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.

sigma delta converter and how to form an n-bit output?

Status
Not open for further replies.

svensl

Full Member level 1
Full Member level 1
Joined
Mar 25, 2005
Messages
99
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,286
Activity points
2,202
I would like to know how the digital output (1-bit) from a sigma delta data converter can form an n-bit signal using decimation.

Thanks
 

to understad that you must read the basics of sigma-delta convertor; it`s all about noise shaping and filtering.
 

Hi svensl!

One method of implementation is an up/down counter + accumulator.

The up/down counter acts as a filter. Being triangular in time domain, it is sinc2 in freq domain.

The accumulator acc=acc+count if dig.o/p=1 else acc=acc

At the end of one up/down cycle the n-bit accumulator gives you the adc result! :)

You will have to play around with the decimation period - the up/down counter clock - the no.of bits etc.

Hope this helped.

Giri
 

Thanks for your response Giri. I am not entirely certain if I understood you correctly. If I take the 1bit output stream (0,1) from a SD modulator and decimate the signal with a sinc decimator, I average let's say 4 samples and divide by 4 before dumping the result to the output. So as an example if I have 0101010 then
using a moving average filter I average the first 4 sample 0101=2/4, then move one over and average the next 4 samples 1010=2/4, 0101=2/4 and finally the last four 1010=2/4. I will have 4 numbers 0.5,0.5,0.5,0.5. But how does this represent a 2 bit number, meaning how to get a binary number xxxx.

s
 

Ok. Suppose u want a 4 bit ADC o/p(binary), make a 2 bit up/down counter. (This method I believe is the simplest)

Suppose your signal is in the 50% range, so you get 0101010...

start with acc=counter=0

clock 1 : count=0 acc=0(here SD o/p=0)
clock 2 : count=1 acc=1(because SD o/p=1)
clock 3: count=2 acc=1(sd=0 here)
clock 4: count=3 acc=4(sd=1 here)
clock 5: count=2 acc=4
clock 6: count=1 acc=5
clock 7:count=0 acc=5 : 0101

If your SD mod gives all ones your acc o/p would be 9

And all zeros would give zero


So your full scale value is 9 for a 2-bit counter
Your acc value will be relative to this full scale value of 9.

If you have a large counter the 1 lsb error we get here will vanish.

I hope it is much clearer now :)

Giri
 

    svensl

    Points: 2
    Helpful Answer Positive Rating
Thanks for the clarification. I do know how it works now. In the example you gave you would decimate by 7 since you consider (average) seven 1-bit outputs from the SDM. Can you give further references on different implementation schemes, if you have them handy by any chance?

Thanks.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top