Chapter 4 Function Debugging

Functional debugging is the key to verifying whether the design is qualified, and it is also the stage where engineers spend the most time, and this process may need to be repeated many times. This chapter focuses on the debugging process.

The first job after the PCB board is back is to make a sample. Since the size of the PCB is small and the components are all packaged in 0201, we arrange machine stickers, so that the PCBA produced is more secure. The following picture is the actual picture posted:

After the PCBA is completed, the first step is to burn the program, and then it is step by step, and the function test is started. The function debugging is divided into two parts, the first part is the test and debugging of the functional parameters of the hardware; the second part is the software UI testing. This chapter mainly talks about the first part.

The main contents of the test of functional parameters are as follows:

1. Confirmation of chip working voltage

Before powering on, it is best to use a multimeter to test whether there is a short circuit at the power supply end to prevent accidentally burning the chip. If there is no problem, continue to the next step and test the voltage data of the chip. The following are the actual measured data:

Operating Voltage

The above data shows that the main chip is working normally.

2. Confirmation of charging current

Use a DC power supply of 5V to connect to the PCBA, and test the actual charging current. The current is about 50MA, indicating that this data is normal, and the size of the current can be set by software.

However, during the charging process, some colleagues reported that the chip was burned out, and measured the short circuit between the 5V and GND terminals of the PCBA with a multimeter. After investigation, it was found that it was the cause of the resistor R11, which was 0R, and The actual design requirement is 2.2R.

3. Confirmation of working current

The above measured data meet the design requirements.

4. PCBA RF conduction data

The above shows that the RF parameters of the PCBA are normal.

5. Lithium battery protection

The parameters are normal.

After the above tests, the hardware parameters of the PCBA are basically normal and meet the design requirements. Are you a little smug? Don’t be too happy, the challenge may be yet to come.

After the PCBA test was basically completed, the mold shell material that had been waiting for several weeks finally came out, and I couldn’t wait to arrange the installation, because there is still a very important function that has not been tested, that is the ANC effect. Because the ANC function needs to be installed on the whole machine before it can be listened to. As a result, after installing the whole machine, I listened to it and was very disappointed! Because there is no ANC effect at all, I asked a software engineer, the reason is that the ANC function has not been added, because ANC needs to be actually debugged before it is added to the software.

Give the prototype to the acoustic engineer and participate in the debugging together, which requires an acoustic room and equipment.

ANC debugging

Debugging is a hard job. Repeatedly debug one by one frequency, look at the curve, and then actually test the sense of hearing; there is a little feeling, but it does not meet the requirements. In summary, there are the following points:

1. After the ANC is activated, there is an obvious sense of sound pressure, just like the feeling of suddenly entering a tunnel when taking a train.

2. The effect of ANC is not obvious

3. There is noise after starting ANC.

After repeated debugging in the afternoon, the expected effect was not achieved, the work was forced to be interrupted, and the reason was started to be self-examined. The acoustic engineer gave suggestions:

  1. Check whether the speakers meet the requirements.
  2. Does silicon wheat meet the requirements.
  3. sealing problem.

Reinstall the next day, first verify the above three questions,

Horn curve

The curve of the speaker is good, the difference between high and low frequency (100HZ and 1KHZ) is 8-10DB, which meets the requirements. The sensitivity of the silicon microphone -38DB+/-1DB is also normal, and there is no problem with the sealing. Doing ANC debugging, after two hours of optimization, the problem still exists:

1. Sound pressure sense

2. There is ANC effect, but not very good

3. There is noise after starting ANC

The problem fell into a stalemate, and I couldn’t find a solution for a while, and I didn’t find the real reason. The debugging of ANC couldn’t go on for a while. After two days of calming down, let’s find the problem of noise from the hardware first, and test it in a non-ANC mode. Signal-to-noise ratio, the results are as follows:

Noise Floor & SNR & Distortion

The above data is not bad, there is no problem with the hardware! I began to doubt the software, but there was no evidence. When doing research and development, I often encountered some problems. The software said it was a hardware problem, and the hardware said it was a software problem. They quarreled with each other! No matter who the problem is, it must be well-founded to be convincing. It took a few days to sort out the ideas, from the schematic diagram to the datasheet of the chip, I carefully read and checked all the details, and found nothing At the same time, I also communicated with the FAC of the original factory, but still did not find a clue. I had no choice but to go back to my PCBA to find the reason. I suspected that the silicon microphone might be interfered, so I removed the MIC for the call. At that time, I found that the call function was normal. Now I understand that the software has reversed the call MIC setting: the position of the call MIC is set to the noise reduction microphone, and the position of the noise reduction microphone is set to the call MIC. So I found the software and FAE to communicate. After confirming, it is true that these two microphones can be modified according to needs.

After modifying the above parameters, reinstall the prototype according to the process, and adjust the ANC effect again, and the problem has been significantly improved. The following is the ANC curve of the ANC prototype:

It is finally possible to move on, and it is much easier, because the functional prototype can be delivered!

The next chapter will talk about how to go to mass production.

Leave a Reply

Your email address will not be published.