Computer Numerical Control of Machine Tools and Processes

Professor A Roy Choudhury

Department of Mechanical Engineering
Indian Institute of Technology Kharagpur
Lecture 13
Interpolators Linear
Welcome viewers to 13th Lecture of the online course “Computer control of machine tools
and processes”. So in this particular lecture we will be discussing somewhat about
interpolators and to a large extent linear interpolators. What is the purpose of having
interpolators? First of all, we are having a continuous control system, we are going to cut
profiles and there we might have to cut straight lines, I mean we might have to cut along
straight lines or along a circular arc or more generally, we might have to cut along complex
curves also.

Not only spatial curves, but also three-dimensional that means 3-D curves. First of all, if we
have to cut along a straight line, it apparently seems to be quite an easy task okay, cutting
along the straight line seems to be the easiest possible or the simplest possible movement, so
why do we have to have this interpolator, linear interpolator for this? This is because in case
of interpolation the program is only providing the machine controls with an initial point and a
final point. So in between the points or the locations through which the cutter is going to
move relative to the work piece, these are not provided.

Also, the velocities that are supposed to be developed by the different axis of motion that is
also not known, so this way two specific tasks of the interpolator emerges; one is, the
interpolator has to make the cutter maintain a definite velocity along the profile of the work
piece, so along the path the cutter has to move at the feed rate programmed. Second, in order
to move along the path the cutter has to have a definite ratio existing between the different
axis velocities. If it is linear motion, these axis velocities are constant, therefore the ratio also
remains constant, this is the way in which linear interpolator can be specified, so let us have a
look what we have.
(Refer Slide Time: 03:17)

So linear interpolators can be made by software or hardware and hardware interpolators may
be made by a number of digital differential analyzers which we have already studied in the
previous lecture on logic circuit and decoder. It is also possible to make software counterparts
of this particular interpolator, which is sometimes referred to as reference pulse interpolator.
So interpolation as we just now discussed, it might not always be restricted to linear and
circular interpolation but also may involve parametric curve interpolation. And these curve
interpolators maybe off-line or online.

(Refer Slide Time: 04:08)

So first of all, let us take a look at a numerical problem, this will help us in establishing the
different relations which exists between different parts of an interpolator, what is this
problem? This problem states that find out the basic length unit, the interpolator frequency f
and the number of interpolator pulses sent through the AND gate of the X axis control loop
for the command in the line number 30 for a CNC closed loop Contouring control or
continuous control machine with encoder number of holes 200, leads screw pitch 4
millimeters, number of starts = 1. There is position down counter and there is an up down
counter also in the control loop, there is a digital to analog converter and M is the permanent
magnet DC motor, A the amplifier, T the tachogenerator and E the encoder.

We are already conversant with the sort of control loops, I will just reiterate 1 or 2 very
important points, that is: interpolator pulses are coming in and first of all they decrement the
position down counter, it is the custodian of the position. So if we only allow a specific
number of pulses to be passed through, the down counter keeps track of the balance of the
pulses of the interpolator and the encoder, while the interpolator is filling up the up-down
counter with higher and higher number, the encoder is going on decrementing the up-down
counter content.

So that, when the motor is rotating the lead screw at the required RPM, that means the
programmed RPM, at that time the up-down counter content becomes steady so that in linear
interpolation a steady state speed is reached. This thing we have already discussed, so I do
not want to elaborate on it once more, so let us concentrate upon the problem. We can
formulate the problem solution this way that is we will find out the basic length unit, we will
find out the feed along the X axis, we will find out the rate of pulses output by the encoder
and we will find out the number of pulses output by the encoder for line number 30 and that
way we will solve the problem.

What exactly is the problem? The interpolator pulse rate is required, what is the train rate of
pulses coming along the line of interpolator pulses that line which is drawn on the left-hand
side, these pulses what is the frequency for this particular command? Let us have a look at the
command,

N20 G01 X20 Y40 F100

N30 X50 Y80

so the initial point for the next line would be (20, 40). From the point (20, 40) in line number
30 we are reaching X = 50 and Y = 80, which means X is moving from 20 to 50, 30
millimeters and Y is moving from 40 to 80, which means 40 millimeters, so 30 millimeters
along X and 40 millimeters along Y and the feed is of 100 millimeters, so let’s go to the
solution.

