[Jihai APM32E103VET6S MINI development board review] RTC test

常见泽1

[Jihai APM32E103VET6S MINI development board review] RTC test [Copy link]

 

APM32 has a built-in RTC as shown below

It says that you can choose LSI, so I want to test the accuracy of using LSI.

I have used ST products, and LSI products are not very accurate either. Some products require software calibration before they can be used.

Look at the APM32E103 manual

There is no typical value for LSI. I think the difference in LSI may be very large.

Look at the APM32F103 manual again

It is worthwhile for LSI, so APM32E103 should have been cut down in some way.

The program code is shown in the figure

I tested it with a stopwatch for about a minute.

The error is really large, almost more than twice.

lugl4313820

I tested it with a stopwatch for about a minute.

The error is really large, almost more than twice.

I learned a lot. Thanks for sharing!

Honghuzaitian

Is this large error caused by the MCU itself or is it related to the program or surrounding components?

常见泽1

Honghuzaitian posted on 2022-10-14 01:05 Is this large error caused by the MCU itself or is it related to the program or surrounding components?

The official program uses LSI APM32E103, which does not have LSI maximum and minimum values, while APM32F103 does.

怪叔叔

I just checked the Geehy official website, the E103xE data sheet is https://www.geehy.com/uploads/tool/APM32E103xCxE%20%E6%95%B0%E6%8D%AE%E6%89%8B%E5%86%8C%20V1.3.pdf, and the accuracy of LSI is provided.

怪叔叔

How to get accurate timing with LSI:

Analysis of the reasons for inaccurate data:

The default division factor is 32768. If you choose to use the external LSE clock 32.768KHZ, you can get a more accurate 1S clock timing. The calculation method is fLSE/division factor = counting frequency. 32768Hz/32767 is approximately equal to 1Hz.

When LSE is replaced by LSI, the clock of LSI varies between different chips, with a typical value of about 40KHz. You can choose a frequency division that is easy to measure, use the second interrupt to output the divided clock frequency, and invert the accurate LSI. Then adjust the frequency division coefficient based on the calculated LSI to get a relatively accurate 1Hz signal.

1. Use the program provided by the original poster to adjust the frequency division factor to 40. It is estimated that the second interrupt frequency can be around 1KHz. Use an oscilloscope to calculate or make a timer capturer yourself, capture the frequency, and then infer the real frequency of the LSI.

2. Modify the frequency division coefficient according to the actual frequency of LSI to obtain relatively stable RTC clock data

常见泽1

Uncle Weird posted on 2022-10-14 17:51 I just checked the Geehy official website, the data sheet of E103xE https://www.geehy.com/uploads/tool/APM32E103xCxE%20%E6%95%B0%E ...

The version provided in the forum is different from this version. Only the new version has

常见泽1

Strange Uncle posted on 2022-10-14 18:06 How to get accurate timing made by LSI: Analysis of the reasons for inaccurate data: The default division factor of the host is 32768. If you choose to use external...

There are many ways to do software compensation.

It's just that without calibration, the difference is not that big after using ST