Help me with calculation of SNR

• Aug 29, 2007

• #1

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Hi
i need help regarding calculating SNR,

My signal is coming out of a RF receiver which demodulates it from higher freq to baseband,

the signal is noisy and i want to calculate its SNR,

i can generate the original received signal with simulations as well, or i can get the original transmitted signal form another chip(of same kind)

now how should i caluclate SNR, use the signal from simulations or use the one from the tx pin of receiver of the IC, and compare it with the noisy received signal.

Also there is this issue of how to get SNR by comparing a noisy signal with a noise free signal from the oscilloscope

any help is appreciated
br

 

• Aug 30, 2007

• #2

B

brmadhukar

Advanced Member level 3

Joined

Jun 21, 2002

Messages

839

Helped

42

Reputation

84

Reaction score

11

Trophy points

1,298

Location

India

Activity points

6,783

Re: Calculation of SNR

you need to find the mean and variance from the received signal and calculate the SNR. That is fairly standard method.

Hope this helps

 

• Aug 30, 2007

• #3

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Re: Calculation of SNR

well actually
i thought of the same in the beginning, i was thinking about taking the ratio of mean of the signal to its std deviation as SNR but the signal is zero mean and SNR is in -db which doesnot make sense, as the signal sure looks alright with slight fluctuations though :s

 

• Aug 31, 2007

• #4

A

arkamu

Junior Member level 2

Joined

Mar 1, 2007

Messages

23

Helped

0

Reputation

0

Reaction score

0

Trophy points

1,281

Activity points

1,381

Re: Calculation of SNR

bg

 

• Aug 31, 2007

• #5

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Re: Calculation of SNR

What is this document?
spam!!?

 

• Aug 31, 2007

• #6

B

bulx

Full Member level 4

Joined

Aug 7, 2004

Messages

223

Helped

47

Reputation

94

Reaction score

17

Trophy points

1,298

Activity points

1,520

Re: Calculation of SNR

Guess you want to measure SNR at RF reciever output.

Measure power of signal at RF recvr output using some equipment, you get signal+noise power (x). Now switch off transmission and repeat, you get noise power. Assuming all noise is additive and uncorrelated to signal, calculate SNR as x/y - 1. Agree?
-b

 

• Aug 31, 2007

• #7

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Re: Calculation of SNR

not a bad idea,

but can You calrify why x/y-1, where did this -1 comes from,



and also, the approach i have taken is that i calculated the psd of the signal going out from the tx pin of transmitter chip,


then i went to the receiver chip in the other building and found the psd of the received signal, then :

SNR= max(PSD at tx pin)- max(PSD at receive pin) (db)

is this approach ok?

 

• Aug 31, 2007

• #8

D

darock

Member level 2

Joined

Mar 4, 2007

Messages

48

Helped

17

Reputation

34

Reaction score

1

Trophy points

1,288

Activity points

1,543

Re: Calculation of SNR

-1 comes from the noise variance..

Do not transmit anything..
find the power of the received signal which gives you the noise power NP.


Now transmit the required signal
The signal power SP is
SP=Total Received power(RP) -NP
SNR=SP/NP
=RP/NP -1

 

• Sep 1, 2007

• #9

B

bulx

Full Member level 4

Joined

Aug 7, 2004

Messages

223

Helped

47

Reputation

94

Reaction score

17

Trophy points

1,298

Activity points

1,520

Re: Calculation of SNR

you method should work, but you have make some corrections

darock is, of course, correct, and so this has to be included into your calc, unless SNR is high.

secondly, PSD gives power at a certain frequency. Dont know why you take max. What max is this, is it maximum across frequency or max of different measurements done at different times?

Since we are required to find complete power, we sould sum up (integrate) PSD across all significant frequencies.
-b

 

• Sep 5, 2007

• #10

L

lmjandy

Junior Member level 1

Joined

Apr 9, 2007

Messages

16

Helped

0

Reputation

0

Reaction score

0

Trophy points

1,281

Activity points

1,351

Calculation of SNR

as bulx said

switch off transmission and repeat, you get noise power.

but if i can't get noise power,
and can get signal+noise power (x) only, how can i calc SNR???

 

• Sep 5, 2007

• #11

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Re: Calculation of SNR

actually i am getting same problem,

i can only access the signal once it has been converted back to baseband frequency, so if i stop trasnmissioon and monitor that prin, all i will get will be noise which has been passing through the AGC and will have same apparent power as the signal and this baffles me too,
how to get to noise power then?

 

• Sep 6, 2007

• #12

B

bizoo

Member level 5

Joined

Jun 22, 2005

Messages

90

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

2,057

Re: Calculation of SNR

Hi guys, there is a bit of blow here, the power meter i have been using has been erroneous, and it is giving big deviations from its previous values now


All i have the signals themselves inside the oscilloscope, and their PSD, is there a way of finding SNR with psd and not replying on these darned power meters

 

• Sep 10, 2007

• #13

B

bulx

Full Member level 4

Joined

Aug 7, 2004

Messages

223

Helped

47

Reputation

94

Reaction score

17

Trophy points

1,298

Activity points

1,520

Re: Calculation of SNR

I assume you have samples of the received signal for a long duration (1000s of symbol periods). I also assume you can find the begining of the symbol period. Now average the symbols over a long period and then you should get the signal amplitude fairly accurately, since the noise should average out. Now knowing the signal shape, amplitude and period, you can find its power. You can now subtract the recieved power from the signal power and get the noise power.

This of course assumes noise is additive and uncorrelated to the signal.
-b

 

• Feb 5, 2011

• #14

K

kapil411

Newbie level 5

Joined

Aug 11, 2007

Messages

9

Helped

0

Reputation

0

Reaction score

0

Trophy points

1,281

Activity points

1,355

Re: Calculation of SNR

hello

can any1 help me
i'm using cadence spectre

how many cycles i must simulate ADC to calculate SNR!!!!!
& how to calculate SNR,SNIAD !!!!!
which function i should use psd or dft function in caluculator!!!!!

 

• Feb 8, 2011

• #15

samiurrehman

Member level 2

Joined

Aug 30, 2010

Messages

49

Helped

2

Reputation

4

Reaction score

2

Trophy points

1,288

Location

Pakistan

Activity points

1,729

Re: Calculation of SNR

kapil411 said:

hello

can any1 help me
i'm using cadence spectre

how many cycles i must simulate ADC to calculate SNR!!!!!
& how to calculate SNR,SNIAD !!!!!
which function i should use psd or dft function in caluculator!!!!!

Click to expand...

I have the same problem, can someone upload some tutorial documents related to calculator functioning in Cadence.

 

• Feb 8, 2011

• #16

K

kapil411

Newbie level 5

Joined

Aug 11, 2007

Messages

9

Helped

0

Reputation

0

Reaction score

0

Trophy points

1,281

Activity points

1,355

Re: Calculation of SNR

samiurrehman said:

I have the same problem, can someone upload some tutorial documents related to calculator functioning in Cadence.

Click to expand...

how to come to conclusion about input frequency of ADC?????????
how to select bin for FFT???????
how many cycles 1 shud simulate inorder to find noise shaping??????