(Refer Slide Time: 07:47)

So we can solve the problem this way, we first understand that the basic length unit must be
equal to leads screw pitch divided by the encoder number of holes, which is 20 microns 0.02
millimeters. Amount of movement along X, we found out just now, it is 30 millimeters,
movement along Y = 40 millimeters, this also we have found out just now. Displacement
ratio is equal to velocity ratio for straight-line movement, what does this mean? It means that
whenever motion is taking along a line, a part of it along X, a part of it along Y, in that case
the velocity along X and velocity along Y will be having the same ratio as the displacement
along X and displacement along Y in case of linear motion.

So as we are having linear interpolation here, so Vx/Vy = Δx/Δy=3/4 and we also know since
the feed velocity is 100 millimeters per minute therefore, the resultant velocity along the
cutter path must be √𝑉𝑥2 + 𝑉𝑦2 = 100. From this, we can easily find out the value of Vx = 60
millimeters per minute, so once Vx has been found out from the basic length unit which has
been calculated, we can find out the rest of the terms required.

For example, when the table is moving at 60 millimeters per minute along the X axis, the
encoder must be sending 60/0.02 pulses per minute, why is this so, because 60 millimeters
per minute to the encoder is not exactly 60, but it is equal to the number of basic length unit it
has to generate in the corresponding time, so let us see 60/0.02 must be equal to 3000
pulses/minute. So in order todevelop a speed of 60 millimeters per minute or rather when the
table develops the 60 millimeters/minute, encoder sends 3000 pulses/minute, so now we
understand that the encoder pulse rate is 3000 pulses/minute.

So at steady state for linear velocity, the encoder pulse rate must be equal to the interpolator
pulses rate, what does this mean? We already established that when we are having linear
velocity, a steady state has been reached and at that time interpolator pulses rate becomes
equal to the encoder pulse rate. Since encoder pulse rate has been found out to be 3000 ppm,
interpolator pulse rate also must be 3000 ppm, and so answer is interpolator pulse rate must
be 3000 pulses per minute. And how many pulses are being passed through at the AND gate,
it must be corresponding to 30 millimeters because 30 millimeters is the movement along X,
divide 30 millimeters by 0.02 and you will get 1500 pulses, so 1500 pulses are being sent
through so this is the answer to that question.

(Refer Slide Time: 11:39)

Now let us see a DDA based interpolator. In the DDA based interpolator as we have
discussed previously there is an input pulse rate f, it enters a feed DDA which is having m
number of bits and inside this m bit counter a number S is stored, so S has to be smaller in
size compared to 2m - 1 that that can be the maximum size of S. The output pulse rate of any
such DDA is always equal to f multiplied by Δ x (which is inside that particular counter)
divided by 2n if n be the number of bits of that DDA. So the pulse rate output of the feed
DDA goes to X DDA and Y DDA and they send out their respective pulse rates to the X
motor and the Y motor, this is the basic structure of the linear interpolator.
 𝑓×10×𝑉×∆𝑥
And output pulse rate is ultimately found out to be , where f = the input pulse
𝐿×2𝑚+𝑛

frequency, V is the programmed feed rate along the cutter path, Δx is the incremental X
motion, L is the length segment of the path. This is the pulse rate out of the X DDA, and S
also has an expression, 10 × V / L, these things we have already defined and f has a value of
2m+n/ 10, so let us take a numerical problem.

