I am working with 12Bit ADC on custom board with MK22FX512AVLQ12 controller. I am trying to sample VREFH channel (channel no. 29) but I am obeserving considerable fluction of 7 counts (Please see the attached Sheet).
The supply voltage to the controller is stable 3.3V with no ripples. I have connected VDDA to VREFH and VSSA to VREFL. There are capacitors of value 0.01uF, 0.1uF (as mentioned in the reference manual) and 1 uF connected between the pins VREFH and VREFL.( Please see the attached snippet)
12 bit resolution, software trigger mode
ADC module clock: Bus clock with frequency 20MHz and frequency divider of 2
Longsampling disable, continuousConversion disabled, Hardware averaging disabled.
I am not able to understand the reason behind fluctuation in VREFH.
Please help in resolving the issue.
Why do you consider a fluctuation of 7 units "considerable"? This is a range of 7 units (0.17% variation) or, if you're measuring in a range of 0V to 3.3V, this is a variation of 5.6mV.
Normally, these results would be considered really good for an MCU, you've clearly done a good job with your PCB layout and power suplly. Personally I think a 10mV variation to be reasonable for a well designed PCB and you've almost halved that (actually, when I look at the graph in the spreadsheet the normal variation is between 4090 and 4095 you're doing quite a bit better that the 5.6mV).
If you're looking for less variation, then you should be looking at an external ADC with much higher precision (and much higher cost).