Mechanical Measurement Systems

Prof. Ravi Kumar

Department of Mechanical and Industrial Engineering
Indian Institute of Technology, Roorkee

Lecture – 24
Compensation

Hello! I welcome you all in this course on Mechanical Measurement Systems. Today, we
will discuss about the compensation in measuring systems, compensation in the process
of improving the dynamic characteristics of measuring system.

(Refer Slide Time: 00:32)

Topics to be covered in today's lecture are, compensation of a first order system,

compensation with the second order system and I will take one example if the time is
left.
(Refer Slide Time: 00:42)

So, in a first order system, as we know the response of the first order system can be
written as.

(Refer Slide Time: 00:48)

This is the governing equation for the first order system and in this equation, if I want to
improve the dynamic response of the first order system, this λ has to be changed. In fact,
lambda has to be reduced; time constant has to be reduced in order to improve the
response of the system right.
So, in order to do that, suppose we are getting output form certain input on certain
measurement.

(Refer Slide Time: 01:29)

There is a suppose some fluid and the temperature is the fluid is being measured with the
help of a thermocouple, thermocouple is giving output let us say output is, so output of
this thermocouple is let us say V 1 right.

Now what we will do; we will impose another voltage V 2 in this output and this V 2 is
here and there is a display right. And in between, we will provide an RC circuit. So, there
is a resistance and there is a capacitance in the circuit and this is also how for connected
to this one.

Now; the current which was earlier going directly to the display as V 1, now it is going
as a V 2 to the display and this V 1 has been manipulated to be 2 by using this RC
circuit. The output of this output of current is of this is suppose is I 1, then is it will be
splitting here in I 2 and I 3 and then it will be going to again, it will combined here and
then again it will become I 1 and then it will go to the display.

Now, resistance R is R is the resistance of the resistor is and C is the capacitance of the
capacitor. So, I 1 = I 2 +I 3 right. Now I 2 now for finding out the I or I 2, it is going to
be (V 2- V 1) / R. This is the value of I 2.
Now regarding the value I 1, I 3 I 3 is going through this. So, for I 3 it is going to be C
D again current is flowing. So, it is not V 2 by minus 1, it is V 1- V 2, the current is
flowing in this direction, so, potential difference, so V 1 -V 2. So, C D again, V 1 -V 2;
this is I 3.

Now, if we combine these two, then we will get V 2 and V 2 = I R and; V 2 is going to
be in this case, now V 2 will also have certain resistance R. So, this system will also have
certain resistance R. So, this we can take R o to differentiate between these two R and R
o.

(Refer Slide Time: 04:52)

So, here, again I will write I 1, I 2 is going to be equal to (V 1- V 2) / R o and I 3 is going

to be equal to C D (V 1 -V 2) because resistance of display unit also has to be taken into
account right.

Now here, I is equal to, I 1 is = I 2 + I 3. Now this D is differential of this. For a

capacitor, for a capacitor q = C V. So, I =C dV / dt. From here we have got this. So, it is
derivative of V 1 - V 2.
(Refer Slide Time: 05:41)

So, I 1 = I 2 + I 3 and this is going to be equal to (V 1 -V 2 / R o) + C D (V 1 -V 2). And

this V 1 is equal to this current multiplied by R, V = IR this is not this is not V 1, V 2; V
2.

. So, what we have done here I am just repeating. This is current I 1 and it is splitted to I
2 and I 3. We have calculated the value of I 2 from this, resistance R o and that is (V 1 -
V 2) / R o. We have calculated the value of I 3 as C D( V 1 - V 2).

Once we have the value of I 2 and I 3, we are just added them and we got the value of I 1
and voltage V 2 is nothing but resistance multiplied by I 1 and this is the total resistance
multiplied, this is the resistance and this is the current this is the final equation we have
got.

Now using this equation, we can find the value of V 2 /V 1 using this equation, we can
get the value of V 1 / V 2 by V 1 sorry V 2/ V 1 as α(1 + λo D) / (1 +α λ o D).
(Refer Slide Time: 07:13)

Now, α what is α? Now αhere is, α here is R /(R + Ro), R is the resistance of display unit
and Ro is the resistance added we added through RC circuit and λo = R α. Now, if we
have these two values, now we have these values because R and R o are known to us
then we can find the value of λ and here α is required and λo is required, both values are
with us.

So, and this V 1 αλ o D is equal to V 2 multiplied by this.

(Refer Slide Time: 08:41)

.
Or we can write or we can write (1 + α λ D) V = α Ki q i or qi is let us say this
temperature difference to ϴi, it is temperature difference. So, it is ϴ i. So, this is the
input value right and this is the response of the system λwe choose, the value of because
the λo = Ro C. So, λ we should choose in such a manner that it is equal to the λ o or if it is
equal to Ro C. So, we will put here λ o. So, that we have to choose like that right and then
we will get this response of the system.