(Refer Slide Time: 13:39)

During DDA based linear interpolation, the cutting tool Centre has to move from point (20,
10) to the point (120, 60) at a speed of 3 meters per minute. The coordinates of the points are
in millimeters, the table size is 600 × 600 millimeters, the maximum table speed is 12 meters
per minute for a minimum distance of 50 millimeters and the basic length unit (BLU) is given
to be 10 microns. Find out the bit sizes of all 3 DDAs, the S word in the feed DDA that
means the S word which has to be put in this particular case, the values of Px and Py, the
clock frequency and the output frequency of feed DDA and of course the output frequency of
x and y axis DDAs.
(Refer Slide Time: 14:45)

So first of all let us see the bit sizes of all three DDAs. So we understand that 600 millimeters
is the table size on either X or Y direction, if this be so then if the basic length unit is 10
microns, in that case we can easily find out what will be the maximum size of Δ x that means
maximum possible incremental motion. So if we suppose that the maximum possible
incremental motion is equal to the length of the table itself, which is 600 millimeters
therefore, 600 divided by basic length unit is maximum Δx or incremental x motion.

So if 600 millimeters is our motion, then 600 ×100 is the maximum number which has to be
accommodated in the X feed DDA X DDA okay, so let me read it. 60,000 being equal to 600
divided by 10 microns is the maximum number which has to be accommodated in the X
DDAcounter. Now X DDA counter is of n bits, so the maximum number it can accommodate
is 2n-1, so simply by calculation we can find out if n = 16, then 2n- 1 is 65536 and this is the
smallest counter which can accommodate 60,000, so we come to the conclusion that the size
of the counter inside X and Y DDA must be 16 each, 16 bit counter inside X and Y DDA.

In the same way, we understand that the feed DDA which has S word inside which is of m bit
size, it can contain or rather it has to contain the maximum possible feed rate number or S
word. In that case, the maximum possible number is given by the condition given in the
problem, the condition is the maximum table feed or rather table speed is 12 meters per
minute for a minimum distance of 50 millimeters, so we understand that if we take a table
speed of 12 divided by the corresponding length of 50 millimeters, we will be able to
compute the maximum value of the S word.
So the highest velocity V is 12 meters per minute and lowest distance L is 50 millimeters
therefore, maximum S word is 10×V/L, so this thing is expressed in length units per minute,
so this is equal to 2400. So in order to accommodate 2400, which is the maximum possible S
word, we have to have m = 12, so 212- 1 will be able to accommodate 2400. Having
understood that, we understand that m = 12 and n = 16 so that the input clock frequency
which is equal to 2m+n/10 pulses per minute, we can easily compute it to be 228/10 pulses per
minute = 447392 pulses per second. So f has been computed as we have already calculated
the values of m and n.

(Refer Slide Time: 19:34)

Coming to the 2nd part of the problem, for the particular case that has been given, incremental
x and incremental y in basic length units, they are 10,000 and 5000 respectively, how did we
find this out? We found out the initial point and the final point, the differences Δx and Δy ,
divided it by the basic length unit and obtained these two values. Basic length unit
incidentally just to remind you, was 10 microns, so now the length of cut must be the square
root of the sum of squares of these two and therefore, it is equal to 11180 in basic length unit
okay. So why are we finding out the length of cut because 10×V/L has to be calculated for
this particular case.

So output frequency of feed DDA for this movement, so what is this feed DDA? That is the
first DDA which was existing, let us have a quick look this is the feed DDA, we will find out
the output frequency of this one then multiply this by the corresponding Δx/2n, Δy/2n value
and that way we will get these two frequencies. So output frequency of the feed DDA for this
movement is f ×S/2m since we know all the values now, we put 447392 as the value of f, we
put S=10×V/L = 10×300,000/11180 that means velocity is 3 meters per minute given for a
particular case that is divided by the basic length unit.

So V = basic length unit per minute divided by length expressed in the same length units that
means basic length units; this further divided by 2m, which is 212 and that way we obtain
29309.43 pps. We are having this in pps because f is in pps (pulses per second). So having
obtained the output frequency of feed DDA, we can easily obtain the frequencies of x and y
DDAs. How is this so: because feed DDA multiplied again by the same form that is f×Δx/2n ,
so we have feed DDA frequency multiplied by Δx divided by 2n where n = 16 will give us
4472 pulses per second, this is Vx.

