

    Thank you very much for your answers, are very clarifying. I now have some other questions though.

    Precisely for helping to maximize the usable code range, I'm using an external ADC reference of 2.5 V. It is appropriated to use a CM of 1/2*Vref(ADC) = 1.25 V to set the bias point in RLD?

    Should I use this CM voltage as reference in analog input connections for proper bias point setting? Like this:

    I think that if I use AGND as in the User Guide I will be losing negative codes because I'm not using bipolar supply, or is the RLD setting the bias point internally of the ADS1299?

    If possible could you give me some documentation with the proper use of RLD shield? I don't understand if this is the same RLD electrode signal with a 180° out of phase angle, or is another signal provided with a specific specially shielded type of electrode or any of those, or both! I'm guessing the phase angle is induced by the inverter voltage follower that appears at the user guide, so I lean towards first case, same RLD electrode signal. But i would like to know the advantages of doing so also.

    Again, thank you very much, all your help is very appreciated. You truly are a Mastermind.

    Best regards,

    Julián.

  • 0 in reply to Julian Grisales
    TI__Mastermind 43915 points

    Hi Julian,

    The common mode range limitation is set by the power supply which is why the RLD uses the midpoint from the two power supplies to set the RLD voltage. You need to use high impedance resistors to set the DC operating point, caps are not going to work. See the picture I included using the ADS1291 above.

    Using a single supply is fine. The bias point is taken from half of the two supplies.

    Shield drive we have seen done both through inversion of the RLD and noninversion. See the U2 amplifier configuration for the ADS1298ECG-FE eval board for a possible configuration.

    Regards,

    Tony Calabria

  • 0 in reply to Tony Calabria
    Intellectual 300 points

    Hi Tony,

    I think I'm not understanding you here, you're telling me that the common-mode range limit is set by the power supply and the RLD uses the midpoint between two power supplies to set the DC bias point. First, which power supplies? I think you're talking about (AVDD + AVSS) / 2 which in my case with unipolar supply equals to (AVDD + AGND) / 2. But for the ADC reference I'm using an external source of 2.5 V with the REF5025, connected to the VREFP pin of the ADS1299. If I set the DC operating point to midpoint of power supply and then I use this midpoint to bias the body, shouldn't the ADC readings be greater than the ADC code range, giving only lectures of EFFFFF? What I mean is that the midpoint of power supply is 2.5 V which is the same as the ADC reference I have chosen, giving this value as bias point for the body measurements.

    I'm also using RLD electrode, as I understand of what you have just told me, this electrode is going to set the bias point for the analog-inputs (electrodes) by a high impedance resistor, but in a previous post I understood that this configuration was necessary only when you're not using a RLD electrode at all. So, should I connect the analog inputs to RLDOUT by this 10MΩ resistor even when a RLD electrode is connected to it? Seems legit accordingly to the datasheet in page 18 figure 21, where the differential inputs are connected to a CM voltage.

    I have saw that is also very common to use a filter at the input of every channel (I guess is for EMI), should I build that? I also have look they use some clamp diodes, I don't see those really necessary, but if you recommend using them I could add this to the schematic. The finally interface between the electrodes and the ADS1299 will look like in the next image for two channels, I'm using the common-voltage of 1/2*Vref(ADC) for CM which is of 1.25 V. Again, is this correct, do you recommend using anything else?

    Thank you very much Tony, you have been a great help, I really appreciate this.

    Best regards, 

    Julián.

  • 0 in reply to Julian Grisales
    Intellectual 300 points

    Somebody? I really need some light here!

    Thanks in advance :)

  • 0 in reply to Julian Grisales
    Prodigy 110 points

    Hello Julian,

    i must admit that some of your questions I didn't understand very well, but here is my overall opinion about it:

    I think you should always use the RLD electrode to eliminate the common mode noise that comes from the body. In doing so, not only you get a cleaner signal, but also you can use that electrode as bias for multiple channels, and you don't have to use the differential amplifier at the inputs.

    Also,, the didoes are important for input protection in case of failure of the ads1299. Be careful to make sure to use input resistors that wont allow for a current higher than 50microA to go into the patient's body.

