ADC Selection

[Fibo]

Hello,

I have some questions on how to go about selecting an ADC for an application. I would like to take a voltage measurement across a shunt.

1mOhm Shunt
500-1000 samples per second

If possible I'd like to be able to see 0-200A with a resolution of 5mA?

I think I'm just getting information overload from Google, but maybe a 24bit Sigma-Delta? I'm not sure... not sure how to start

thanks!

 

[Fibo]

On Tuesday, April 29, 2014 10:07:45 AM UTC-5, RobertMacy wrote:

On Mon, 28 Apr 2014 20:21:56 -0700, Fibo wrote:



...snip....



The 200A will probably never happen, and if it does it'll only be around

for a few msec.



I need to make two of these measurements as close to simultaneously as I

can get. At first I was thinking about using a sigma-delta, something

like the LTC2442, but I can't tell how quickly the MUX switches, or how

close (in time) the voltage readings are to each other. Do you know what

parameter on the datasheet tells me that information?



http://www.linear.com/product/LTC2442



or maybe a SAR with a mux input is a better option... maybe an 18bit SAR

with a MUX input...



thanks!



is this AC only? if so, use a cheap soundcard input, you get at least two

channels, simultaneous [within the delay of your input filtering

differences] know it's usually better than one tenth of 1/48000 ?? what

2uS, plus you can calibrate down closer. very useful if you're trying to

preserve phase to measure power through this thing. uh, be careful of AC

mains isolation etc.

No it's a DC application, power monitoring

 

[RobertMacy]

On Mon, 28 Apr 2014 20:21:56 -0700, Fibo <panfilero@gmail.com> wrote:

...snip....

The 200A will probably never happen, and if it does it'll only be around
for a few msec.

I need to make two of these measurements as close to simultaneously as I
can get. At first I was thinking about using a sigma-delta, something
like the LTC2442, but I can't tell how quickly the MUX switches, or how
close (in time) the voltage readings are to each other. Do you know what
parameter on the datasheet tells me that information?

http://www.linear.com/product/LTC2442

or maybe a SAR with a mux input is a better option... maybe an 18bit SAR
with a MUX input...

thanks!

is this AC only? if so, use a cheap soundcard input, you get at least two
channels, simultaneous [within the delay of your input filtering
differences] know it's usually better than one tenth of 1/48000 ?? what
2uS, plus you can calibrate down closer. very useful if you're trying to
preserve phase to measure power through this thing. uh, be careful of AC
mains isolation etc.

 

[Tim Wescott]

On Mon, 28 Apr 2014 20:21:56 -0700, Fibo wrote:

On Monday, April 28, 2014 5:38:49 PM UTC-5, Tim Wescott wrote:
On Mon, 28 Apr 2014 15:15:53 -0700, Fibo wrote:



Hello,



I have some questions on how to go about selecting an ADC for an

application. I would like to take a voltage measurement across a
shunt.



1mOhm Shunt 500-1000 samples per second



If possible I'd like to be able to see 0-200A with a resolution of
5mA?



I think I'm just getting information overload from Google, but maybe
a

24bit Sigma-Delta? I'm not sure... not sure how to start



thanks!



200A across a 1mOhm shunt generates 200mV and dissipates 40W -- can you

stand that much power in your shunt?



200A / 5mA = 40,000. That implies that a 16-bit ADC will have
sufficient

resolution, assuming that you're not misusing the word. You can expect

both precision and accuracy (yes, they're different words) to be worse

than the resolution.



You want to take your 200mV +/- 5uV signal from the shunt resistor and

amplify it by an order of magnitude to match the reference voltage of

your ADC. Assuming it's a 2.5V reference, that's a gain of 12.5.
You'll

need an amplifier whose offset and temperature drift are commensurate

with that 5uV resolution, even when it's sitting next to something with

200A flowing through it. Even if your resistor has one end at "ground"

you should contemplate how meaningful "ground" is when there's 200A

running around. Then you should make a Kelvin connection to the shunt,

and feed that into a differential amplifier that can take whatever sort

of common-mode voltage you've got on the shunt.



