Sonsivri
 
*
Welcome, Guest. Please login or register.
Did you miss your activation email?
March 28, 2024, 09:33:59 21:33


Login with username, password and session length


Pages: [1]
Print
Author Topic: [request for suggestions] on a/d converter  (Read 4599 times)
0 Members and 1 Guest are viewing this topic.
sphinx
Hero Member
*****
Offline Offline

Posts: 899

Thank You
-Given: 604
-Receive: 260



« on: February 19, 2017, 08:52:15 08:52 »

i am looking for suggestions on a/d converter for a project volt ampere meter i wanted suggestions for. since i am not that
familiar with a/d converter i maybe don't know some manufacturers that makes them. i am looking for a 16 bit 0-65535
preferably 0-5v range where i can get around 8-10SPS feels like sufficient for my project with serial interface
i don't know if you need more info

/sphinx

i forgot to mention most preferable would be a 4 channel a/d but 2-channel will work too
« Last Edit: February 19, 2017, 09:28:30 09:28 by sphinx » Logged

laws of physics are not laws at all, just assumptions and formulas that work as long as we don't figure something new that wrecks the calculations. the infinite onion try to peel that one
titi
Active Member
***
Offline Offline

Posts: 229

Thank You
-Given: 3679
-Receive: 834



« Reply #1 on: February 19, 2017, 09:29:17 09:29 »

Hi sphinx,
For measuring current it is generally a complex task, because you need to amplify the low voltage across the shunt, so generally you will have some problems of offset due to the OP Amp, you need also a good voltage Ref for a good precision and stability of the measure.
At the end you will have some trimming to do to compensate AOP offset.

The solution I use is an low cost ADC of Microchip the MCP3421, its is an ADC 18-Bit I2C Interface with variable resolutions (15 SPS for 16 bit, 60 SPS 14bit, 3.75 SPS 18bit), it include of good Ref of 2.048v +-0.05% and a PGA with selectable GAIN (1,2,4,8)
I use it to measure the current in low side (shunt is between GND and the load) with a SHUNT of 0.1 Ohm 1% with PGA x8.
If you need an other channel for reading voltage it is possible to use the MCP3422 with 2 channels or MCP3424 4 channels (exists in SOIC package more easy to solder).
MCP3422

The only problem, is that the MCP3421 chip exists only in SMD SOT-23-6, but if SMD is not a problem it is the best candidate for this.

Some other chip can be used but needs good Ref or have not a PGA, like : LTC2400 (not cheap), ADS1115 (easy to find on eBay), ADS1118 ...

Best regards

See: http://interface.khm.de/index.php/lab-log/connect-a-mcp3421-18-bit-analog-to-digital-converter-to-an-arduino-board/

See: http://www.esp8266-projects.com/2015/04/18-bit-adc-mcp3421-i2c-driver-esp8266.html




A project that seams at what you want to do:


Site: http://www.elektroda.pl/rtvforum/topic2575274.html
« Last Edit: February 19, 2017, 10:18:53 10:18 by titi » Logged
mars01
V.I.P
Hero Member
*****
Offline Offline

Posts: 535

Thank You
-Given: 692
-Receive: 1756



« Reply #2 on: February 19, 2017, 11:01:03 11:01 »

i am looking for suggestions on a/d converter. i am looking for a 16 bit 0-65535
preferably 0-5v range where i can get around 8-10SPS

Hi,
If you want good answers you will need to share your design input info.

What is the intended range of the voltage, amperage of your power supply? I think you mentioned in another topic that you are using it to measure a power suply parameters.
Why do you need a 16bit ADC?
Are there any special needs? Because most people who are designing lower than 100V power supply's are happy to use a 12bit ADC, sometimes even 10bit ADC can be enough (especially if you use oversampling which will help with the noise, too).
« Last Edit: February 19, 2017, 11:03:53 11:03 by mars01 » Logged
sphinx
Hero Member
*****
Offline Offline

Posts: 899

Thank You
-Given: 604
-Receive: 260



« Reply #3 on: February 19, 2017, 11:04:32 11:04 »

the design is for 0-30 volt and 0-5 ampere
my thought that 16-bit would be nice for a bit more accuracy
Logged

laws of physics are not laws at all, just assumptions and formulas that work as long as we don't figure something new that wrecks the calculations. the infinite onion try to peel that one
mars01
V.I.P
Hero Member
*****
Offline Offline

Posts: 535

Thank You
-Given: 692
-Receive: 1756



« Reply #4 on: February 19, 2017, 12:28:13 12:28 »

Your biggest enemy when it comes to accuracy is the noise. Noise from surrounding components. Noise from reference voltage source. Noise from the ADC itself. Noise from the PCB layout. Thermal noise from resistors. And so on.

