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.

architecture choice for 24 bit ,1Khz DAC

Status
Not open for further replies.

mitrobge

Member level 3
Joined
Feb 19, 2005
Messages
55
Helped
4
Reputation
8
Reaction score
1
Trophy points
1,288
Location
GREECE
Activity points
1,626
D/A of 24 bit resolution able to handle signals of 1 KHz ...

For these simple specifications which architecture would you prefer?

Current steering (binary,thermometer,segmented)

requires high matching=>too much area.

Sigma Delta is the only solution?

thanx
 

you will have to opt for a hybrid architecture...for 24 bit you can not go for a single architecture for all bits.
 

i think sigma-delta is the best solution.
 

What is a sigma-delta exactly??

--Knack
 

i wonder, if because of the low frequency other architectures (with high sampling frequencies) could be useful besides sigma delta?
 

at very first i would like to say that you have not goven the specs in totality..next thin blind faith in sigma delta doesnt worth as you just want to work on 1Khz it will be better to devide your design in thermometer coded then r-2r ladder and then some lsbs for current steering or any other no of bits you can decide by urself accoring to your other specs.
 

The problem about segmentation is that INL is independent of segmentation and depends only to number of bits (in my case 24 bits is too big).Thus, to acheive a good INL (0.5LSB) requires too big transisto crs to acheive low standard deviation of currents. So, to obtain good current matching, current sources should be enormous (resulting in a chip of cm order).I recall that Area is inversely proportional to the square of standard deviation.However,with segmentation DNL will drop. A 100% fully thermometer decoder will have very low DNL.

That' why i think about sigma delta which is inherently linear.
Is sth else?
 

sigma-delta is better choice
 

Yes. Delta sigma is the only choice.
Looks like I have not seen 24-bit using other architectures. Am I right?
Current sterring, R-2R all require large devices when it comes to MSB. Delta Sigma, instead, does not require such ridiculous multiplication.
i do think up to 16-bit, we can use other structures. More than than, we opt for delta sigma.

BTW., is the name delta sigma or sigma delta?
I think the real name should be delta sigma, am i right?
 

yeah,,delta sigma,,ive try it before.
 

hi
please post some docs on sigma delta DAC design
Thanks
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top