How to solve the abnormal ecg channel output of VSM biosensor MAX30001?

Infineon / Mitsubishi / Fuji / Semikron / Eupec / IXYS

How to solve the abnormal ecg channel output of VSM biosensor MAX30001?

Posted Date: 2024-01-17

The ecg channel output of MAX30001 is abnormal

I made a development board using max30001g. It can read and write the registers of Max30001g normally.

I connected ECGP to the positive terminal of the signal generator, ECGN and ground to the negative terminal of the signal generator. The DC output of 600mV was not input, and the final result was the same and 0 (the signal generator is normal)

Input a sine wave with offset 600mv and Amplitude 800mv. The result is not 0, but the output is very strange. The result is as follows (ECGP is connected to the positive terminal of the signal generator, and ECGN and ground are connected to the negative terminal of the signal generator)

There is no isolation in the register between AFE and ECGN or ECGP

Your connections and settings for the signal generator are wrong. The differential mode AC amplitude of the input signal of MAX30001 must be within ±32.5mV, and your signal has reached 800mV. There are also requirements for the common mode component of the input signal (near VMID), and the differential mode DC amplitude is also required (within ±650mV). These settings you have exceed or are at the boundary of the range (generally the DC voltage of the input ECGP and ECGN signals are the same). Please check carefully with the manual.

I tested both the differential mode DC amplitude and AC amplitude according to the manual.

What I am most curious about is that no matter I give any positive or negative DC amplitude (within ±650mV, plus or minus a few millivolts to tens of millivolts) (the signal generator is ok and the signal generator can be measured directly between ECGP and ECGN voltage), the value I output has always been 97 points, 99 points, etc., but when I change the amplitude of the DC, my waveform will rise or fall to a certain amplitude, and then quickly return to 979899 points.In other words, no matter how the tributary amplitude changes, my output is fixed.

For AC measurements, I used square waves and positive-selected waves (the signal is 10Hz and the sampling rate is 128SPS). No matter how the amplitude changes, they keep jumping between 90 millivolts and negative 90 millivolts (the signal period is also wrong. ) (as shown on the left). The only time the period is correct is when the 30Hz square wave is input (of course the amplitude is not correct either) (as shown in the picture above)

I tried to directly measure my own ECG signal, and the results were the same as those of testing DC.

What would cause such a result (I once considered whether I wrote the program shift incorrectly, but I did not find the error)

Please tell me, is the ground of your signal generator connected to the earth, or is it floating ground? Is the ground of your board connected to the earth, or is it floating ground? I want to confirm whether the ground at both ends of your board is directly connected to the earth. I'm not asking you to connect directly here, but just to confirm. In addition, take an oscilloscope or multimeter, connect the negative pole to the GND of the board, and the positive pole to the signal input, and check the DC differential voltage?

The signal generator is a three-wire socket connected to the domestic 220V AC power, and the ground wire is connected to the earth.

The ground of the board is connected to the negative pole of ECGN and the signal generator, and the positive pole is connected to ECGP. There is no float ground'

I have tested it and the DC differential voltage is ok.

Can you confirm if your signal generator is floating ground? Although it has a three-pin plug connected to the ground wire, this only means that the shell of the signal generator is grounded. Whether its GND output itself is floating ground, you need to read its manual. This is actually very important and involves issues such as Lead Bias and VCM later. I recently bought one. Although it has a three-prong plug, it is floating ground.

The simplest way is, if you have MAX30001EVSYS, input your signal into the evaluation board. Can you see the normal waveform on the computer? I have some doubts about this. Because of your connection method, we basically don’t use this for testing.

I assume that your signal generator is definitely not floating ground (I am just assuming here, you have to confirm it in detail), then the signal generator CH+ outputs DC 0V, AC a small sinusoidal signal of about a few mV, and then CH- is grounded. , connected to the earth. Then for the MAX30001, the GND of the chip must not be connected to the ground, that is, the GND of the chip must be floating ground. At this time, it is required to use a battery to power the chip, and the chip cannot be connected to the computer through the USB interface (because the ground of the computer is connected to the earth). At the same time, the Lead Bias or Vcm of the chip must be enabled to bias the positive and negative input signal channels, and the DC voltage difference between the positive and negative terminals of the input channel and the GND of the chip should be biased to about Vmid.

Thank you for the reply! I don't have the MAX30001EVSYS board. I once used a lithium battery to power it, and its output was still messy.

I have also tested without a signal generator, empty circuit or short circuit ECGP and ECGN. I also tried to test the ECG directly, but my output was still messy.

The internal lead bias and calibration of my chip are not turned on. Is there any successful register configuration that I can learn from? I have also made an ad7768 board before and tested it with this signal generator. I think I can stop worrying about this signal generator. I can still read the register correctly, and the value of the INFO register is 0x521000. Is my MAX30001G chip genuine? And if the register chip is read correctly, it shouldn't be broken, right?

This problem has been bothering me for a month. My internal register configuration has been connected to the ECG AFE. No matter how the simple DC signal changes, my output will not change (although the AC signal will change, but it becomes nonsense). How can I test to know if my video is fake or bad?

Sorry, as an application engineer, it is not within my scope of responsibility to judge the authenticity of the chip. I only care about the correct use of the product. AD7768 and MAX30001 are not the same usage, and they are not the same product. If the register can be read normally, then the chip's communication is normal. Although you have repeatedly stressed that it has nothing to do with the connection method of the signal generator, in fact we have encountered many cases, 90% of which are caused by incorrect connection and configuration of the signal generator. So I still want to rule this out. The GND of your board is 0V, and the input ground of ECGN is 0V, which is beyond the specifications of the data sheet. Besides, to test on humans, lead bias must be applied. Can you please test, the signal generator CH- is grounded, CH+ outputs DC 0V, AC several mV signals, the board is changed to floating ground, and lead bias is used, can a normal waveform appear?

I tried what you said. The specific parameters are as shown below (some sinusoidal trends). The problem is summarized as follows:

When I tried a DC signal, the input was still basically 0 (no matter it was negative several hundred mV to positive several hundred mV)

When I change the DC signal from a positive value to a negative value, my output does not change the sign, and the output will change for a while and then return to the original value (similar to waveform observation, the line will rise and then quickly return to the original value) value)

When I tried 10Hz sine in the past, I basically couldn't see the waveform. Now when I input 1Hz, this positive wave trend will appear, but I found that the sampling rate of the waveform shown in the picture is all 256Hz (but the actual sampling rate in the left picture is 256Hz , right picture 128Hz)

When my input peak-to-peak value is 10mVpp, it looks like a sine wave (left picture). But when my peak-to-peak value is 28mVpp, the wave is deformed (right picture), and the overall amplitude is within 12mV. I have also used the positive wave selection test with a peak-to-peak value of 5mVpp, but basically nothing can be seen (the noise amplitude is too large), and the noise is between plus and minus 6mV like the two figures below.

What would cause a major disturbance in torture? And I also turned on the low-pass filter, 256Hz corresponds to the 40Hz low-pass filter. The one corresponding to 128 is about 28Hz low-pass filter

A low-pass filter may not be enough, a 50Hz notch filter is required.

Review Editor: Liu Qing

#solve #abnormal #ecg #channel #output #VSM #biosensor #MAX30001