I don't think that 30V/65536 = 457uV accuracy is attainable in normal conditions, or useful. Better to stick to a 12bit converter which will allow you an accuracy with 30V/4096 =~ 7mV step. Which can be improved with a bit of oversampling  Wink

You can use a board like this https://www.adafruit.com/product/1083 with the ADS1015 ADC http://www.ti.com/lit/ds/symlink/ads1013.pdf

Source: http://www.ebay.com/sch/i.html?_from=R40&_trksid=p2050601.m570.l1313.TR0.TRC0.H0.XADS1015.TRS0&_nkw=ADS1015&_sacat=0

LE: Oversampling info from TI: http://www.ti.com/lit/an/slaa323/slaa323.pdf

  
« Last Edit: February 19, 2017, 12:43:23 12:43 by mars01 » Logged
sphinx
Hero Member
*****
Offline Offline

Posts: 899

Thank You
-Given: 604
-Receive: 260



« Reply #5 on: February 19, 2017, 01:04:51 13:04 »

looks like mcp3421 mcp3422-4 might be a good idea to work with because i can change the resolution from 12 to 18 bits, it's not that expensive for trials
and so far my experimentation with proteus look promising. i need to get the hardware so i can test in real life.
Logged

laws of physics are not laws at all, just assumptions and formulas that work as long as we don't figure something new that wrecks the calculations. the infinite onion try to peel that one
mars01
V.I.P
Hero Member
*****
Offline Offline

Posts: 535

Thank You
-Given: 692
-Receive: 1756



« Reply #6 on: February 19, 2017, 03:01:00 15:01 »

Those Microchip parts have an internal Vref of 2.048V so the input range is below the 0...5V range that you mentioned. And worth to mention that for the positive range that you are targeting you actually have half a digital range available (because it is implied you will use the ADC as single ended not differential).

For example if you set them to work as 12bit ADC and PGA = 1 you'll get conversions of 0 ... 2.047V (actually not exactly Vref = 2.048V since they say that the resolution at 12bit is 1mV) to digital between 0 and 2047 which means 11bit.

LE: the TI part is the same, though.
« Last Edit: February 19, 2017, 03:42:50 15:42 by mars01 » Logged
sphinx
Hero Member
*****
Offline Offline

Posts: 899

Thank You
-Given: 604
-Receive: 260



« Reply #7 on: February 19, 2017, 03:55:39 15:55 »

i knew that about the reference, since i am no master at programming and i would have liked to make the software in preferably in c like ccs,
i am not going to write libraries for that it's a bit over my head. i found some software in the libraries for arduino stole a bit here and there
and i got it working. i guess i gotta find some similar adc working for 0-5v as reference where i can it to work. this is just for a home project
so i figured mcp3424 out and i got it working now i got a next step to find another chip i can get that to work that has 0-5v as reference.

i am working on this trying my best, i sure appreciate all input and its not easy find something when one don't what to look for
when there is lots of parameters like spi i2c sps resolution vref ... ad so on

all your help is very much appreciated
« Last Edit: February 19, 2017, 04:22:45 16:22 by sphinx » Logged

laws of physics are not laws at all, just assumptions and formulas that work as long as we don't figure something new that wrecks the calculations. the infinite onion try to peel that one
Pages: [1]
Print
Jump to:  


DISCLAIMER
WE DONT HOST ANY ILLEGAL FILES ON THE SERVER
USE CONTACT US TO REPORT ILLEGAL FILES
ADMINISTRATORS CANNOT BE HELD RESPONSIBLE FOR USERS POSTS AND LINKS

... Copyright © 2003-2999 Sonsivri.to ...
Powered by SMF 1.1.18 | SMF © 2006-2009, Simple Machines LLC | HarzeM Dilber MC