> > One night in 1951, I had worked late on the CCNY ham transmitter, W2HJ, > when I found that I was locked into the building. I had the modulator > apart and I didn't know Morse well enough to trust keying the final, so > I soldered tin can bottoms to adjacent turns of the output tank. The > final was a an Eimac 304TH with a KW input, so getting too close would > have zapped me good. I arranges a piece of grounded screen wire between > me and the improvised capacitor plates and shouted, The resulting phase > modulation was enough to get through to a kind soul who phoned the > police who, in turn, got Buildings and Grounds to let me out. I had no > ticket, but nobody blinked when I used the station's call letters. > > A slope detector is frequency sensitive, so a phase modulator comes > through with a 6 dB/octave slope, but it's still intelligible. > > Jerry > -- > Engineering is the art of making what you want from things you can get.- Hide quoted text - > > - Show quoted text -Sounds like a good Carl and Jerry Story :-) Mark
Generating phase modulation.
Started by ●December 7, 2009
Reply by ●December 7, 20092009-12-07
Reply by ●December 7, 20092009-12-07
Seems to me the most troubling thing here is the desired result is to perform an operation (modulation) on 100 sinusoids and then sum them together; but the proposed signal flow chart -- as I understand it -- is to sum them together first, then formulate the now much-more-complex operation of modulating them as an ensemble. Yes, you can use something like an allpass filter to change their phase. But how do you modulate this filter, that is, how do you make it time-varying in the desired way? I am saying there must be some good system reason for forming this pre-summing of sinusoids before modulation, otherwise you would not be doing the processing in this particular order. Steve
Reply by ●December 7, 20092009-12-07
>Am I sensing a upcoming mockery? =) > >Not at all. Puzzlement, perhaps.Thanks for your patience!>I'm not sure what "it" is. You have a bunch of sinusoids whose phase you>want to adjust individually. Neither their frequencies nor their >amplitudes, nor their phases relative to some reference change with >time.Well, as you say, I think what is lacking is a strict definition of what I mean by phase in this context. What I mean is at t=t0, sinusoid i has phase phi0_i. Rather than the sinusoids having phase phi1_i =phi0_i + omega_i*(t1-t0) at time t=t1, I would like them to have phi1_i_alt not equal to phi1_i, in effect, omega_i = omega_i(t).>Why not build a network for each frequency that alters its phase >to your liking? If it affects amplitude, you can restore that with a >multiplicative constant. That's simple enough so that I'm puzzled that >you ask. If it's not that, I'm puzzled about what you want.This means I have to explicitly compute the time evolution of each sinusoid with (smoothly) time-varying frequency. Well, that might not be too bad, how do I choose what frequency altering function to use? I "simple" raised cosine or gaussian would perhaps work.>Note that phase is defined only for frequencies that are exactly equal. >Although we use the term loosely in other contexts, the meaning has to >be explicitly defined for each context. > >Jerry >-- >Engineering is the art of making what you want from things you can get. >�����������������������������������������������������������������������Sorry for the confusion, perhaps that is inevitably in the communication between a beginner and an expert.
Reply by ●December 7, 20092009-12-07
>Seems to me the most troubling thing here is the desired result >is to perform an operation (modulation) on 100 sinusoids >and then sum them together; but the proposed signal flow chart -- >as I understand it -- is to sum them together first, then formulate >the now much-more-complex operation of modulating them as an ensemble. > >Yes, you can use something like an allpass filter to change their >phase. But how do you modulate this filter, that is, how do >you make it time-varying in the desired way?Right, I'm guessing that would be equally troublesome.>I am saying there must be some good system reason for forming >this pre-summing of sinusoids before modulation, otherwise >you would not be doing the processing in this particular order. > >Steve >Maybe I didn't quite get you here but there is no need to sum them together first and then perform the phasemodulation. As of now I'm trying to modulate them first and then sum them together.
Reply by ●December 7, 20092009-12-07
Demus <sodemus@hotmail.com> wrote:> spp wrote,>>the proposed signal flow chart -- >>as I understand it -- is to sum them together first, then formulate >>the now much-more-complex operation of modulating them as an ensemble.>Maybe I didn't quite get you here but there is no need to sum them >together first and then perform the phasemodulation. As of now I'm trying >to modulate them first and then sum them together.Ah. Your phrase was "a signal consisting of ... 100 sinusoids", which I interpreted to mean a single signal with 100 sinusoid components. In this case your problem may be pretty simple, such as implementing a variable delay line on each carrier to change its phase. Steve
Reply by ●December 7, 20092009-12-07
Steve Pope wrote:> Demus <sodemus@hotmail.com> wrote: > >> spp wrote, > >>> the proposed signal flow chart -- >>> as I understand it -- is to sum them together first, then formulate >>> the now much-more-complex operation of modulating them as an ensemble. > >> Maybe I didn't quite get you here but there is no need to sum them >> together first and then perform the phasemodulation. As of now I'm trying >> to modulate them first and then sum them together. > > Ah. Your phrase was "a signal consisting of ... 100 sinusoids", > which I interpreted to mean a single signal with 100 sinusoid components. > > In this case your problem may be pretty simple, such as implementing > a variable delay line on each carrier to change its phase.Why variable? The initial and desired phases are both said to be known. Fixed phase shifters should suffice. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●December 7, 20092009-12-07
Jerry Avins <jya@ieee.org> wrote:>Steve Pope wrote:>> In this case your problem may be pretty simple, such as implementing >> a variable delay line on each carrier to change its phase.>Why variable? The initial and desired phases are both said to be known. >Fixed phase shifters should suffice.Because the word "modulation" implies something is varying. The phrase "phase modulation" implies the phase is varying. Steve
Reply by ●December 7, 20092009-12-07
> >Why variable? The initial and desired phases are both said to be known. >Fixed phase shifters should suffice. > >Jerry >-- >Engineering is the art of making what you want from things you can get. >����������������������������������������������������������������������� >Well, maybe I didn't make that clear but there are TWO conditions on the phase... phi_t0_i = A_i, phi_t1_i = B_i BUT phi_t1_i - phi_t0_i =/ (t1-t0)*omega_i So there is no appropriate constant phaseshift to apply.
Reply by ●December 8, 20092009-12-08
On Dec 8, 9:02�am, "Demus" <sode...@hotmail.com> wrote:> Hello, > > As I understand it, phase/frequency modulation is usually achieved by > generating by directly modifying the argument of a sinusoid. > > However, if the task was to frequency modulate an existing signal which > itself we don't generate, is there a standard way doing so, e.g. by > superposing a modulating signal (probably a piecewise continuous signal) on > the existing signal? > > In my case, phase modulation might be the more proper term. > > Thanks in advance!Phase modulation is just FM but without the integrator on the baseband. Hardy
Reply by ●December 8, 20092009-12-08
Jerry Avins wrote:> I soldered tin can bottoms to adjacent turns of the output tank. The > final was a an Eimac 304TH with a KW input, so getting too close would > have zapped me good. I arranges a piece of grounded screen wire between > me and the improvised capacitor plates and shouted, The resulting phase > modulation was enough to get through to a kind soul who phoned the > police who, in turn, got Buildings and Grounds to let me out. I had no > ticket, but nobody blinked when I used the station's call letters.Ladies and Gentlemen: the original MacGyver! :) Rick Armstrong
