STM32L432 Internal Temp Sensor

[PCarn.1]

I'm trying to get the internal temperature sensor working on a custom board with an STM32L432KC. VDDA is 2.5V and is stable. The raw readings I'm getting are about 200 counts higher than I would expect on a 12bit config. The reading doesn't change based on temperature, usually I can see a 5 degree change if I shoot compressed air at the chip. I'm guessing I'm somehow reading a different ADC channel but I'm not exactly sure how to debug that. Attached are the register values after boot up.

[PCarn.1]

I'm also using 

__LL_ADC_CALC_TEMPERATURE

to get the temperature. It currently returns 79 degrees.

[AScha.3]

So what you get from ADC ? (in hex)

Maybe, its ok - because typical values are at 3.0 Vdda ->

but you have 2.5Vdda .

+

btw

Did you calibrate ADC ?

 

[waclawek.jan]

+1 to what @AScha.3  said above; some more possible sources of problems here.

JW

[PCarn.1]

I understand the calibration values might be off. But I should still see the raw counts changing with major temperature changes, but I don't.

[waclawek.jan]

> I understand the calibration values might be off.

Why do you think the problem are the calibration values? I pointed you to a list, did you exclude all other sources of problems?

> I should still see the raw counts changing with major temperature changes, but I don't.

So, you read out the same value over and over?

JW

 

[PCarn.1]

I don't think it's the calibration values. There is some variation in my readings but they are not affected by temperature changes. I have other ADC channels that are working so I know my ADC is on. There's a problem specific to the temperature sensor channel that I can't figure out. It almost seems like I'm reading a different a different ADC channel.

[waclawek.jan]

> There is some variation in my readings

How much?

> but they are not affected by temperature changes.

At all? And what are those temperature changes? Blowing air won't change internal temperature much, unless the chip is really hot on the surface.

Have you checked that all other items of the list I linked to are OK?

JW

[Andrew Neil]

---

@waclawek.jan wrote:

Blowing air won't change internal temperature much, unless the chip is really hot 

---

Or the air is significantly hotter (or colder) than the chip ...

[Andrew Neil]

---

@PCarn.1 wrote:

usually I can see a 5 degree change if I shoot compressed air at the chip.

---

Some more info on this "usually" - same chip? same board?

[PCarn.1]

12-bit resolution I'm seeing 1714 counts +/- 3 counts. I've set the temp sensor up on STM32G474s, F217s, L475s, and G431s, on all of them blowing compressed air on the chip can cool it down from it's ambient temp of 30C to 25C. 

Almost all of those points on the list can be ruled out when taking into account that the other ADC channels are working and not being concerned with conversion to degrees C yet. 

[waclawek.jan]

> I'm seeing 1714 counts +/- 3 counts

In the screenshot in your initial post I see DR=1439.

JW

[PCarn.1]

The one point in that list that does apply can be checked by the photo of the registers posted above. Ch17 is enabled.

[waclawek.jan]

Also, judging from ADC_CR.ADCALDIF being set, and the fact that ADC_CALFACT contains nonzero value only for the differential field, you probably did not perform single-ended calibration.

JW

[waclawek.jan]

Also, the 'L432 may run cooler (have lower power consumption) than the other chips, so blown air may have less influence at its die temperature.

JW