Once again, feed DDA output frequency multiplied by the Δy/216 gives us 2236 pulses per
second, this is Vy. Now the vector sum of Vx and Vy is to be 3 meters per minute, so if Vx =
4472 basic length unit per second, it is correspondingly equal to 44.72 millimeters per second
= 2683 millimeters per minute. In a similar manner, if we compute Vy, it will come out to be
1341.5 millimeters per minute and therefore, if we combine these two vectorially, their result
will come out to be 3 meters per minute.

So therefore we have found out first of all : m, we have found out n, n is 16, m is 12, f has
been found out as 447392 pulses per second, the output frequency of the feed DDA has been
found out to be 29309.43 pulses per second and Vx and Vy frequencies have also been found
out, okay. Let us take one multiple-choice question.
(Refer Slide Time: 23:58)

In this multiple-choice choice question what is given, you are joining DRDO which we
already know to be Defense Research Development Organisation and they ask you to design
a grenade with programmable delay before detonation. Grenades, when you take out some
pin from that, after that a definite time is permitted beyond which an explosion or detonation
occurs. So here we are talking about a grenade where the detonation time or the detonation
delay is programmable. Inside the grenade, there is to be a chip with a fixed clock at 1024
pulses per second, which is input to a DDA.

So all these things are inside the grenade because DDAs and clock - small clocks, et cetera
they are miniaturized and they can easily fit into grenade, so which is input to a DDA and the
first overflow of the DDA will cause detonation, so it is simply like this. I have a clock, I
have a DDA, I am putting in some number and the first overflow pulse which comes out, first
output pulse it is going to cause detonation. The grenade will have buttons to load the counter
of DDA with 3 numbers for 3 detonation delay times; 2 seconds, 4 seconds and 8 seconds,
what should be the minimum DDA size to make this work?

So we know f, we do not know the number of bits inside the counter of the DDA and we
know the required pulse rate output at the detonation side. So the choices of answers are :
(DDA size we know it is 2n- 1, so that way we have given it in that format

210- 1

213- 1
218- 1

and none of the others of course,

so let us try this out.

(Refer Slide Time: 26:27)

So the rate at which the detonations are supposed to take place that is half per second because
if 2 s is the time delay for detonation, that means the first detonation is going to take place
after 2 seconds have passed and therefore, if one detonation takes place in 2 seconds
therefore, per second I am having half a detonation this is the rate of output pulses required
for this particular case. In the same way, if 8 seconds are supposed to pass before the
detonation therefore, 1/8 detonation per second is taking place. So we have the frequency of
detonation to be ½ , ¼ and 1/8 per second.

So let us assume the number of bits in DDA to be x, so what does this mean ? that the counter
size in the DDA is equal to x bits, so the maximum number that it can contain is equal to 2x-
1. So now let us look at the lowest output pulse rate that can be obtained from this DDA,
what do we mean by lowest output pulse rate? 1024 is fixed, the frequency then hardware
size we are already assuming to be of x bit, so 2x denominator is fixed and therefore, the x1
which gets multiplied as the number contained in the counter, if it is the lowest possible value
x1 inside the counter, it can only take up integer values that means its minimum value can be
1, if it is 0 then there is no point having that DDA.
So minimum value of x1 is 1 and then you can take other positive integer values. So, if we
are planning to get the slowest detonation rate at 1/8 per second, let us equate 1024×1/2x that
means I will put a 1 bit there inside, so let us work out and see how much that comes out to
be.

(Refer Slide Time: 29:04)

We have here 1024 is the frequency multiplied by only 1 here divided by 2x equal to the
frequency that we want. If I put the smallest number, I should get the lowest frequency. So
that way, 2x = 1024 × 23, and therefore, x becomes equal to 13. So if you have the number of
bits inside this counter to be 13, then we can definitely ensure that this is going to have
detonation times very easily by putting 1024 multiplied by 2/2x, this is going to give us
double of this, therefore we are going to get ¼ , so on and so forth, if you put 4we are going
to have ½ , so this way we can make this particular grenade work okay, thank you very much
thank you.