UCD3138OL64EVM-031: AD12 ADC READING at -40'C

Hi, sir,

At very low temperature, the contact from the device to socket may be not very good. There may be some voltage drop there. I think the accuracy is better if the device is soldered on PCB.

Thanks,

Sean

Hi  Sean：

thanks for your response;

we are using UCD3138RGCX in a new design and found low temperature ADC accuracy issue, that's why we did the test on EVM so it could be duplicated by others easily; 

on our board we did the same test, results like below;

UCD3138RGCX is powered by  TLV70233DBVT which is powered by PS with 4.3V; we bias AD12 , AD13 around 0.4V with 100nf decoupling cap and 1.69k/12.1k resistor divider from 3.3V; 

code

AdcRegs.ADCCTRL.bit.MAX_CONV = 1;
AdcRegs.ADCCTRL.bit.SINGLE_SWEEP = 1;
AdcRegs.ADCCTRL.bit.ADC_EN = 1;

AdcRegs.ADCSEQSEL0.bit.SEQ0 = 12;   IT'S 13 FOR AD13'S TEST
AdcRegs.ADCAVGCTRL.bit.AVG0_CONFIG = 2; /* Average 16 samples */
AdcRegs.ADCAVGCTRL.bit.AVG0_EN = 1; /* Module0 averaging enabled */

at 25'c, the ADCRESULT[0] reading is around 665 with AD12 dc voltage reading 0.40416V;

at -40'c,the ADCRESULT[0] reading is around 609 with AD12 dc voltage reading 0.40382V;

at 25'c, the ADCRESULT[0] reading is around 666 with AD13 dc voltage reading 0.40496V;

at -40'c,the ADCRESULT[0] reading is around 607 with AD13 dc voltage reading 0.40512V;

the result is almost the same as we got on UCD3138OL64EVM-031;

Hi, Sean,

Thanks for your explanation. If ADC is out of your spec at very low temperature, you may consider to calibrate ADC. You can send me an email (s-xu@ti.com) for your application, and I can share some ideas with you. 

Thanks,

Sean