Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Why INL DNL analysis of ADC gets wrong with 00 logic value

Status
Not open for further replies.

yefj

Advanced Member level 4
Joined
Sep 12, 2019
Messages
1,190
Helped
1
Reputation
2
Reaction score
3
Trophy points
38
Activity points
7,181
Hello, bellow we have two cases of DNL INL analisys . the first one without 00 logic acts great the INL is the the sum of the previos DNL.
In the second case when i include 00 logic and we have 4 members. Its just get wrong?
Why its ruining the DNL INL behavior? Thanks.
1602356534494.png


1602356446072.png
 

The other bit-transitions are from:to, so
the calculations for (say) '10' involves the
values from the predecessor ("from") state
'01'. Nonlinearities require that you have a
line, which means two points, to criticize
against.

What is the predecessor for '00'? I can't think
of one.

I think the confusion is that there is no DNL
for '00' and the DNL for '01' has the implicit
reference to '00' value, '10' to '01', ....
 

Hello Dick_freebird, by the formula
DNL(N)=actual(N+1)-actual(N)-Savg /Savg
so DNL of 00 need 00 and 01 values not the previos of 00.
So by the formula we can do 00 of DNL.
Why if we have 00 our INL is not longer sum of previos DNL?
 

DNL for code 0 is undefined. And INL start with the transitions from code 0 to code 1 and there it is 0.
 
  • Like
Reactions: yefj

    yefj

    Points: 2
    Helpful Answer Positive Rating
Hello Sutapanaki, Bello we have another example.
When i disregard the 000 line and start from 001 INL DNL works great.
Regarding the offset which is actual(first)-idial(first) ,my offset first is 000 or should i calculated offset from 001?
Thanks.

1602490382932.png
 

The DNL of an N-bit ADC/DAC is always defined from 1 to 2^N-1. There is no sense in calculating the DNL of code 0 since DNL(i) = [Vt(i)-Vt(i-1)]/LSB, where Vt(i) is the transition level voltage of code i.
 

Is this for an ADC or a DAC. If it is an ADC and you say that for code 0 you have voltage 0.02, then what do you have below 0.02? If it is a DAC, then you don't have really a problem. Yes, there is an offset, but the way DNL, INL are calculated, offset doesn't matter. Offset is something that affects all codes, while DNL/INL are on a code to code basis.
 
Hello Sutapanaki, The second table is a DAC,you say that if its DAC then we count 000 too?
Why it willwork in this case?

In ADC i saw that if we dont calculate UNIFORM without 000 code then it works great and we have INL as a sum of DNL.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top