Kinetis MicrocontrollersのトピックRe: ADC Stability

ADC Stability

ali_asadzadeh — Tue, 17 Sep 2013 09:16:03 GMT

Hi,

I have connected a 10K multi turn pot to FRDM-KL25Z platform, I'm using keil and tried this code to read the Pot voltage, But my readings are not stable enough, they will jump around 10% of the value, How can I achieve the 14.5 ENOB from the ADC?

Here is my main code

int analog_in; char txt_buff[20]; SystemCoreClockUpdate();                      /* Get Core Clock Frequency */ uart0_init(); ADC_Init(); Calibrate_ADC();  while(1) {   analog_in=0;   for(int i=0;i<16;i++)   {   analog_in+=adc_read(0);   }   analog_in=analog_in/16;   sprintf(txt_buff,"analog = %d\r\n",analog_in);   uart0_puts(txt_buff); }

The code for ADC driver is attached,

Original Attachment has been moved to: ADC.h.txt.zip

Original Attachment has been moved to: ADC.c.txt.zip

Re: ADC Stability

chris_brown — Tue, 17 Sep 2013 13:26:46 GMT

Hi Ali,

The ADC on the KL25 requires a very low source resistance to attain the best results possible. I can't tell you specifically what to do to get to your requested ENOB, but I can tell you that you should make sure your source resistance is well below 5 kOhms. My recommendation would be to use a unity gain operational amplifier circuit to drive your ADC pin. Of course, even in this scenario, you will want to select an operational amplifier that has a low output resistance. In addition, you will want to make sure that the RC time constant of the operational amplifier output pin is less than 1ns. All of this information is documented in the KL25Z data sheet in chapter 6.6.1 (ADC electrical specifications) under note 4 of Table 19 16-bit ADC operating conditions.

Thanks,

Chris

Re: ADC Stability

ali_asadzadeh — Wed, 18 Sep 2013 06:35:34 GMT

I have added a unity gain op amp in front of The Pot, Using LM358, I get this results,It's better but They vary a lot too.

analog = 11639

analog = 11621

analog = 11801

analog = 11661

analog = 11669

analog = 11725

analog = 11677

analog = 11800

analog = 11577

analog = 11647

analog = 11578

analog = 11575

analog = 11423

analog = 11674

analog = 11613

analog = 11221

analog = 11517

analog = 11778

analog = 11635

analog = 11660

analog = 11674

analog = 11717

analog = 11903

analog = 11687

analog = 11817

analog = 11643

analog = 11710

analog = 11676

analog = 11703

analog = 11538

analog = 10996

analog = 11639

analog = 11702

analog = 11613

analog = 11727

analog = 11797

analog = 11083

analog = 11784

analog = 11682

This data gives me this info

Min value= 10996

Max Value = 11903

Mean Value= 11630

Max Error= 5.5%

And here is the Graph

Still I got very High errors ( it's worse than a 8bit ADC!) what is the problem?

Re: ADC Stability

ali_asadzadeh — Sat, 28 Sep 2013 07:11:18 GMT

Has anybody get stable results? I'm Still struggling getting a salable result,

Re: ADC Stability

Capps314 — Mon, 06 Jan 2014 13:58:06 GMT

I have had similar problems with the SAR ADC on a K20. Double check your calibration routine, if you can set a break point after it is completed and check the CALF bit manually. I had to use the COCO bit, instead of the CAL bit, on a K20 to make sure the calibration was completed. I also software average hundreds of ADC conversions before using a value. Also, the time constant of the input RC on the ADC pin is very important, make sure it is under 1nS.

Re: ADC Stability

yazexie — Fri, 24 Oct 2014 02:19:30 GMT

Ignore the calibration routine?