    Please somebody correct if I'm wrong. I'm sorry that I can't help you much more, but I have a lot of doubts myself since I'm using the ads1298 for the first time and there are not many references about EMG applications. 

    In that sense, can please someone describe how could I adapt the schemitic below of the user guide for EMG?

    Thank s in advance!


    Márcia Vagos  

  • 0 in reply to Marcia Vagos
    Prodigy 130 points

    Dear all,

    Can anyone help me on this.. which diodes and opamp to use in my design of ADS1198...

    Diode shall i use BAT54S schottky diode and opamp TLC2272..

    Also if anybody suggests part numbers as well that would be useful..

    Thanks in advance..

    Regards,

    Deebak chandar T

  • 0 in reply to Julian Grisales
    Prodigy 10 points

    hi can u send me full diagram of 12 lead ecg using AD1298

    Thanks & regards

    chintan

  • 0 in reply to Marcia Vagos
    Prodigy 70 points

    Hi Marcia,

    I'm trying to use the ADS1298 for EMG acquisition. I wonder if you ended up using the schematics you posted on December 2012 and if it works, especially on the interface part, between the ADS1298 and the electrodes. 

    I would kindly like to know what type of electrodes you chose to use and how efficient they were.

    Thanks in advance!

  • 0 in reply to Youssef Boksmati
    Prodigy 110 points

    Hi Youssef,

    I'm very sorry for the delay in the reply. To answer your question, the schematics I used were as shown in the picture below. That system was assemble and some one else tested it for EMG signal acquisition and it worked, so I guess the schematics are okay. Note that I used schottly diodes and resistors and capacitors with miniaturized packages to minimize noise.

    The type of electrodes I used were textile electrodes made of conductive fabric, but I guess that any type of conventional electrodes will work better! I dont know how was the SNR with the ADS1298 because i did not test it personally, but it problably will also depend a lot on the layout of the circuitry and on the precision of assembly. I am sorry but I cannot help you further.

    Best regards and good luck,

    Márcia

  • 0 in reply to Deebak Chandar1
    Prodigy 110 points

    Hi Deebak,

    there are many diodes with similar characteristics that you can use, so it depends on the application requirements. For example, for my application I used BAS70XY diodes and OPA336 op-amps, but BAT54S seems to be quite similar. There is a wide variety to choosen from:

    http://www.nxp.com/documents/data_sheet/BAS40_1PSXXSB4X_SER.pdf

    http://www.nxp.com/documents/data_sheet/BAS70_1PS7XSB70_SER.pdf

    http://www.ti.com/lit/ds/symlink/opa336.pdf

    You have different packages with different configurations and sizes, it all depends on how you want the layout and size of your system to be.

    Good luck :)

    Márcia

  • 0 in reply to Marcia Vagos
    Prodigy 70 points

    Marcia,

    Sorry for the delay but I just red your reply.

    Thank you so much for your answer! The schematics you sent will help me and all those who are trying to do EMG measurements with the ADS1298 chip!

    I will try to understand the schematics and will optimize it in order to use metallic electrodes instead of textile electrodes. I'll then get back to you with my version of the circuit and some explanations.

    Thanks again Marcia! Let's rock the EMG communitee!!

    Youssef

  • 0
    Prodigy 55 points

    Hi Julian,

    I want to add a pre-amp on the EEG electrode to work with ADS1299, and I find this picture

    it is similar to my design. I am confusing the connections of pre-amp now, so can you post the complete picture or send it to me? 

    Thank you !!

    Hao

  • 0 in reply to Marcia Vagos
    Prodigy 70 points
    Hi

    I am working on multichannel EMG signal acquisition for classification of hand movements.Did you get answers to your above query?Even I am facing same problem about the pin connections of ADS1298 for EMG application.
    If you can just help me out with the connections,it will be a great help.

    Thank you
  • 0 in reply to Marcia Vagos
    Prodigy 70 points
    Hi Marcia

    I am working on Multichannel EMG signal acquisition.I have similar questions in my mind as yours.Did you get answers to your above queries?