Find the transfer function for this signal flow graph

AI Thread Summary
The discussion centers on finding the transfer function for a signal flow graph using Mason's method. Participants express uncertainty about the method's application and suggest using software like MATLAB or freeMASON for quicker solutions. Key points include the need to properly label edges and identify forward paths and loops for clarity. Adjustments to connections, such as moving W4 and W5, are discussed to simplify the graph. Overall, the conversation highlights the complexity of the problem and the potential benefits of software assistance in solving it.
dat
Messages
6
Reaction score
1

Homework Statement


da.jpg


Homework Equations


Manson method

The Attempt at a Solution


I used Manson method. I am not sure this is the best result. can someone use MATLAB or some software to solve this quickly ??
tyt.png
 
  • Like
Likes berkeman
Physics news on Phys.org
I think the man is called Mason, and sorry: I think you are your method is wrong.

Now, if you number the points from x to y: 1 . . 6, the transfer function from 1 to 2 seems to be 1. But later you will have a problem with the W4 from 1 to 3. You must move this W4 connection from 1-3 to 1-2. You can do this if you substitute the W4 by W4/W1, so that the new 1-2 function becomes 1 + W4/W1.

In the same way you must move the W5 function from 3-5 to 4-5, so that the new 4-5 function becomes
1 + W5/W2. So now the 2-3 and 3-4 loops in series have been made nice and clean.

You will have to do some more movements, to unfasten the loops, but I hope you understand the "system". Having done so, the rest of the problem is Mason's.
 
Last edited:
  • Like
Likes Greg Bernhardt and dat
My general comment is that the first thing to do is to label the edges so that all forward paths and loops can be clearly identified. Then do the substitutions in each clearly identified path. As far as I can see, the forward path gains and the loop gains make sense. But it is very difficult to identify the paths being calculated and I got tired of looking at it.
I am not aware of any software package that helps these calculations except one called freeMASON by Nikolai Schwerg https://www.radioeng.cz/fulltexts/2012/12_03_0892_0897.pdf The formula manipulations are equivalent to some standard matrix calculations, so there may be some software like Mathematica or Maple that would help.
 
  • Like
Likes dat
Hesch said:
I think the man is called Mason, and sorry: I think you are your method is wrong.

Now, if you number the points from x to y: 1 . . 6, the transfer function from 1 to 2 seems to be 1. But later you will have a problem with the W4 from 1 to 3. You must move this W4 connection from 1-3 to 1-2. You can do this if you substitute the W4 by W4/W1, so that the new 1-2 function becomes 1 + W4/W1.

In the same way you must move the W5 function from 3-5 to 4-5, so that the new 4-5 function becomes
1 + W5/W2. So now the 2-3 and 3-4 loops in series have been made nice and clean.

You will have to do some more movements, to unfasten the loops, but I hope you understand the "system". Having done so, the rest of the problem is Mason's.
I understand what you mean, your method is the first method I learned. I can solve this problem by only doing your method, not using Mason method. But Mason method I had used is the faster way to solve and it assume that you don't have to simplify the graph.
 
Hesch said:
I think the man is called Mason, and sorry: I think you are your method is wrong.

Now, if you number the points from x to y: 1 . . 6, the transfer function from 1 to 2 seems to be 1. But later you will have a problem with the W4 from 1 to 3. You must move this W4 connection from 1-3 to 1-2. You can do this if you substitute the W4 by W4/W1, so that the new 1-2 function becomes 1 + W4/W1.

In the same way you must move the W5 function from 3-5 to 4-5, so that the new 4-5 function becomes
1 + W5/W2. So now the 2-3 and 3-4 loops in series have been made nice and clean.

You will have to do some more movements, to unfasten the loops, but I hope you understand the "system". Having done so, the rest of the problem is Mason's.
oopss
 
FactChecker said:
My general comment is that the first thing to do is to label the edges so that all forward paths and loops can be clearly identified. Then do the substitutions in each clearly identified path. As far as I can see, the forward path gains and the loop gains make sense. But it is very difficult to identify the paths being calculated and I got tired of looking at it.
I am not aware of any software package that helps these calculations except one called freeMASON by Nikolai Schwerg https://www.radioeng.cz/fulltexts/2012/12_03_0892_0897.pdf The formula manipulations are equivalent to some standard matrix calculations, so there may be some software like Mathematica or Maple that would help.
can you just put the graph in the software and find the tranfer funtion to get the result ?? I only want to know the right answer
 
dat said:
can you just put the graph in the software and find the tranfer funtion to get the result ?? I only want to know the right answer
Sorry, I have never used that software and don't have it.
 
  • Like
Likes dat
FactChecker said:
Sorry, I have never used that software and don't have it.
thank you for your concern
 

Similar threads

Back
Top