ehild said:
Hi Stevenb,
I really grew up in the stone age using slide rules and math tables, but I did not hear about signal flow graphs - it looks interesting- but I do not understand how to draw such one and how the method works, can you suggest some material about it?
ehild
I'm actually post-stone-age myself, but not by much of a margin. I started studying engineering around 1980, so I was taught by some stone-age mentors, which I feel very fortunate about, by the way.
I was lucky to learn electronics from one of the old analog design masters, just before he retired. He was in his eighties at the time, and wearing a hearing-aid bigger than today's cell phones. I'm pretty sure he chiseled out a stone wheel or two in his time. I still have the notes from his class on analog circuit design, which I took as a senior in college. I was further blessed in grad school to be assigned twice as a TA for the lab associated with his lectures. There is no better way to learn than to try and teach others, and doing it under the piercing gaze of such a master just added to the pressure. So this was quite an education.
Anyway, Prof. H. was a firm believer that the SFG approach is the best for analog circuit design, even after the wider availabiltiy of computers. He felt it provides insight to the inner workings of feedback. Somehow he eventually convinced me of this and I've used this approach throughout my own career. The only thing I've added to the process is checking all formulas derived from Mason's gain rule, using symbolic processors like Mathematica, Maple and Maxima. Hence, I get the confidence of accuracy from the new ways, while keeping the insight of the old ways.
You are right. The SFG approach is one that most people have not heard about, for some reason. Perhaps it's because there is a pretty steep learning curve to get proficient enough to use it to it's full advantages. Hence, it's a hard sell to get people to appreciate it. There are so many ways to solve linear systems that many people think it's just another way, but few give design insight like this technique.
The original paper on this is by S. J. Mason, from which we get Mason's Gain Rule.
S. J. Mason, "Feedback Theory - Some Properties of Signal Flowgraphs", Proc. of the Inst. of Radeo Eng., 41, pp. 1144-1156; September 1953
His follow up paper is available on-line here:
http://dspace.mit.edu/bitstream/handle/1721.1/4778/RLE-TR-303-15342712.pdf
And some on-line documents are here. The second one looks really good.
http://www.ece.tufts.edu/~srout01/ee12-2008/pdfs/lecture14-ogata.pdf
http://www.ives.edu.mx/bibliodigital/Ingenierias/Libros%20Ingenieria/Automatica/the%20engineering%20handbook/Section%2016/Ch96.pdf
If you are interested, you are welcome to PM me your email, and I can send you a PDF of some pages from my electronics notebook (20 MByte file !). I wish I could post it here, but it's just too big.
(EDIT ! Scratch that, I read R1=0.8 Ohms instead of R1=0.5 Ohms)
