Accurate Measurement Using PSoC5LP Delta-Sigma ADC

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
YaIs_283501
Level 3
Level 3
First like given
        I have a question about Delta-Sigma ADC. We were evaluated for each resolution 20bit, 19bit, 18bit of by DelSig_I2C project use of AN84783, but does not seem to be normally distributed If you look at the histogram of 18bit. Effective resolution's are significantly lower for this, but I can solve what if? Evaluation environment of I is as follows. -Use development board: CY8CKIT-050 · I have been short-circuited input http://jp.cypress.com/?rID=82678   
0 Likes
5 Replies
Bob_Marlowe
Level 10
Level 10
First like given 50 questions asked 10 questions asked

In the english manual for AN84783 here: www.cypress.com/ you find on page 10 a note concerning noise. Did you check and measure all of that discussed major noise sources and verified that each of them (and the sum) is smaller than your wanted noise level. Remember that 20 bit resolution @ Vref = 1.024V  is equivalent to ca. 1µV/Lsb. Checking noise and ripple is done best using an oscilloscope, AC nd high gain and look for the level of the overshoots.

   

 

   

Bob

0 Likes
YaIs_283501
Level 3
Level 3
First like given
        Thank you for reply. Though it is normal distribution with a resolution of 20bit, it is not a normal distribution it's a resolution of 18bit. By not become a normal distribution even when the resolution of 20bit if there is noise in the input? I think noise is the lowest input of the ADC because it short.   
0 Likes
Bob_Marlowe
Level 10
Level 10
First like given 50 questions asked 10 questions asked

...and the shorting wire is not able to catch up some noise? How did you wire the cap for bypassing the Vref??

   

We are talking about µV, this is quite less than what you get when you near your hand to a high impendance input. You shorted the inpiut pins. Shorted to what? Ground?? Shorted, but left floating??? The application note clearly states that even the noise margin of a resistor would be too much and that you need a precision voltage source to be able to measure the performance of the ADC correctly.

   

And precisely: 1.024V / 218  ~ 4µV/lsb

   

 

   

Bob

0 Likes
lock attach
Attachments are accessible only for community members.
ETRO_SSN583
Level 9
Level 9
250 likes received 100 sign-ins 5 likes given

Can you post you measured histogram plots for all three ranges ?

   

 

   

Attached a guide to meauring for the case input shorted. A scope irrelevant,

   

and in fact its noise path typically 8 bit digitizer, so even with averaging not

   

useful in accurate 20 bit measurments, let alone 16, for accurate signal path

   

noise measurements pre ADC inputs for these kinds of resolution.

   

 

   

Regards, Dana.

0 Likes
ETRO_SSN583
Level 9
Level 9
250 likes received 100 sign-ins 5 likes given

Assumed that you setup as follows(per ap note, note DC source not ground referenced) -

   

 

   

Of course you can make the measurements with 0 Vdc (differential in) by shorting

   

inputs, but use ohmmeter tool to establish short path R, which in turn adds thermal

   

noise at the input. Lastly note that when making rail measurements, you must have

   

the inpuit buffer of ADC bypassed in its properties, otherwise you are outside its CM

   

range.

   

 

   

0 Likes