ADC noise on STM32F413

[AlexisC]

Hello,
For a project (control motor application), we are using ADC from MCU STM32F413. We encounter some noise on ADC measurments. We observe to high standard deviation on our ADC result.

We show in errata doc of mcu  we an issue: "Internal noise impacting the ADC accuracy" and so we follow the AN4073 but standard deviation on adc measurment is still to high.

Finally, our analysis show that program execution from flash disturb our adc measurements. When we put code in RAM memory. We improve significantly our standard deviation and so noise. Can you have a feedback about this observation that flash execution disturb ADC measurement?

Also, we use an other MCU STM32F469 for an other project (same application, our software is similar) and on this we not observe to high standard deviation and so noise on our adc measurement but we see also in errata of MCU STM32F469 issue: "Internal noise impacting the ADC accuracy". Can you confirm me this issue in STM32F469 because I read the AN4073 and I don't see the STM32F469 is impacted.

Regards,

Alexis

[Issamos]

Hello @AlexisC 

As you can see in the same errata you are using, the flash parameters and configuration may have an impact on ADC accuracy. That's why is always better to benchmark your application.

For the STM32F469, you can use the AN4073. This AN is compatible with all STM32Fxx. The list of applicable products give some of the boards that the application used to test the characteristics.

Best regards.

II

[AlexisC]

Hello Issamos,
Thanks for your reply.

Just on other question: is it possible that the issue "Internal noise impacting the ADC accuracy" on STM32F469 is less impacated than STM32F413 by internal noise?
Because on STM32F469 we don't observed high noise with this MCU compare to STM32F413?

[Issamos]

Hello @AlexisC 

Yes, the noise doesn't have the same impact on all the boards. It can be different between boards due to different characteristics (design, components,...)

Advice: when you want to ask a new question, close the last topic by choosing a best answer and ask your question in another post to give him more visibility.

Best regards.

II

[AlexisC]

Hello,
I know that hardware board impact noise. I just wan't you know for the internal noise in ADC mentionned in errata document for STM32F413 and STM32F469; if you (ST) have characterize by a benchmark the noise to know if STM32F469 have less noise than STM32F413  on ADC. It will be interesting for us.

Thanks,

Alexis

[Issamos]

I'm not an ST employee and I don't have any idea about that. But I think if there is such a file, you can find it looking in the website of ST.

Best regards.

II

[Johi]

Have you considered the option of external noise? I had a similar problem and by changing the pin I use for analog readings the noise was much smaller. As it happened, the pin used was close to the RMII output of the MCU and crosstalk was at the base of my issue. 

[AlexisC]

Thanks for your reply.

We are pretty sure that our noise came from perturbation with flash acces as is mentionned in application note AN4073. Because when we put our code in the RAM instead FLASH, noise on our ADC measurements disappears. Also, we observe the noise in all ADC pin in our application.

[Issamos]

Hello @AlexisC 

I share the same opinion with you. Is relates probably to the flash as mentioned in the AN4073.

Best regards.

II

[AlexisC]

Hello,

Before closed this subject, can someone from ST employee can reply about my question if STM32F469 is less impacted to this internal issue than STM32F413.

Thanks,

Alexis