ADS1298: ECG signals look like

Hi Alex,

Thank you for your quick answer.

Your answer help me to clear things a bit now. As i known, as an ecg its amplitude should be 1mV - 3mV, well maybe a little bit more. But in my case it is about 18 mV on the R-peak, which is weird for me. You meant that it is because of the more higher amplitude the more higher signal to noise ratio (SNR). i know that this ratio can be deal in Sigma-Delta ADC, right? But can you pls explain me that more specific that why i do have 18 mV, what is the roll of signal to noise ratio here? I am not a master at this area.

Yes, i mean the PQRST-Peaks of an normal ecg. But as you said, those could be seen by increasing the sample rate.

Hi Van Tu,

Happy to help! 

What is your gain setting? I've known ECG signals to be as high as 5mV, but I agree that 18mV is a little high. So there's probably either gain somewhere in the signal chain (you could probe the input pins to check magnitude after your ECG cables, the internal PGA gain is enabled (default is 6) or something is incorrect in the code to voltage calculation. Check your LSB and code to voltage conversion equations using the BIOFAQ, perhaps your LSB size is 10x what it should be? 

Yes. The amount of noise in the system will be static, meaning that it is at a certain level and will not change. If your ECG signal is small, then the noise will have a greater impact on the measurement. If your ECG signal is large, then the same noise level will have less of an impact on the measurement. So having a larger ECG signal is actually a good thing.

Hi Alex,

My gain setting is 6. I dont think there is something incorrect in the code to voltage calculation because i got the Test Signal right with this calculation.

Oh right, your explaination  is really helpful.

To prove it, i will post my register settings and test signal, also the way that i calculated the voltage from hex codes, so that you can check it for me. Thank you very much.

Lets say we have a hex code: 3d3c, which is 15676 in dezimal. gain is 6, and Vref = 2,4 V, so

Hi Alex,

So what would you say then? Can i use the noise ratio problem to explain the 18mV peak?

one additional question: as you see i got the test signal right with 1 Hz frequency. But the frequency of my ECG signal seems wrong with data rate at 500 SPS right?

Hi Van Tu,

I don't understand your question. The gain setting of 6 is multiplying the signal magnitude by 6, explaining why you are seeing 18mV instead of 3mV like you expected. This is not necessarily a bad thing, as it improves your signal to noise ratio. If you want to plot the "real" ECG, simply divide your results by 6. 

I don't see anything wrong with the ECG frequency - are you expecting something different than what you are seeing? 

Hi Alex,

The thing is i already divided my results by 6 and got 18 mV.

According to the test signal, a periode of the test signal has exactly 500 Samples, so we have 1 Hz at 500 SPS data rate. The same to ecg signal, a priode of an ecg signal has 10 Samples, so how many Herts do we have at the same data rate 500 SPS? Could you pls tell me that?

Thank you very much for your answers.

Hi Van Tu,

I have made a mistake, there are issues with how you are converting your codes to voltages. 2.4V needs to be multiplied by 2, and the denominator should be ((2^24) -1) please see this post for clarification:

https://e2e.ti.com/support/data-converters/f/73/p/772488/2856893#2856893

Frequency of the ECG is 50Hz. 500SPS/10samples/peak

Hi Alex,

Thank you,

It actually gives the same result for LSB as i used the calculation above. So i dont know what could be a mistake.

Yes, 50 Hz is my result as well, but i thought 50 Hz is too big for an ecg signal. I think i am wrong.

Hi Van Tu,

My apologies, you are correct - there is no change. 

I agree that 50 beats per second is much too high, I would expect closer to 50-100 beats/minute. Was this from a simulator or a patient? 

The frequency of the test signal looks to be correct, nearly 1Hz - but the magnitude is not within spec. The magnitude should be 1mV with a tolerance of 2%. 0.747mV is closer to 25% tolerance. 

What are your TEST PACE pins connected to AVDD or Floating? If connected to AVDD, try to lift the pins and let them float. This is a known issue that sometimes leads to incorrect test signal magnitudes. 

I would focus on getting the test signal to show correctly before worrying too much about the ECG, since it is a known magnitude it should be a good baseline to work with. 

Check your ECG cables. Some electrodes are "active" such that they have internal gain stages built into the cables. 

The last thing to check is order of operations in your code to ensure that everything is being multiplied, and divided correctly. 

Hi Alex,

Thanks for your quick answers again,

These signals are from my body.

If we look closer to test signal, we have 2 mV peak to peak, so the magnitude is actually 1 mV, (-1,25 mV - 0,75 mV). I think its because of the off set in the device right? I think this is not a problem if i'm not wrong. The TEST PACE pins are floating by the way.

For the ECG signals i will see what i can do. But i have no idea actually.

Hi Van Tu,

Correct again - I apologize, I think I'm trying to find problems where there are not any. 

The last thing I can suggest is to connect another known signal source (i.e. signal generator) to the inputs and see if you get the correct output. This will verify that there is not an additional gain stage that is unexpected. If this test comes back without problems, then you know that the issue is limited to your ECG cables/electrodes. 

Hi Alex,

I really appreciate your help a lot.

Ok, i will figure out what i should do the best.

Thank you and have a nice day.

Van Tu, Hoang

Hi Van Tu,

I am happy to help. Please let me know how it goes and good luck!