 |
 |
| View previous topic :: View next topic |
| Author |
Message |
temtronic
Joined: 01 Jul 2010
Posts: 9633
Location: Greensville,Ontario
|
|
 Posted: Thu Jul 09, 2015 5:17 am |
 |
|
I have to agree with Mt. T about PCB layout, power supply design, etc. in dealing with ADC chips!!
Simple test, connect Vin+ and Vin- to ground. sample and display 256 times. You should get 0 everytime or within +-1 or 2 bits. If not you have an 'analog' problem. Noise, layout, bad ground, EMI, the list is long....
Do the same test using a battery to see how the 'top end' of the ADC performs as well.
Designing rock stable ADC of more than 10 bits is half engineering, half 'art'. You can easily spend weeks getting the PCB 'almost right...
Jay |
|
 |
Ttelmah
Joined: 11 Mar 2010
Posts: 20061
|
|
| Posted: Thu Jul 09, 2015 7:35 am |
|
|
First, the routine should be returning signed. You are handing it a signed value but declaring the routine as returning unsigned (the default). Keep your signing tidy. Look at how I do the sign conversion. What I do is declare both a signed variable, and an unsigned variable to sit in the same memory space (the union). Then clock the data into the unsigned variable (it is unsigned as far as the logic goes), test the sign bit, extend this into the high bits as required, and then return the result as a signed variable.
Then remember what I said about /32.
The chip returns 5 bits after it's last useable bit. They call these extra bit 'sub Lsb' bits.
Then you are back to operating CS. Doing it this way you need to be operating in the Figure 5 clocking. Pulsing CS low, and testing if SDO (SDI on the PIC) is taken low (which signals that the conversion has completed). It'll go to sleep after the conversion has completed.
Seriously, again, why not use the SPI routines. Much simpler, and faster.
SPI version using CS:
| Code: |
#include "18F2520.h"
#fuses INTRC, NOWDT, NOPROTECT, BROWNOUT, PUT, NOLVP // Internal oscillator
#use delay(clock=1000000) // 1Mhz
#define A2D_CSI PIN_C1
#define A2D_SCLK PIN_C3
#define A2D_SDO PIN_C4
#use spi(MASTER, DI=A2D_SDO, CLK=A2D_SCLK, BAUD=100000, BITS=32, MODE=1, stream=LTC)
signed int32 Read_A2D(void)
{
int1 eoc;
union
{
signed int32 rval;
unsigned int32 raw;
} temp=0;
do
{
output_low(A2D_CSI);
eoc = input_state(A2D_SCO);
output_high(A2D_CSI); //clock in the EOC bit
} while (eoc); //wait for EOC to drop
output_low(A2D_CSI); //now drop select for the transfer
temp.raw=spi_xfer(LTC,0,32); //get 32bits
output_high(A2D_CSI);
//Now need to sign extend the value
if (bit_test(temp.raw,29))
{
//-ve value
temp.raw |=0xE0000000; //turn on top 3 bits
}
else
{
temp.raw &=0x01FFFFFFF; //only use low 29bits
}
return temp.rval/32;
}
void main(void)
{
signed int32 A2D_value;
signed int32 oversample;
int8 ctr;
setup_adc_ports(NO_ANALOGS|VSS_VDD); // No analogs
setup_adc(ADC_OFF|ADC_TAD_MUL_0);
output_high(A2D_CSI);
setup_comparator (NC_NC_NC_NC); // Disable comparator
//Now need to trigger a dummy read to start the conversion
A2D_value=Read_A2D();
while(1)
{
delay_ms(200); //delay before reading the first time
oversample=0;
for (ctr=0;ctr<64;ctr++)
oversample+=Read_A2D();
//Now perform simple *64 averaging.
A2D_value=oversample/64;
// display the adc value on lcd
//remember this is a _signed_ value
}
}
|
There was one big mistake in my previous code, I was only converting to a 16bit signed, not 32. Corrected here.
Now the unit returns a range of Vref (-0.5Vref to +0.5Vref), for a 25 bit result. 24bits for each half Vref range. Each count is just under 0.2uV. Now seriously you post the Vref to four decimals (implying a 5 digit DVM), but a _single_ count on your DVM, will be 500 counts on this chip!.....
Also, this will see signals far faster than your DVM. Most DVM's take a large fraction of a second to sample. This chip works hundreds of times faster, so the DVM, will not 'see' signals that this will record.
You are a long way from understanding just how hard 24bit operation is!.
I've corrected the sign extension, and added 64* averaging to this.
Note that I'm taking advantage of the fact that you can read the state of the SPI input pin when it is configured as SPI, without changing anything.  |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 242
Location: chennai
|
|
| Posted: Sat Jul 18, 2015 12:48 am |
|
|
Thanks Ttelmah. I have used your code.
Below are the results.
Value changes like 24212, 23622, 26192, 26997, 22319, 27021, ...
@temtronic: I used the Ttelmah code and IN+ and IN- are shorted to ground.
Results are -1195, -1203, -1226, -1206, -1205, -1181, -1190, ... |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 20061
|
|
| Posted: Sat Jul 18, 2015 12:57 am |
|
|
| Just 220uV difference between the two terminals. Varying by about +/-about 4.5uV. Not bad. |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 242
Location: chennai
|
|
| Posted: Tue Jul 28, 2015 12:26 am |
|
|
I am here with another problem with LTC2410.
Supply voltage: 3.3V
IN+ = 0.696V; IN- = 0.696V.
REF+ = 0.250V; REF- = 0V.
Output code from LTC2410 is saturated for this input voltage. Why is it so? Please help.
When i increase the voltage at IN+ by 50mV, the code remains the same. |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 20061
|
|
| Posted: Tue Jul 28, 2015 12:41 am |
|
|
Look at the data sheet.
What is the input range specified as relative to the Vref?.
"ANALOG INPUT RANGE –0.5VREF TO 0.5VREF"
Of course it is saturated. |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 242
Location: chennai
|
|
| Posted: Tue Jul 28, 2015 12:59 am |
|
|
In datasheet it say,
*The differential input voltage VIN = IN+ – IN–.
**The differential reference voltage VREF = REF+ – REF–.
Vin = 0.696 - 0.696 => 0V
Vref = 0.250-0 => 0.250V
-0.5* 0.25 => -0.125
+0.5*0.25 => +0.125
Vin is in between the -0.5 x Vref to 0.5* Vref.
Please correct if im wrong. |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 242
Location: chennai
|
|
| Posted: Tue Jul 28, 2015 9:51 pm |
|
|
output code from the LTC2410 is always
0b11111111111111111111111111111
Why it is behaving like this. Am I missing something? |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
