SOLVED: How can I reduce ADC jitter

[Demon]

- did not touch VREFs,
What do you mean by this , you have changes pos vref to fvr buffer1- used FVR 4x with 5 VDC in,
which will turn out to be a mistake since pots use vdd as top input , by doing this you have lost 20% of their travel range...

I meant, did not touch VREF-, sorry.

If I use VREF+ = VDD, my ADC readings stop at 1000 or so.
If I use VREF+ = FVRx4, my readings reach 1023.

Weird.

...- used FOSC / 128,way too slow , while not out of acceptable range fosc/32 is more reasonable...

I've switched to 32 as well.

...used average of 16 ADC readings,

not sure about that , its not necessary either if every thing else is ok ...

If I don't use averages, I jump all over like this:

OldADC1:00847  ADCin:00844  Diff:00003
OldADC1:00844  ADCin:00847  Diff:00003
OldADC1:00847  ADCin:00844  Diff:00003
OldADC1:00844  ADCin:00848  Diff:00004
OldADC1:00848  ADCin:00845  Diff:00003
OldADC1:00845  ADCin:00848  Diff:00003
OldADC1:00848  ADCin:00844  Diff:00004

...- used USART to show results
- added PAUSEUS 1 after all ADC operations - THIS NEEDS TO BE VERIFIED IF IT'S NECESSARY, it was just to make sure things worked.
it does nothing useful ...

I removed them.

...my schema , added another supply rail straddling 10k pot , without series wiper resistor to an2
not no none nil caps across an inputs

typical raw readings

R 827    669
R 826    669
R 826    669
R 826    668
R 825    669
R 827    669
R 826    669
R 826    669
R 826    668
R 825    668
R 827    670
R 826    669
R 825    668
R 827    669
R 827    669

Envious emote...

...
DEFINE ADC_BITS 10 ' 10-bit Analog to digital
DEFINE ADC_SAMPLEUS 5 ' Set sampling time in uS...

Reduced mine from 50 down to 5 as well


I just noticed that my temps on the 16F18855 goes over 100c when ADC runs; that's insane. I have no other routines running. If I stop ADC, it comes back down to 30ies.




I took everything non-essential off the breadboard (just kept my 2 power LEDs):

#CONFIG
    __config _CONFIG1, _FEXTOSC_OFF & _RSTOSC_HFINT32 & _CLKOUTEN_OFF & _CSWEN_OFF & _FCMEN_ON
    __config _CONFIG2, _MCLRE_ON & _PWRTE_OFF & _LPBOREN_OFF & _BOREN_ON & _BORV_LO & _ZCD_OFF & _PPS1WAY_OFF & _STVREN_ON & _DEBUG_OFF
    __config _CONFIG3, _WDTCPS_WDTCPS_11 & _WDTE_OFF & _WDTCWS_WDTCWS_7 & _WDTCCS_LFINTOSC
    __config _CONFIG4, _WRT_OFF & _SCANE_available & _LVP_OFF
    __config _CONFIG5, _CP_OFF & _CPD_OFF
#ENDCONFIG

DEFINE OSC 32

DEFINE  ADC_BITS 10                 ' 10-bit Analog to digital
DEFINE  ADC_SAMPLEUS 5              ' Set sampling time in uS

DEFINE  HSER_RXREG PORTC
DEFINE  HSER_RXBIT 7
DEFINE  HSER_TXREG PORTC
DEFINE  HSER_TXBIT 6

DEFINE  HSER_RCSTA 90h              ' Enable serial port & continuous receive
DEFINE  HSER_TXSTA 24h              ' Enable transmit, BRGH = 1
Define  HSER_BAUD 115200
DEFINE  HSER_CLROERR 1              ' Clear overflow automatically
DEFINE  HSER_SPBRGH  0
DEFINE  HSER_SPBRG  68

