STM32F373CCT6 SDADC: Incorrect Readings with Internal VDDA Reference on One Board After Environmental Testing

[Madhavaraja]

Subject: STM32F373CCT6 SDADC: Wrong Readings with VDDA Reference on One Board After Environmental Testing

Hello ST Support Team,

We use the STM32F373CCT6 to read four analog inputs with SDADC1 and SDADC2. One of our boards developed a problem during environmental testing (temperature change, dry heat, cold, damp heat cyclic and steady state), where it ran in full-function mode throughout.

What happened

The board worked correctly until day 2 of the damp heat steady-state test (40 °C). Then all four channels read with an offset (for example, a 100 V input read as 50 V). After a power cycle it worked for 2 to 3 hours, then read full scale (3.3 V) even with no input. Reflashing the firmware did not help.

A couple of days later we debugged it with the same firmware (internal VDDA reference). With no input, each channel should read 1.65 V (confirmed with a multimeter at the MCU pin), but the firmware reads 1.85 V on all four channels. With the reference set to internal 1.8 V, the readings are correct.

Configuration

• Single-ended offset mode, gain 0.5
• MCU clock 72 MHz, SDADC clock 1.5 MHz
• VDDA (pin 9) and VDDSD (pin 24): 3.3 V
• Formula: voltage = (float)raw_value * (SDADC_VREF_VOLTAGE / 32768);

Questions

1. What could cause wrong readings with the VDDA reference while the internal 1.8 V reference reads correctly?
2. Can you share the SDADC internal architecture (especially the VDDA connection to the reference path) or a document to help us debug this?
3. Please advise on a solution.

Thank you for your support.

Best regards,

MADHAVARAJA RS

ILLUMINEN TECHNOLOGIES

 

 

[mƎALLEm]

Hello, ​@Madhavaraja and welcome to the ST community,

1- Is that issue is specific to one board or you have the same issue for all boards perceiving the same environment (temperature change, dry heat, cold, damp heat cyclic and steady state)?

2- What do you mean by “damp heat”? with humidity? if yes, did you apply coating to protect the electronics from the moisture?

Can you share the SDADC internal architecture (especially the VDDA connection to the reference path) or a document to help us debug this?

Nothing to share apart from what was provided in the documentation: datasheet, reference manual, errata sheet ..

PS: 

A couple of days later we debugged it with the same firmware (internal VDDA reference). With no input, each channel should read 1.65 V

We can’t rely on conversion of a floating input. You need to apply an external voltage.

[Madhavaraja]

Thank you for your response and support.

I have attached the complete failure analysis document for your reference. The details related to your previous question are included and highlighted in the document.

Kindly review the document and help us identify the root cause of the failure and suggest a suitable solution.

As we are planning to start production next month with around 1,000 units, your timely response and support would be greatly appreciated in resolving this issue before production.

If you require any further information or clarification, please let us know. We will be happy to provide the necessary details.

Thank you for your support.

[mƎALLEm]

Hello ​@Madhavaraja ,

1- Just to warn you this is not a ST peer to peer support channel. So your data and schematics are exposed to public.

2- U9 OpAmp is supplied by 5V and especially U9A OpAmp is connected to AIN0 (I suppose the ADC input). Are you sure you are not exceeding the allowed voltage for the ADC input?

3- Ok you have another behavior when you exposed the boards to an other environment (heat, humidity etc ..), but there are other components are involved. Who told you the issue is related to STM32? Did you monitor the U9A OpAmp output on TP19? Is the voltage output delivered by U9A OpAmp is what you expect in that environment? if you don’t have the evidence we cannot go further than that.