Now, for this response of the system, here you can see, the time constant has been
modified to α λo and alpha is less than 1. So, the response of the system has improved,
but on the other side sensitivity of the system has been sacrificed here. So, we have
improved the performance of the system. So, time constant has reduced, but it is the
same time a static sensitivity of the instrument has also reduced.

So, it is always a conservation law which the applicable everywhere; if here giving
somewhere, if you are gaining somewhere, we are gainng in the form of quick response
of the instrument and we are sacrificing the sensitivity of the instrument. Now same
thing can be done for second order system.

Now, second order system, if you want to improve the dynamic performance of second
order system, then damping has to be reduced. The damping is the main issue in the
second order instrument. Even if the system is under damped in this case is suppose the
damping ratio is 0.05 or 0.1, a lot of oscillations are going to be there. So, that has to be
compensated or it has to be improved or the dynamic performance of instrument has to
be improved for a second order system.

Now, for the second order system, in the output form the primary sensing element or
from the thermocouple, tip of the thermocouple, it will be the form of some millivolts
and this output will go to now I will make a simpler circuit; this is R o, this is resistance
we are adding and this is R, resistance of display unit and in series of that we take L and
C.
(Refer Slide Time: 11:31)

So, we make it in the RLC circuit and this is V 1 across this, we are getting V 1 and for
V 2 across this, we are getting V 2. So, in the existing circuit, Ro L and C are added.

Now, due to addition of this, the response of the system will change, the response of the
system will change again. We will calculate the value of V 1. V 1 is there is a current
right.

So, V 1 = [Ro + R + L D + (1 / C D)], L D sorry multiplied by L D and multiplied by i.

So, this is the V1 and for V2 = V1 - i R o or just we can remove this from here, then we
will get the value of V 2.

So, now, we have two values V 1 and V 2. Now we can take the ratio of these V2 / V1. If
we take ratio of these two V 2 / V 1=[( LCD 2 + RCD +1)/ (LCD2 + (Ro +R) CD +1)]; if
we take the ratio, the ratio is going to be like this.

Now, we have to further compare this, I will rub this off.

(Refer Slide Time: 14:00)

V 2 /V 1 = (D / ω1)2 + 2 ζ1 (D / ω1)+1. This ζ1 and (D / ω1)2 + 2 ζ1 (D / ω1)+1; only

difference is this damping ratio.

Now, if you compare these two equations, we will get ω 1 as √(1 / LC) from here. And if
we take the value of ζ1 = 2 by sorry 2 by not is 2 it is R, it is R /[2√(L / C)], R /[2√(1 /
LC)]. That is the value of the ζ1 and for ζ2, for ζ2 = {(R+ Ro) / [2√(L / C)]}, √(L / C) right
and now, if we write the response of the system.

(Refer Slide Time: 15:54)

.
That is initial response of system (D / ωn)2 + 2 ζ1 (D / ωn)+1 or D square by or [D2 / ωn2 2
ζ1 (D / ωn)+1] qo that is V 1 = K, some ϴi. This is the initial response of the second order
system.

Now, here the ω1 = ωn. If we take ω1 = ωn , that is the one boundary condition and this is
ζ1 = ζ right and then we can replace this equation for second order by this equation.

(Refer Slide Time: 17:01)

And here we can see that the effective damping, the ζ1 = ζ, ω1 = ωn and this equation can
further be modified as this one and here the effective damping has increased from ζ2 o ζ1
to ζ2 ight.

So, here in this case, what does happen just by putting a LRC circuit, we have increased
the damping of the system and these techniques this technique has been used or is being
used in many of the electrical appliances. We are simply just by putting RC circuit LRC
circuit, this we are using in measurement also; just by putting this RC circuit and RC
circuit by putting RC circuit in the first order system, we have improved the response of
the system because in the first order system, the main drawback or the main restriction in
the first law first order system is the response of the system.

So, in order to make the response quicker, we have reduced the time constant. Now here
in this case in the second order system, we have improved the or increase the damping
ratio thus the system was under damped and the zeta value was very low. So, in this case,
we have just introduced the RLC circuit in the in the in the output and this RLC circuit
has increase or increase the damping of the system. So, damping ratio has increased right
and system has become more stable as far as the dynamic response is concerned.

Now, I will take one example of thermocouple because it is of interest for the mechanical
instruments. A thermocouple with a time constant of 0.3 seconds and static sensitivity of
0.04 millivolts per degree centigrade, so it has time constant a thermocouple has time
constant of 0.3 seconds and static sensitivity and let us say K 1 is equal to 0.04 millivolts
per degree centigrade. This is for thermocouple.