;--- Setup registers -----------------------------------------------------------

BAUDCON.3 = 1                       ' Enable 16 bit baudrate generator

FVRCON = %10000011                  ' FIXED VOLTAGE REFERENCE CONTROL REGISTER
'   bit 7   FVREN: Fixed Voltage Reference Enable bit
'               1 = Fixed Voltage Reference is enabled
'   bit 1-0 ADFVR<1:0>: ADC FVR Buffer Gain Selection bit
'               11 = ADC FVR Buffer Gain is 4x, (4.096V)(2)

'ADCON0 = %10000100                  ' ADC CONTROL REGISTER 0
'   bit 7   ADON: ADC Enable bit
'               0 = ADC is disabled
'               1 = ADC is enabled
'   bit 2   ADFRM0: ADC results Format/alignment Selection
'               1 = ADRES and ADPREV data are right-justified
'               0 = ADRES and ADPREV data are left-justified, zero-filled

ADCLK = %00001111                   ' ADC CLOCK SELECTION REGISTER
'   bit 5-0 ADCCS<5:0>: ADC Conversion Clock Select bits
'               001111 = FOSC/32

ADREF = %00000011                   ' ADC REFERENCE SELECTION REGISTER
'   bit 4 ADNREF: ADC Negative Voltage Reference Selection bit
'               0 = VREF- is connected to AVSS
'   bit 1-0 ADPREF: ADC Positive Voltage Reference Selection bits
'               11 = VREF+ is connected to FVR_buffer 1
'               00 = VREF+ is connected to VDD

ANSELA = %00000010                      ' Pin A1 = ADC (B5K)
ANSELB = %00000000
ANSELC = %00000000

TRISA = %00000010                       ' Pin A1 = ADC input 1
TRISB = %00000000                       ' Pin B7 = ...not available, ICSPDAT
                                        ' Pin B6 = ...not available, ICSPCLK
TRISC = %11000000                       ' Pin C7 = RX      *** Datasheet requirement, INPUT ***
                                        ' Pin C6 = TX      *** Datasheet requirement, INPUT ***

ADCinput            var WORD
ADCdiff             var WORD
OldADC1             var WORD

    Pause 500                           ' Let PIC and LCD stabilize
    ADCinput = 0
    ADCdiff = 0
    OldADC1 = 9999

Mainloop:
 
rem                 ADC1, B5K pot

    adcin 1, ADCinput
    ADCinput = ADCinput >> 6                                ' tried left justified for fun
    
    if ADCinput < oldADC1 then                              ' Calculate change in ADC
        ADCdiff = oldadc1 - ADCinput
    else
        ADCdiff =  ADCinput - oldadc1
    endif

    IF  ADCdiff > 2 then                                    ' Check for Diff over 2
        hserout [   "OldADC1:",     DEC5 oldadc1, "  ", _
                    "ADCin:",       DEC5 ADCinput, "  ", _
                    "Diff:",        DEC5 ADCdiff, 10]                             
        while TXSTA.1 = 0               ' Check TRMT bit
        wend
        oldADC1 = ADCinput
    else
        if ADCinput <> oldadc1 then                         '   ADC value diff than previous
            if ADCinput = 0 then                            '     Reached end of rotation
                hserout [   "OldADC1:",     DEC5 oldadc1, "  ", _
                            "ADCin:",       DEC5 ADCinput, "  ", _
                            "Diff:",        DEC5 ADCdiff, "  ", 10]                             
                while TXSTA.1 = 0               ' Check TRMT bit
                wend
                oldADC1 = ADCinput
            else
                if ADCinput = 1023 then                     '     Reached start of rotation
                    hserout [   "OldADC1:",     DEC5 oldadc1, "  ", _
                                "ADCin:",       DEC5 ADCinput, "  ", _
                                "Diff:",        DEC5 ADCdiff, 10]                             
                    while TXSTA.1 = 0               ' Check TRMT bit
                    wend
                    oldADC1 = ADCinput
                endif
            endif
        endif
    endif

  GOTO Mainloop

end

I know I can join those Ifs to check for end of rotation, I just wanted to keep code simple to make sure everything was getting picked up.

