ADS1293EVM: Base Line Wandering in ECG Signal

Hi,

This ticket is an extension from the ticket -

https://e2e.ti.com/support/data-converters-group/data-converters/f/data-converters-forum/1128782/ads1293evm-channel-3-signal-is-noisy-in-5-lead-ecg

In above thread/ticket, I suggested some possible solutions and registers settings that you may need to try and observe&compare the signal integrity. Please take a look and let me know.

Thanks

ChienChun Yang said:

This ticket is an extension from the ticket -

https://e2e.ti.com/support/data-converters-group/data-converters/f/data-converters-forum/1128782/ads1293evm-channel-3-signal-is-noisy-in-5-lead-ecg

In above thread/ticket, I suggested some possible solutions and registers settings that you may need to try and observe&compare the signal integrity. Please take a look and let me know.

Thanks

That ticket is resolved. ChienChun Yang Now kindly help me solve this please.

Hi 

I read  "connecting to a human being'?   So, all screenshots from here are from human subject(s)?

FYI, ADS1293EVM is not designed to connect to human body.

I see you usually acquire for 15 seconds. How much voltage does the baseline drift over 15 seconds? Could you adjust the y-axis scale? It could be small, but it's just the y scale is very fine when you set it to auto. Make sure the cables are steady not moving.

And, could you confirm that you don't see this wander/drift when using an ECG simulator?

Do you have or have you tried any ECG simulator that could output respiration signal? If yes, how did the results look like?

Do you enable and use the RLD electrode?

Do you try ask the subject hold breath and stay still to see whether the wander/drift improve?

Could you estimate or use the EVM GUI to estimate the drift/wander frequency?

Thanks

ChienChun Yang said:

I read  "connecting to a human being'?   So, all screenshots from here are from human subject(s)?

Yes all screenshots are from a particular human subject.

ChienChun Yang said:

FYI, ADS1293EVM is not designed to connect to human body.

Point noted. To minimize risks, I am disconnecting the AC Power adapter of my laptop while working on a human subject. And I have taken consent from the subject for experimental participation. 

ChienChun Yang said:

I see you usually acquire for 15 seconds. How much voltage does the baseline drift over 15 seconds? Could you adjust the y-axis scale? It could be small, but it's just the y scale is very fine when you set it to auto. Make sure the cables are steady not moving.

Yes the Y axis is set to auto and I got your point that the scale is very fine when in auto. So what range do you recommend me to set the Y axis scale? I will  make sure that the cables are steady and not moving.

ChienChun Yang said:

And, could you confirm that you don't see this wander/drift when using an ECG simulator?

Yes this drift is totally absent when using the ECG simulator.

ChienChun Yang said:

Do you have or have you tried any ECG simulator that could output respiration signal? If yes, how did the results look like?

Unfortunately I do not have any ECG simulator which can output the respiration signal. So I can not tell about the results.

ChienChun Yang said:

Do you enable and use the RLD electrode?

Yes I enable and use the RLD electrode. From my understanding the RLD injects a known voltage to the body to prevent it from remaining in floating condition (the body in floating condition acts as an antenna to picks up all sort of noises). So I enable the RLD and connect it to  IN4.

ChienChun Yang said:

Do you try ask the subject hold breath and stay still to see whether the wander/drift improve?

I have not asked the subject to do so. But in future tests I will do so to check whether the wander improves.

ChienChun Yang said:

Could you estimate or use the EVM GUI to estimate the drift/wander frequency?

The frequency is variable in different experiments. I have not maintained a log for the same. I will maintain that and post it here.

Finally recommend me a range to set for the Y axis so that I will use that to plot the graphs and post here.

Hi,

There is no definite Y axis range. You will need to observe the signal/data from your end to adjust the Y-axis range to keep the signal within the plot window.

I may suggest start from y axis range that are 2 times of the max peak-to-peak magnitude of your ECG signal.

