ADC produces very different values

[cyrusthevirus]

Not yet. That is the plan for tonight. I did not get a chance last night. Can I use the circuit above or do I need to go below 3.3 volts? BTW with the temp sensor, if I heat it up, the adc values go up. Not sure if that means anything.

 

[Mosaic]

The Vdd voltage of the PIC is often the default V ref unless u specify otherwise. I haven't read the datasheet tho.

To be certain that you are getting correct ADC values...just feed a known voltage < Vdd and compare the expected ADC value with the result from the PIC. If that matches then your problem lies elsewhere.

 

[cyrusthevirus]

Well, I tired an know voltage and even rewired another pic and I am still getting the same value to my LCD, which is different than my prototype. Can someone review the code? I am not sure what to do next. Two pics, same value, but different than the prototype.

 

[Mosaic]

You didn't answer the question.

Did the known voltage calculate to match the ADC value?

 

[cyrusthevirus]

But my ADC is calculated. If I follow what you put in the first post what value do I watch? How do I know that I am sampling at 8-bits?

Do I change this:

void showADC(void) {
	binary_temp =0;
	int voltage =0;
	z=0;
	for (z =0; z<64 ; z++) 
		{
			delay_5msec();
	   		adcvalue = readADC(4);    //select input channel an4 pin 21
   			binary_temp = binary_temp+adcvalue;
		}   
	binary_temp = binary_temp>>6; //get average value
	voltage =0;
	voltage =binary_temp*32;
	voltage = voltage/100;   // display 2xADRESH for thermo vlaue
	binary_to_ASCIIconvert(voltage);
	binary_temp =0;
	}

to this?

void showADC(void) {
	binary_temp =0;
	int voltage =0;
	z=0;
	for (z =0; z<64 ; z++) 
		{
			delay_5msec();
	   		adcvalue = readADC(4);    //select input channel an4 pin 21
   			binary_temp = binary_temp+adcvalue;
		}   
	binary_temp = binary_temp>>6; //get average value
	voltage =0;
	voltage =binary_temp/3.3; //3.3 ref voltage (I expect binary_temp to be 2V)
	voltage = voltage*256;   // 8-bit sampling
	binary_to_ASCIIconvert(voltage);
	binary_temp =0;
	}

Or is the code changed in the readADC?

 

[be80be]

Post your whole code

 

[Mosaic]

Even if your ADC samples at 9 -16 bits , once u use the hi 8 bits (left aligned) you are working with an 8 bit sample giving 2^8 = 256 discrete values.

I see u are using the ADRESH data which implies you are using no more than 8 bits, otherwise u would be using the ADRESL data as well.

DID u configure the ADRESH/L registers to be left aligned or right aligned?
Getting that wrong will confuse the ADC results.

When i say 'calculate' the binary ADC value i MEAN MANUALLY do it.

So if u feed in a fixed 2V to the adc input AND u have a 3.3Vdc ref then your ADC value is supposed to be (2/3.3) * 256 ( for an 8 bit adc value)
= 155 .

You can verify this within your code by adding a simple test and lighting an LED. Simply test if for a 2V input: 153<ADC_8bit_SAMPLE<157. Light the LED if it passes.

Then u can apply say a 1V or 3V signal and see if the LED does NOT light, thus verifying the ADC is operational in hardware.

 

[cyrusthevirus]

I tried to post the raw files, but that did not work. I am going to attach them as files.

 

[cyrusthevirus]

Okay, there are a lot of files, but the are all below. I am very grateful for everyone's help. I am so close to finishing this project until I hit this snag.

 

[cyrusthevirus]

Even if your ADC samples at 9 -16 bits , once u use the hi 8 bits (left aligned) you are working with an 8 bit sample giving 2^8 = 256 discrete values.

I see u are using the ADRESH data which implies you are using no more than 8 bits, otherwise u would be using the ADRESL data as well.

DID u configure the ADRESH/L registers to be left aligned or right aligned?
Getting that wrong will confuse the ADC results.

When i say 'calculate' the binary ADC value i MEAN MANUALLY do it.

So if u feed in a fixed 2V to the adc input AND u have a 3.3Vdc ref then your ADC value is supposed to be (2/3.3) * 256 ( for an 8 bit adc value)
= 155 .

You can verify this within your code by adding a simple test and lighting an LED. Simply test if for a 2V input: 153<ADC_8bit_SAMPLE<157. Light the LED if it passes.

Then u can apply say a 1V or 3V signal and see if the LED does NOT light, thus verifying the ADC is operational in hardware.

Okay, but I am still not sure what value to compare to light the led. For the voltage, can I use a lm317 with and input voltage of 5V and (2) 220 ohm resistors to get a contact output of 2.5v for testing?

 

[Mosaic]

Yes u can...just measure the voltage with a DMM 1st...it must be < the PIC vdd or bye bye ADC. Prhaps u should put a 1K in series withthe ADC input b4 u give it your ref voltage.

2.5V will work out to 2.5/3.3 * 256 = 193 or 194 ADC value.

 

[cyrusthevirus]

Okay got your post after I started testing. I can test more tomorrow night and want to test the other two wired pics.

For 2.58V I should get a ADC value of 2.58/3.3 * 256 = 200. I am getting 385

For 3.3V I should get a ADC value of 3.3/3.3 * 256 = 256. I am getting 1010.

DID u configure the ADRESH/L registers to be left aligned or right aligned?
Getting that wrong will confuse the ADC results.

Could this be the problem? How do I configure them?

 

[be80be]

LOL
ADC is 0=0 and 3.3 would 1024 on a 3.3 chip just like it's with any chip
Have a look at page 199 of the data sheet https://www.electro-tech-online.com/custompdfs/2010/11/39881D.pdf

 

[cyrusthevirus]

LOL
ADC is 0=0 and 3.3 would 1024 on a 3.3 chip just like it's with any chip
Have a look at page 199 of the data sheet https://www.electro-tech-online.com/custompdfs/2010/11/39881D-1.pdf

Okay so that is a good sign (i think) how about the other value?

 

[be80be]

For 2.58V I should get

should be more like 771

Your reading more like a volt

I would put a pot with one side to gnd and top side to vdd and the wiper to your adc input

You can set it at any volts from 0 to 3.3 volts

Like this

 

[cyrusthevirus]

I used a LM317 with two 220ohm 1/4W resistors. Vdd out on my DMM reads 2.58V. The pot is a good idea and I believe I have an extra, but I would still have to use my DMM to check the output correct? What would the pot provide that the LM317 is not?

Do I need a common ground between the voltage source and the ADC?

 

[be80be]

You can test a larger range of volts. And I think you'll see whats going on.

See your reading a word and trying to change it into a byte.

That's where your getting your error

Do I need a common ground between the voltage source and the ADC?

Yes

 

[cyrusthevirus]

You can test a larger range of volts. And I think you'll see whats going on.

Okay, thanks for explaining. I will test it tonight and post my results.

See your reading a word and trying to change it into a byte.
That's where your getting your error

Sorry I do not understand.

I had a common ground, just wanted to make sure they did not have to be isolated.

 

[be80be]

Let me look at your ADC code and I'll post some thing to show you. I'm not sure if you have this set right

bit 9-8 FORM1:FORM0: Data Output Format bits
11 = Signed fractional (sddd dddd dd00 0000)
10 = Fractional (dddd dddd dd00 0000)
01 = Signed integer (ssss sssd dddd dddd)
00 = Integer (0000 00dd dddd dddd)

You Data Output Format bits are set for Signed fractional

 

[cyrusthevirus]

I have some new, much shorter code. Let me get it and post it for you.