Analog Output scaling

[bed]

For my current Project I decided to measure and display the Voltage for an accu which is providing round about 3.7 Volts for a stepup module which supports the Project with 5 Volts.

The easy way is just measure the 3.x Volts and display the floating Value.
GreatCownBasic doesn't have floating point arhytmetrics Libray.
I you ask me, that is not needed anyway for most tasks.
So I implemented a simple somewhat pragmatic approch and publish it here for discussion.
The princip is scaling it up to the border of Integer.
Because Integer doesn't know any thing about Values < 1 we have to scale up. I did a rough measurement and my approach is correct enough, your milage will vary.

DIM ad_val,Voltage as integer 
DIM VOLTS,COLONVOLTS AS BYTE

'read 10-bit value to Variable
 ad_val = ReadAd10(an0)
'Voltage = (30000 /1024) * ad_val / 600 ' exact calculation
' but this would be much easier
VOLTS = ad_val / 200
COLONVOLTS = ad_val - VOLTS
GLCDPrint  65, 10, VOLTS
GLCDPrint  70, 10, "."
GLCDPrint  75, 10, COLONVOLTS
GLCDPrint  90, 10, "Volts" 

Have Fun!

I wonder if it is a good Idea for a new forum, called snippets?

[bed]

After rethinking I'found a flaw when measuring a Voltage like 3.01 Volts it would printed as 3.4 and not as 3.04
so one must handle it making 2 extra IF Queries
Such as

IF COLONVOLTS < 100 THEN
 GLCDPrint 70, 10, ".0"
 GLCDPrint  80, 10, COLONVOLTS
END IF
....

So my intention to save some code is absurd, I fear (200 was not correct either, 205 would fit better)

[stan cartwright]

Similar idea for Readad. I didn't like using goto. More decimal places would not mean more accurate.

  adcval=ReadAD(an0)
;
  adcval=adcval*100
  adcval=adcval/51
  volts=str(adcval)
  if adcval<10 then
    volts="0.0"+left(volts,1,1)
    goto volt_format_done
  end if
  if adcval<100 then
    volts="0."+mid(volts,1,2)
  else
    volts=left(volts,1,1)+"."+mid(volts,2,2)
  end if
  volt_format_done:
  GLCDPrint (50,10,volts,1) ;print Voltage

[bed]

Yeah, thank you. I do use this now. My first thought was to omitt String manipulation as far as possible. On the other hand unused RAM is wasted RAM :-)

[William Roth]

Your voltage display value will not be accurate given a 5V' ADC reference and a 3.7V input to the ADC.. There is a somewhat common formula for scaling a range of numbers that can be used with ADC. The formula is:
.
scaled_value = (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min;
.
Where x = the raw ADC value.

Since the ADC reference voltage is 5.00V, the output needs to be scaled so that so that 0 to 1023 IN returns 0 to 500 OUT.

The GCB code might look something like this:

#chip 16f1829, 32
#config FOSC_INTOSC , WDTE_OFF , PWRTE_OFF , MCLRE_OFF

 ;Software I2C For LCD
#define I2C_MODE Master
#define I2C_DATA PORTB.6
#define I2C_CLOCK PORTB.5
#define I2C_DISABLE_INTERRUPTS ON

#define LCD_IO 10
#define LCD_I2C_Address_1 0x7E ; default to 0x4E
#define LCD_SPEED FAST
#define LCD_Backlight_On_State  1
#define LCD_Backlight_Off_State 0
CLS

Dim ad_val  as  word
Dim Volts as byte
Dim hundredths as byte

Do

     ;get raw data
     AD_VAL = Readad10(An4)

    ;Scale/Map where 0.00 to 5.00V in  = 0 to 500 out
    AD_VAL = [long]AD_Val * 500 / 1023 ;casting to prevent overflow

   'Modify for display
    Volts = AD_VAL / 100   ;Get volts
    Hundredths = (AD_Val - (Volts * 100)) ;Get hundredths

    ;display the voltage as X.xx
    Locate 0,0
    Print Volts      ; 0 to 5
    Print "."        ;decimal point
    If hundredths < 10 then Print "0"
    Print hundredths
    LCDSPACE 2          ;clear any trailing data
Loop

[bed]

Yes, but no.
My Hardware ist beeing powered by a single Li Ion Accu, which is stepped up to 5V.
So the Reference is 5,x V not as exact as a real Reference, but my sample has 5.05 Volts which is near to perfect.
Or did I got you wrong? And you mean even when Power supply is 5Volts we have to calibrate?

But good to have such a formula in hand if it comes to that situation.

[William Roth]

I simply ran your code as posted and noted the errors. With your suggested code an input voltage of 2.0V displays "2.214". 3.70V displays "3.202".

  Volts = AD_VAL / 200

This is not scaling. It is simple division. While it will return 0 to 5 given an AD Value of 0 to 1023

COLONVOLTS = AD_VAL - Volts

I have no idea what this is supposed to do. It most certainly does not scale anything. And given that
"Volts" will be an integer between 0 and 5 , all it does is subract one of these integers from the raw "unscaled" ADC value returning a meaningless number.

Example1: Apply 2.5 V to the ADC Pin. The ADC will return ~511 given a 5.0V supply reference.
Volts will be 2. Colonvolts will be 511 - 2 or 509. Since colonvolts is a byte, it will OVERFLOW and actually be 251. 251 is meaningless

Example 2: Apply 3.7V to the ADC Pin. The ADC will read ~758 given a 5.0V supply reference.
Volts will be 3. Colonvolts will be 758 - 3 = 755. Since Colonvolts is a byte, it will OVERFLOW and actually be 243 . 243 is meaningless.

It is only by accident that the code is somewhat accurate at any given voltage. I suggest you rethink your method.

William

[bed]

William, you are right.
My first attempt was not thought through.
I thought about to use integer and not strings do build an reasonable Display.
Integer was an error anyway should have used word and ColonVolt was for the part <1Volt for ommiting string, which was not correct, also.
I have to admit that I worked with ADC and Controllers in whole worked >30 years ago as a hobby and and now after this long period I a bit overhelmed from the possibilities that the chips have developed to.
But William, now we have a good explained solution for this small problem, which surely will found by others newbies like me :-)