--



Tim Wescott

Wescott Design Services

http://www.wescottdesign.com

The 200A will probably never happen, and if it does it'll only be around
for a few msec.

I need to make two of these measurements as close to simultaneously as I
can get. At first I was thinking about using a sigma-delta, something
like the LTC2442, but I can't tell how quickly the MUX switches, or how
close (in time) the voltage readings are to each other. Do you know what
parameter on the datasheet tells me that information?

http://www.linear.com/product/LTC2442

or maybe a SAR with a mux input is a better option... maybe an 18bit SAR
with a MUX input...

thanks!

Unless otherwise stated, it's the conversion time. Sigma-delta
converters do a lot of tricks to increase accuracy at the expense of
latency and oddball filtering -- it's a good idea to read the data sheet
VERY CLOSELY, and make damned sure you know what you're looking at.

I'll try not to give you the full "use engineering units" scold, but "as
close to" isn't very informative -- one person's "so close it doesn't
matter" may be a year, while someone else's may be a picosecond. If you
really want to get as close to simultaneous as humanly possible, then
build a pair of sample & hold circuits (I assume you need to take
measurements on two separate resistors). Sample at the ideal instant,
then ADC convert at your leisure (relatively, because S&H circuits
drift). Simultaneity will be determined wholly by the characteristics
and matching of your S&H circuits.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

 

[Bob Masta]

On Tue, 29 Apr 2014 10:22:45 -0700 (PDT), Fibo
<panfilero@gmail.com> wrote:

On Tuesday, April 29, 2014 10:07:45 AM UTC-5, RobertMacy wrote:
On Mon, 28 Apr 2014 20:21:56 -0700, Fibo wrote:



...snip....



The 200A will probably never happen, and if it does it'll only be around

for a few msec.



I need to make two of these measurements as close to simultaneously as I

can get. At first I was thinking about using a sigma-delta, something

like the LTC2442, but I can't tell how quickly the MUX switches, or how

close (in time) the voltage readings are to each other. Do you know what

parameter on the datasheet tells me that information?



http://www.linear.com/product/LTC2442



or maybe a SAR with a mux input is a better option... maybe an 18bit SAR

with a MUX input...



thanks!



is this AC only? if so, use a cheap soundcard input, you get at least two

channels, simultaneous [within the delay of your input filtering

differences] know it's usually better than one tenth of 1/48000 ?? what

2uS, plus you can calibrate down closer. very useful if you're trying to

preserve phase to measure power through this thing. uh, be careful of AC

mains isolation etc.

No it's a DC application, power monitoring

You may still be able to take advantage of sound card
technology (simultaneous sampling on 2 channels). You can
either use the same sort of chips they use, or modify an
existing sound card by adding a diff-amp input as in
<http://www.daqarta.com/dw_ggnn.htm>.

Actually, it may be even simpler, in your case, if you only
need DC voltages above ground. The circuit discussed above
is to allow bipolar inputs to be referenced to the chip's
Vref as 0V, eliminating the AC coupling caps. But if you
simply eliminate the caps you should be able to measure 0-3V
or so directly. You will need to insure that your
measurement ground is the same as the chip/computer ground.

Note that the device mentioned is only about $15 from HDE
(via Amazon), so you may want to buy a couple to have a
backup in case your initial hacking goes astray. This may
still be a lot cheaper than trying to buy the separate ADC
and supporting chips, not to mention all the development
time.

Best regards,


Bob Masta

DAQARTA v7.50
Data AcQuisition And Real-Time Analysis
www.daqarta.com
Scope, Spectrum, Spectrogram, Sound Level Meter
Frequency Counter, Pitch Track, Pitch-to-MIDI
FREE Signal Generator, DaqMusiq generator
Science with your sound card!