EMRFD Message Archive 11937

Message Date From Subject
11937 2015-12-04 15:58:12 kerrypwr Fourier Transform Machine
I found this fascinating; only a genius could design a mechanical spectrum analyser/waveform generator a century ago;

Kerry VK2TIL.


11938 2015-12-04 19:21:34 Ashhar Farhan Re: Fourier Transform Machine
This was made by Albert Michelson, the great scientist/inventor who also tried to measure the speed of light. Along with Edward Morley, they figured that ether, the proposed medium for electromagnetic waves didn't exist. It was this that grabbed Einstein's attention and the rest was a phenomenal turning around in our understanding of everything.

The spectrum analysis through fourier analysis is actually an approximation of the real thing. quantization error of amplitude and sampling errors of time make it a useful but not the real thing. The reality is analog. We tend to forget this more often than not. Digital is all in the Mind. The matter is Analog. 

No matter? Never mind.

- f

11953 2015-12-07 02:44:11 n2cqr Re: Fourier Transform Machine
I may be missing something here, but it looked to me as if they were only adding (summing) two sines.   They were not really mixing (multiplying), correct?  So this would be a spectrum analyzer only for complex waveforms that result from the linear addition of sines or cosines?  
11954 2015-12-07 03:19:03 Thomas S. Knutsen Re: Fourier Transform Machine
Yes, they are adding sines. 20 of them in this particular model, but more are possible.
The whole field of Fourier theory, and analysis is that all signals are combinations of sines. As such, all complex waveforms are possible. 
This particular machine have some limits, like most digital fourier analyzers do, that it only allows integer frequency bins. This is unlike the Fourier transform integral (not FFT or DFT) that allows non-integer bins, that gives quite interesting results when used on real signals. 

Its a interesting, and quite visible way to do the synthesis and analysis of the frequency components in a given signal. Wonder when someone comes up with a kickstarter for a table model of it...

73 de Thomas.


2015-12-07 11:44 GMT+01:00 n2cqr@yahoo.com [emrfd] <emrfd@yahoogroups.com>:
 

I may be missing something here, but it looked to me as if they were only adding (summing) two sines.   They were not really mixing (multiplying), correct?  So this would be a spectrum analyzer only for complex waveforms that result from the linear addition of sines or cosines?  

11955 2015-12-07 08:14:14 Nick Tsakonas Re: Fourier Transform Machine
on a recent visit to the Science Museum in London I saw another machine called "Harvey harmonic analyser".

see this link from the museum (it take some time to load)

it is interesting but I never managed to find more info how it works and what it does exactly and the layout does not help to visualize its operation.

73,
Nick,sv1djg





11957 2015-12-07 09:34:00 atlantaswl Re: Fourier Transform Machine
Multiplied trig functions (what mixers are doing) can also be expressed as the sum or difference of trig functions.  The result is equivalent.

73, Fred


11964 2015-12-07 22:46:33 John Re: Fourier Transform Machine
The patent, with explanation, is at
http://www.directorypatent.com/GB/335969-a.html
click on Download PDF

John
KK6IL

11965 2015-12-08 03:12:15 Nick Tsakonas Re: Fourier Transform Machine
Thanks a lot John, I am going to study it.
Nick
SV1DJG

>The patent, with explanation, is at
>http://www.directorypatent.com/GB/335969-a.html
>click on Download PDF
>
>John
>KK6IL




11969 2015-12-08 10:01:15 Andy Re: Fourier Transform Machine
"I may be missing something here, but it looked to me as if they were only adding (summing) two sines.   They were not really mixing (multiplying), correct?"

Fourier analysis does not involve multiplying (mixing) signals.  It is a strictly linear analysis or linear summation.

Andy