You are on page 1of 24

SIGNAL TO NOISE (S/ N)

RATIO
Submitted to :- D R . J A G TA R S I N G H

GROUP-2
ANKIT SAXENA ( P G M S E 136013)
V I V E K G U P TA
ANAND PRAKASH (PGMSE136016)
VA I B H AV S H A R M A (PGMSE136024)
S H E M D E B BARMA (PGMSE136021)
(PGMSE136027)
I N T R O D U C TIO N
 Taguchi has proposed various performance measures known as
Signal-to-Noise (S/N) ratio for evaluating the performance of
engineering systems.
 The concept of signal (S) to noise (N) ratios has high visibility in
design of experiment circles due to the work of Taguchi to carry out
complete analysis.
 The signal is what you are measuring that is the result of the
presence of your action.
 Noise is extraneous information that can interfere with or alter the
signal.
 It can not be completely eliminated, but hopefully reduced. Noise is
considered random.
Squeal

Noise Wear

Energy
Heat
Transformation

Signal Slow Car

Fig. 1 Block diagram showing concept of S/N ratio


Since noise can not be eliminated (it is random), we are more
interested in the S/N ratio than the intensity of the noise.
S/N O B J E C T I V E ?
 Reduce as much of the noise as possible by carefully controlling
conditions
⚫ Temperature, power supply variations, etc.

 Increase the signal to noise ratio


⚫ More signal vs. noise means a lower Standard Deviation.
 More precise measurement

⚫ Lower St. Deviation means a better LOD


 Lower limits of detection
TYPES OF N OISE …..
 Chemical Noise:
⚫ Undesired chemical reactions.
⚫ Reaction/technique/instrument specific.

 Instrumental Noise:
⚫ Affects all types of instruments.
⚫ Can often be controlled physically (e.g. temp) or
electronically (software averaging).
Thermal (Johnson) Noise:

⚫ Thermal agitation of electrons affects their “smooth” flow.

⚫ Due to different velocities and movement of electrons in


electrical components.

⚫ Dependent upon both temperature and the range of


frequencies (frequency bandwidths) being
utilized.

⚫ Can be reduced by reducing temperature of electrical


components.
 Eliminated at absolute zero.
Flicker Noise:

⚫ Frequency dependent.
⚫ Magnitude is inversely proportional to frequency.
⚫ Significant at frequencies less than 100 Hz.
⚫ Results in long-term drift in electronic components.
⚫ Can be reduced by using resisters that are metallic,
wire wound.
Environmental noise:
⚫ Unlimited possible sources.
⚫ Can often be eliminated by eliminating the source.
 Other noise sources can not be eliminated.

⚫ Methods of eliminating it…


 Moving the instrument somewhere else.

 Isolating /conditioning the instruments power source.

 Controlling temperature in the room.

 Control expansion/contraction of
components in instrument.
 Eliminating interferences

 Stray light from windows, panels on


open instrument.
 Turning off radios, TV’s, other instruments.
There are many different S/N Ratios. The important
S/N ratios are:

 Nominal the Best


 Target the Best

 Smaller the Better

 Larger the Better

 Dynamic

 Classified Attribute.
NOMINAL-THE-BEST :
 This case arises when a specified value is MOST
desired, meaning that neither a smaller nor a larger value is
desirable.

n = 10 log10(square of mean/ variance)

Examples are;
(i)most parts in mechanical fittings have dimensions which
are nominal-the-best type.
(ii) Thickness should be uniform in deposition
/growth
/plating /etching.
Fig. 2 Robust Design without noise
Fig. 3 Robust Design with noise
SMALLER THE BETTER:

 This is usually the chosen S/N ratio for all undesirable


characteristics like " defects " etc. for which the ideal value is
zero.
n = -10 Log10 [ mean of sum of squares of measured data ]
 Also, when an ideal value is finite and its maximum or
minimum value is defined (like maximum purity is 100% or
maximum Tc is 92K or minimum time for making a telephone
connection is 1 sec) then the difference between measured
data and ideal value is expected to be as small as possible. The
generic form of S/N ratio then becomes
n = -10 Log10 [ mean of sum of squares of {measured -
ideal} ]
The S/N ratio
smaller for
the for the better
used
is the
situations where
target value such as
is
zero, response
computer automotive
time,
emission or corrosion.
S/N= -10 Log 1 0 (MSD)

Fig. 4 Automotive emission


 The negative sign is used to ensure that the largest value
gives the optimum value for the response variable and ,
therefore, Robust design.
 Mean standard deviation is given to show the
relationship to the loss function.
 Loss function is an ideal measure for quality of products
as it is shipped by the supplier to the customer.
 Minimizing loss function leads to quality improvement.
Larger the Better :-
 It is used where the largest value is desired such
as weld strength, gasoline mileage or yield.
 Mathematically the largest value is 00 like the
lost function. It is the reciprocal of smaller the
better .

n = -10 Log 1 0 [mean of sum squares of


reciprocal of measured data]
S/N RATIOS FOR DYNAMIC PROBLEMS

Types of Dynamic Problems:


 Continuous - Continuous type ( C - C )

 Continuous - Digital type ( C - D )

 Digital - Continuous type ( D - C )

 Digital - Digital type ( D - D )


CONTINUOUS - CONTINUOUS
TYPE PROBLEM
 Both signal factor and quality characteristics take
positive or negative values.

 When signal m = 0,
Quality characteristic = 0,

 Ideal function y = m

 Scaling factor exists to adjust slope (proportionality


constant) between y and m.
CONTINUOUS - D I G I TA L TYPE (C -D )
 temperature controller
 input temperature setting - continuous

 output of heating unit - ' on ' or ' off

 divide into two separate problems

one for ' on ' function


other for ' off ' function
 each one continuous -
continuous type or nominal
- the - best type problem
D I G I TA L - C O N T I N U O U S T Y P E ( D - C )

 digital to analog converter


 conversion to ' 0 ' and ' 1 ‘

 divide into two separate problems


one for ' 0 ' function
other for ' 1 ' function
 each one continuous - continuous
type or nominal
- the - best type problem
CONCLUSION
 Signal-to-Noise ratio is also one more contribution
of
Taguchi.
 We came across different types of noise which affect the
performance of the system.
 Thermal noise can only be eliminated if temperature is
absolute zero.
 Environmental noise can be eliminated by
certain methods as explained.
 Nominal-the-best is used when a specified value
is MOST desired.
 Smaller-the-better is used for undesirable values such as
defects, automotive emission, corrosion etc.
 Larger the better is used where the largest value
is desired such as weld strength, gasoline mileage or
yield.
 Taguchi strongly recommended this approach
for multiple runs.

You might also like