How to use sweep frequency?

The last product development is the ECG monitoring aspect. Since the amplifier is a dedicated chip of ADI, the circuit principle design is basically a reference circuit recommended by this IC. One of the problems found was related to sweeping frequency. Just as someone in the forum asked about the frequency response, I simply introduced this item to you.

The acquisition of ECG signals is mainly amplification. The ECG signal is very weak and it needs to be amplified by about a thousand times. However, the frequency is only about 1 to 2 Hz (60 to 120 per minute). Each person has little difference. It is not in the hospital or on the way to the hospital. The heartbeat amplitude is mv, very weak, so The monitoring circuit has high requirements for noise suppression; another problem is that the amplitude of the ECG signal varies from person to person, ranging from 0.1mV to 5mV, and all of them are healthy people. There is a trade-off between circuit selection. It may be necessary to adjust the gain.

At the time, I only worked as a drawing board. The principle was designed by another group of engineers. It was slightly different from the data book recommendation. I asked a simple question at that time. They answered that it was tested and can be used, so I didn’t care Go to test and suffer a big loss. Simply understand that the power frequency noise is very large. At the same time, ADI's FAE also says that there are power frequency problems. So I added a simple passive trap circuit behind the circuit, and the frequency was set at 50Hz. This time used a tool: Multisim software. After the simulation of the software, the effect of the notch circuit is as shown below, and it still has some effects.

However, during the actual testing process, it was found that there is still a strong interference around the power frequency band of about 0.5 V. This is a very amazing thing, because according to the data sheet, ADI this IC has a great suppression of noise, and lithium When battery powered, how could such a big disturbance come from? Is notch circuit failure?

Based on the above questions, we purchased a very cheap home-made signal generator. Although it has only 300 pieces, it is powerful and the signal is clean. The frequency sweep function is very helpful for the analysis of our signals.

Sweep frequency interface as shown below:

The instrument has two channels can be set separately, by defining the starting frequency, termination frequency, scan mode, scan time, the channel will output the sweep signal as shown (from the network), and periodically repeated, so called sweep .

When 1mV, 0~100Hz sweep signal is input, periodic frequency response waveform can be obtained on the oscilloscope. If the waveform obtained before the trap circuit is as follows, it can be seen that at around 50 Hz, the signal is not attenuated but instead overflows.

After trapping the circuit, the unwanted frequency band is suppressed.

The problem lies in the IC circuit, which has two stages of amplifier circuits. The first stage test is basically the same as the description of the specification. However, there is a serious deviation in the frequency response and specifications of the second stage, as shown below.

After the simulation, it was found that the minor changes caused by the circuit design caused this effect. In the following two figures, the left picture shows the ADI original circuit simulation results, and the right figure shows the modified circuit simulation results, which are basically consistent with the measured data.

Since the reason is found, then changing the circuit back to the ADI recommended circuit should solve the problem and test it. In the next two figures, the left figure shows the ADI recommended circuit signal response and the right figure shows the error circuit signal response. It can be clearly seen that ADI's recommended circuit does not have high power frequency noise problems. The use of the original circuit does not require the presence of a trap circuit at all.

In the end, although the reasons were found, the process was very tortuous. Effective communication with others was a very difficult task. Moreover, if a worker wants to do something good, he must first sharpen his tools. If I don't use software simulation, there is no such convenient device for sweeping frequency. It is impossible to find problems quickly.

So what lessons? Do not trust anyone! Do not change the recommended circuit!