Averages was why sometimes it would not reach 0 or 1023.

It works, but it has jumps of 3 now. Gonna see if more caps can reduce that.


No clue why ADC is running so hot, specs say 85c is normal operating temp.


The mounting holes are "slightly" warm, never noticed that before.

[Demon]

...The mounting holes are "slightly" warm, never noticed that before.

Attachment 9922 Attachment 9923

I just checked the mounting holes on my test LCD, the holes do the same thing and the LCD hasn't been connected in days.

Copper/bare metal is at a different temp than the rest so the scanner detects it.

[Demon]

Either my PIC is damaged, or something is whacked inside the breadboard.

105c and wasn't stopping.

#CONFIG
    __config _CONFIG1, _FEXTOSC_OFF & _RSTOSC_HFINT32 & _CLKOUTEN_OFF & _CSWEN_OFF & _FCMEN_ON
    __config _CONFIG2, _MCLRE_ON & _PWRTE_OFF & _LPBOREN_OFF & _BOREN_ON & _BORV_LO & _ZCD_OFF & _PPS1WAY_OFF & _STVREN_ON & _DEBUG_OFF
    __config _CONFIG3, _WDTCPS_WDTCPS_11 & _WDTE_OFF & _WDTCWS_WDTCWS_7 & _WDTCCS_LFINTOSC
    __config _CONFIG4, _WRT_OFF & _SCANE_available & _LVP_OFF
    __config _CONFIG5, _CP_OFF & _CPD_OFF
#ENDCONFIG
DEFINE OSC 4
ANSELA = %00000010                      ' Pin A1 = ADC (B5K)
ANSELB = %00000000
ANSELC = %00000000
TRISA = %00000010                       ' Pin A1 = ADC input 1
TRISB = %00000000                       ' Pin B7 = ...not available, ICSPDAT
                                        ' Pin B6 = ...not available, ICSPCLK
TRISC = %11000000                       ' Pin C7 = RX      *** Datasheet requirement, INPUT ***
                                        ' Pin C6 = TX      *** Datasheet requirement, INPUT ***
ADCinput            var WORD

    Pause 500                           ' Let PIC and LCD stabilize
    ADCinput = 0

Mainloop:
    adcinput = 1023
end

[richard]

i simply hate this , all these pins made into outputs , shorted to gnd , no attempt to set them low
at power up are they low? before you short it all out
worst idea ever for breadboarding , i leave all pins as inputs unless they need to be an output
all this care re floating inputs is complete bull shit. for a battery operated device it will save a few microwatts that's it.
for the finished manufactured product sure do it , for bread boarding its dangerous nonsense where every mistake will be a bad one

TRISA = %00000010                       ' Pin A1 = ADC input 1
TRISB = %00000000                       ' Pin B7 = ...not available, ICSPDAT
                                        ' Pin B6 = ...not available, ICSPCLK
TRISC = %11000000

I took everything non-essential off the breadboard (just kept my 2 power LEDs):

and all the dangerous grounds

[Demon]

I realized something was seriously wrong when that nothing circuit was taking 670mA. I added:

PORTA = %00000000
PORTB = %00000000
PORTC = %00000000

Before the ANSELx and TRISx and it's taking 20mA now, 33c with no ADC, just ADCinput set at 1023 and the HSEROUT loop.

[Demon]

... i leave all pins as inputs unless they need to be an output
all this care re floating inputs is complete bull shit....

I try to keep my development code as close to final as possible, cause I know I'll forget to change stuff back. At least this way I can fire and forget the pin initialization.

But yeah, major bonehead move on my part. I put the old PIC back in and it runs at 20mA too. Gonna leave that 1st PIC in, just in case I do some other numbnut thing.

Thanks mucho Richard.


EDIT: Now this warning will follow my code for the rest of the development.

PORTA = %00000000                       '-------------------------------------------' 
PORTB = %00000000                       '       ALWAYS SET PINS LOW FIRST           '
PORTC = %00000000                       '-------------------------------------------'