Maker Pro
Maker Pro

ADCs ... noise is it normal?

Hi all!

I`ve choosed TLC4541 16-bit ADC for my design.
One thing is driving me crazy ..when I connect Vin to Vref (+5V) and do
few measurements one by one i do get quite big noise(error)
here are some samples :

FE4E
FE7E
FE84
FE55
FEC9
FEC9
FE96
FE68
FE56
FE73
FE86

to get acceptable error(nose) [+/-4] i need to do averaging of 1000
samples
again some samples...:


FE85
FE83
FE84
FE85
FE83
FE84
FE84
FE84
FE83
FE85
FE85
FE84
FE85
FE84
FE84
FE84
FE84
FE85
FE85
FE84

my question is:
is it normal for ADCs, or I choosed a bad one?

best regards
DM
 
Hi all!

I`ve choosed TLC4541 16-bit ADC for my design.
One thing is driving me crazy ..when I connect Vin to Vref (+5V) and do
few measurements one by one i do get quite big noise(error)
here are some samples :

my question is:
is it normal for ADCs, or I choosed a bad one?

Have you hung an oscilloscope on the ostensibly stable voltage that
your A/D coverter is sampling? And on your reference voltage input?

With the ground clip of the oscilloscope probe attached to the 0V net
close to the poit where the A/D converter is connected to the 0V net
(pin 3) ...

Note that the LSB on your A/D converter output is only 76.3uV. The 123
LSB range on your raw data is a 10mV peak-to--peak excursion - you will
need to set the scope gain fairly high and use AC coupling if you want
to see this kevek of noise on an oscilloscope,

Unless your circuit has been carefully constructed, it is very likely
that your A/D converter is okay, and correctly representng the rather
noisy signal that you are giving it.
 
no i do not have osciloscope in here but i did like thist

i used tl431 to get 3V and connected it to Vref and Vin.. what i get is
... noise..
FF80
FFB8
FFB8
FFB8
FF8A
FFBE
FFA6
FF80
FFB8

any idea ?

best regards
david
 
R

Roger Hamlett

Hi all!

I`ve choosed TLC4541 16-bit ADC for my design.
One thing is driving me crazy ..when I connect Vin to Vref (+5V) and do
few measurements one by one i do get quite big noise(error)
here are some samples :

FE4E
<snipped>
You need to look at _everything_. Start with your Vref. How is this
generated?. How does it's ground connect to the ground of the ADC?. How is
it smoothed?. Then the same questions for the sample voltage. Then how are
the digital lines routed relative the analog signals?. You are actually
achieving about 10bit accuracy. The ADC you have, on a properly laid out
board, is capable of doing vastly better than this.

Best Wishes
 
M

Meindert Sprang

no i do not have osciloscope in here but i did like thist

i used tl431 to get 3V and connected it to Vref and Vin.. what i get is
.. noise..
FF80
FFB8
FFB8
FFB8
FF8A
FFBE
FFA6
FF80
FFB8

any idea ?

Your PCB layout and power supply are probably not designed properly. I
remember my first-ever 14ADC design: an ADSP2101 on a board, a power and
ground trace pair running from power supply to the DSP, the ADC and the
input amplifier (working in the mV ranges) in exactly that order. I think I
got about 8 bits of noise.....

Meindert
 
B

Bill Sloman

[email protected] wrote
Hi all

I`ve choosed TLC4541 16-bit ADC for my design
One thing is driving me crazy ..when I connect Vin to Vref (+5V) an d
few measurements one by one i do get quite big noise(error
here are some samples
<snip

my question is
is it normal for ADCs, or I choosed a bad one
Have you hung an oscilloscope on the ostensibly stable voltage tha
your A/D coverter is sampling? And on your reference voltage input

With the ground clip of the oscilloscope probe attached to the 0V ne
close to the poit where the A/D converter is connected to the 0V ne
(pin 3) ..

Note that the LSB on your A/D converter output is only 76.3uV. The 12
LSB range on your raw data is a 10mV peak-to--peak excursion - yo
wil
need to set the scope gain fairly high and use AC coupling if you wan
to see this kevek of noise on an oscilloscope

Unless your circuit has been carefully constructed, it is very likel
that your A/D converter is okay, and correctly representng the rathe
noisy signal that you are giving it
 
Top