It is suddenly immersed in bath of hot oil which is 100 degree centigrade. So, t 2 is 100
degree centigrade. The initial temperature of thermocouple measuring and reference
junction was 25 degree.

(Refer Slide Time: 19:37)

So, reference temperature is 25 degree centigrade. So, ϴ = 100 - 25 is equal to 75 degree

centigrade right.

So, there is an input of 75 degree centigrade, what is the output at t is equal to 0.1, 0.3
and 1.
(Refer Slide Time: 19:51)

Then it is being asked, design the composition circuit for thermocouples. So, that a time
constant is reduced to half it is value right.

So, we will start with the response of thermocouple. So, response of the thermocouple V
= (1- e -t / λ) ×K ϴi, So, V is equal to k is 0.04, ϴ is 75, 1- e -t / 0.3.

-t / 0.3
We further simplify this, this is going to be 150 300, 3, 1- e . So, this is going to be
response of thermocouple. So, just we will keep on putting the value of t and we will be
getting different values of output.
We will note it down somewhere. Let us noted down here when t and output t is 0.1,
output is 0.85, t is 0.3 seconds, output is 1.90 millivolts, this is millivolts and t is equal to
1.0 seconds, then output is 2.8 time millivolts.

So, these values we have calculated just simply putting the value of t here. When we are
putting t is equal to 0.1, then it is shift when t is equal to 0.1, then it is 0.85; it is correct.
So, similarly we can find when the values of V by putting different values of t right.

And then compensation circuit comes into the picture. Now we will compensate this with
this is the first order system. So, RC circuit will be applied. So, if you are compensating
this with RC circuit and time constraint time constraint has to be reduced by 50 percent.

So, it means alpha is equal to 0.5 and λo = to Ro C. This equation it is well read the
derivations, we did it we did this equation and this is equal to λ is original lambda this is
0.3. So, Ro C has to be 0.3 because here λ is 0.3, initial value of λ. So, that is equal to Ro
C.

An α = 0.5. So, α = 0.5 = R / (R + R o). The two equations and in these two equations
there are 3 unknowns, in these 2 equations there are 3 unknowns; I mean one is Ro C.

(Refer Slide Time: 23:47)

And third one is R, all 3 are unknown because we know only time constant of
thermocouple and this K 1 and this is a step input which is given and we want to improve
the response of thermocouple by 50 percent or time constant has to reduce by 50 percent
alpha, we have taken 0.5.

Now, we have three unknown, two equations. So, only when therefore, this is trial and
error method. We will keep on assuming one value and then we will get two
simultaneous equations for two values and then trial and error method, we will do that
and suppose R is equal to I will I will just coming to the final value 1 mega ohm, then R o
is going to be how much 1 mega ohm, 0.5.

When Ro is 1 mega ohm, then C is going to be how much; 0.3 farad micro farad. This is
1, this is 1, 1 / (1 + 1), 0.5. Ro is 1 mega ohm and this is 0.3; so, so C = 0.3 × 10 - 6; so, c
is 0.3 micro farad.

So, I am repeating here. First of all we have assume the certain values of R because there
are two simultaneous equations and three unknowns. We have assumed the value of R is
1 mega ohm Ro is also 1 mega ohm because the ratio of R / (R + R o) = 0.5 V and C, the
value of C will get here; it is 0.3 micro farad.

So, we have all 3 values. Now for all these 3 values again, we will calculate the value of
voltage with the revised values, revised value of λ and this is α. So, λ will be L
multiplied by α and static sensitivity will also be multiplied by α. So, static sensitivity
will be reduced. So, here it will be 0.02 right.

But the same time the lambda will be changed to 0.15 and the response of the system
will be q = α K 1. So, α, so first of all I will write the governing equation then final
equation.
(Refer Slide Time: 26:43)

So, q o is V, V = α K q i (1 -e - α λo t) = 0.5 and K is 0.04. So, 0.02 multiplied × ϴ it is 75

- α λo
(1 - e ), then it is 0.15 t or you can see here that the final version of the voltage is
unusually it was see, so 1.5 1 - e - 0.15 t.

Now, here again, we will be putting the different values of t when we are V /, let us say
this is V /. So, we will be getting certain values of V / for the same values of t and once
you calculate these values, it is going to be 0.729 this is for V 1 0 point sorry 1.297 and
1.498.

So, output is affected, but at the same time system, this can be further calibrated, but the
same time response of the system has improved response of the system has improved
because the time constraint has been changed from 0.3 to 0.15.

So, we have found this in this numerical output was required. So, output we have
calculated for compensated circuit and; obviously, the output for compensated circuit is
significantly lower than the output of uncompensated circuit because the sensitivity of
the instrument is compromised in order to improve the a dynamic performance of the
system. That is all for today.

Thank you very much.