Thanks

Thank you for the suggestion. I could not perform the experiments today as the subject is unavailable. But there is another thing which came to my notice just now.

While using a simulator, I do not find any error messages. All the fields are green

But in the images I have posted, you would find that LOD, RLDRAIL, CMOR these three fields show error in various combinations. What could be the reason of these error and could it have anything to do with the baseline wandering?

Thanks

Hi,

Thanks for pointing that out and Yes, they are related/relevant. Please see datasheet 8.3.11.2 Common-Mode Output Range (CMOR), 8.3.12, 8.3.14 and 8.3.15 and page 30.

Are Lead Off detection and/or PACE detection functions related registers enabled? If yes, could you turn those off first and see/observe what outcome is?

Not encouraged to use EVM on human body, but you may as well need to try different settings for RLD_BW and RLD_CAPDRIVE(datasheet page 48).

Also, Not encouraged and no offense, but, you may want to check and try different of the followings-

electrodes

cables

gel

patient skin conditions and/or any fur/hair.

different subjects.

Thanks

ChienChun Yang Thank you for your response. There is another issue that I think might be relevant. 

In the datasheet for the ADS1293EVM, the schematic of the connector is shown like this

Pins 10 and 11, I presume are connected to the body of the DB9 connector and is shown in the schematic to be connected to the GND. But while checking with a Multimeter, I did not find any continuity between the body of the connector and the GND pin test points. Also, in my ECG cable, the shield wire is not not connected to the body of the DB9 plug. So tell me one thing, is it required for the shield wire of the cable to be connected to the GND of the EVM? I am calling it GND because AGND and DGND are shorted in the EVM using R5, R6 and R14.

"did not find any continuity between the body of the connector and the GND pin test points"

Yes. they are not connected to AGND.

And, Yes, AGND and GND on EVM board are common nodes.

Is the ECG cable you are referring here the same one you used for the ECG simulator?  

It may not be required.

Could you try connect the ECG cable's shield wire to one of the GND TP ring on the EVM, and see how the result is?

Thanks.

ChienChun Yang said:

Is the ECG cable you are referring here the same one you used for the ECG simulator?  

Yes the same cable is used. 

ChienChun Yang said:

Could you try connect the ECG cable's shield wire to one of the GND TP ring on the EVM, and see how the result is?

Will check this out and post here. 

ChienChun Yang said:

Please see datasheet 8.3.11.2 Common-Mode Output Range (CMOR), 8.3.12, 8.3.14 and 8.3.15 and page 30.

I found out that my common mode detector is configured this way.

Since I am using 5 electrode ECG, I presume that the input from the chest electrode should also be enabled here right? And should I set CMDET_EN_IN4 to 1? Remember that the RLD output is connected to IN4 (SELRLD is set to 100 i.e IN4). Please help me here.

Hi,

Yes, you may need to play around different settings for CMDET_EN_INx as it may vary from systems to systems designs as well as electrodes.

In addition, please take a look for the following -

You may try different combination settings for datasheet 8.3.11 Common-Mode (CM) Detector and 8.3.11.1 Cable Shield Driving and 8.3.12 Right-Leg Drive (RLD)

e.g. try different settings for 

PACE2RLDIN

8.6.4 Common-Mode Detection and Right-Leg Drive Common-Mode Feedback Control Registers

CMDET_EN: Common-Mode Detect Enable

CMDET_CN: Common-Mode Detect Control

RLD_BW

RLD_CAPDRIVE

If your are designing a 5-lead system? For a 5 lead system, you may take a look of datasheet section 9.2.2 5-Lead ECG Application and note that

(1) The ideal values of R1, R2 and C1 will vary per system / application; typical values for these components are: R1 = 100 kΩ, R2 = 1 MΩ and C1 = 1.5 nF.

Thanks

Obtained the following result after setting PACE2RLDIN to 0 and using a different kind of patch electrode.  

Hi,

Appreciate for providing us the feedback.

If you don't mind sharing. What are the difference between the different patch electrodes?

Thanks