Hi Sir,
Customer are using internal temperature sensor. They used to use F2132 with stable 1~2C variation for every chip. But when porting to FR5849, they found the temperature will be variant 2~5C with different chip at same condition.
They are not sure if it is normal or not. Or any software tips for this?
Best regards
Jacky Xu
Hi Jacky,
very interesting case.
I have observed on MSP430FR5994 and MSP430FR6989 LaunchPads that their temperature sensors can show 2-3 degC temp. differences during 2-5-10 sec. periods. I do not thing that power dissipation on a level below 1mW can affect power thermal and temperature of these MCU so quickly and so much. Idea I currently have: I will check how these chips perform when power by a battery or LDO.
New FR58xx, FR59xx, FR69xx MCUs use different type sensor than the old F2132.
At 25 degC new sensor has Vsensor at about 765 mV and 2.5 mV/degC. These new MCUs device descriptor structures contain calibration values for 30°C ±3°C and 85°C ±3°C for each of the available reference voltage levels.
The old type F2132 sensors, at 25 degC, have Vsensor at 1085 mV (with offset +/- 100 mV) and 3.55 mV/degC. The sensors use different calibration.
<<< You: any software tips for this?
I would recommend to check if your customer is using Vref and Tenperature Sensor calibration for this new type sensors. I am pretty sure he does it.
FR5849 technical reference http://www.ti.com/lit/ds/symlink/msp430fr5849.pdf on page 48 says:
As you see device descriptor calibration values provide +/- 3 degC accuracy.
Adding ADC12_B errors shifts Temperature Sensor total error to 4-5 degC.
Maybe TMP116 is the answer.
**Attention** This is a public forum
