CCS/TM4C123GH6PM: TM4C123 ADC not stable

Hi Praveen,

Praveen Gonsalves said:

To check with the issue I used multimeter to check voltage fluctuations on RCV420 output, the it shows a constant voltage. The same pin when I connect to launchpad, the adc reading keeps on changing. 

Instead on multimeter, did you check the same on DSO? What do you observe?

Praveen Gonsalves said:

The changes in the voltage is in milli volts. Is there anyway to get a stable adc value??

You can add an external RC filter for stable value. As in software you can try normal averaging or moving average of ADC digital values. 

Regards,

Gautam

Hi,

The Launchpad is not an appropriate board to check/use such sensitive IC - the ADC reference is taken from VDD, and the AVSS is connected to GND. A more suitable would be a custom board, with a separate ADC reference and AVSS disconnected from GND. The RCV ground pin should be connected directly to AVSS, without any use of GND connection, and a careful designed board as indicated in fig.1 of the RCV data sheet.

To check and verify if this will have chances to work, just observe on LP a vias which connects AVSS to GND, and the ground connection from RCV should be soldered directly on that vias.

Do not forget the RCV has an output of 5.0V, unsuitable for ADC, so you should dived it first, but the resistor going to gnd should be connected to the gnd of RCV.

Agree w/both posters: Gautam & Petrei.    Use of a scope will reveal "variations" in the "RCV's" analog output - which will NOT be noted upon any "averaging" multi-meter.    (and - if captured & displayed - would "Make you crazy" w/"digit bounce!")   (firm/I past produced digital speedos/tachs (bounce proved unsettling)

You report, "Change of signal in millivolts - but do not provide a quantifiable value.   (there's huge difference between 5mV - and 250/500mV fluctuation - and that data has NOT been supplied!)

Two earlier posters - while advising well - did not direct that the, "Shortest possible connections - between RCV & your MCU - are required to lessen the arrival of "unwanted" signals - which may CAUSE - just what you report."    (you may even try a shielded cable - w/one side grounded - as a noise reduction technique.)

Should the RCV's output impedance NOT match the input impedance of the MCU - that too will invite issues.    (it becomes your job to check - and to determine the degree of "match.")    Usually the addition of a small value, ceramic cap - placed as close as possible to the MCU's targeted ADC pin - will help.   The R-C filter suggested will tend to "smooth variations" but at the cost of "slowed input signal response."

it is possible too - that ANY excess separation between your "Bridge" and the RCV IC - invites noise.   As quick/dirty test - you may attempt to shorten lead lengths - even shield that "Bridge-RCV" assembly.

As Petrei well notes - the LPad is designed primarily for a "price-point" - and ADC performance on "all such boards" is (usually) compromised...   I'd estimate that - w/Greater Attention to signal details - (noted by those two & here) you may reduce that signal variation by MORE than 50%.   (assumes the signal variance is >200mV)

Hello Gautam,

I do not have a DSO to check with it.

Will look into adding the RC filter and averaging the ADC values.

Regards
Praveen

Hello Petrei,

Will consider the design considerations you mentioned while designing my custom board.
Also will see if I can connect RCV420's ground to the vias of LP which connects AVSS to GND.

Right now have used resistors as mentioned in RCV420 datasheet to get output limit till 3v3.

Regards
Praveen