Transmitter Modulators

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: Transmitter Modulators

Post by MrAl »

Hi Robert,


Would be nice to hear from someone with some hands on experience in this field. I'm working totally from theory that is more basic than the theory of modulating like this and have no idea how the actual implementation is done.
LEDs vs Bulbs, LEDs are winning.
Robert Reed
Posts: 2277
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Re: Transmitter Modulators

Post by Robert Reed »

Well heres a little tidbit to rub salt into our wounds. The ITU (International Telecommunications Union) has just okayed a push for an even more complex transmission. This falls under 4G technology and would send one gigabyte of data rate. Obviously the required bandwidth would be at least that and probably more. Now heres where it gets interesting - There is no contiguous spectrum left that can handle that bandwidth, so the proposal will be to fragment it over whatever pieces and bits of available spectra a tower operator can gather to total the the necessary bandwidth needed.
So, for our next discussion - just kidding :grin: - as this one has my brain in knots already and by comparison would be merely modulation 101.

PS even though we didn't accomplish too much, we certainly added a lot of activity to this apparently dying forum.
User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: Transmitter Modulators

Post by MrAl »

Hi Robert,

Actually that sounds a little simpler than this one :smile:
LEDs vs Bulbs, LEDs are winning.
User avatar
CeaSaR
Posts: 1949
Joined: Sat Nov 08, 2003 1:01 am
Location: Phoenixville, PA USA
Contact:

Re: Transmitter Modulators

Post by CeaSaR »

Well, I have just sent an e-mail to Lou Frenzel. I am hoping he joins in, at least
enough to set our spinning heads straight. :lol:

CeaSaR
Hey, what do I know?
Robert Reed
Posts: 2277
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Re: Transmitter Modulators

Post by Robert Reed »

OK Ceasar
I subscribe to ELECTRONIC DESIGN Magazine in which he has regular columns. I will keep my eyes peeled on that end too.
LouisFrenzel
Posts: 1
Joined: Tue Aug 09, 2011 2:49 pm
Contact:

Re: Transmitter Modulators

Post by LouisFrenzel »

Interesting discussion. I assume that the main question here is “what does the spectrum of a BPSK signal look like?” BPSK is phase modulation. Phase and frequency modulation are forms of angle modulation. Both produce an infinite number of sidebands on both sides of the carrier. That’s the bad news. The good news is that most of the power is usually concentrated in a narrower band around the carrier. And the total bandwidth needed for it depends on the nature of the modulating signal.
The modulating signal is usually a binary rectangular wave meaning it contains lots of harmonics. Worst case is a binary signal with alternating 0s and 1s that produce a square wave. This means a fundamental frequency (f) of the binary signal 1/2T where T is the bit time duration. A square wave produces odd harmonics so you get modulating frequencies of 3f, 5f, 7f, etc. Each harmonic will produce a sideband around the carrier.
The harmonics are produced by the steep rise and fall times of the binary signal. To compensate for so many large harmonics you filter the binary signal to slow those rise and fall times in a raised cosine low pass filter. This greatly reduces the number of harmonics and their strength and, in turn, that reduces the signal bandwidth significantly.
As for the spectrum, it is complex because of all the harmonics. The non-filtered spectrum is a major lobe of energy around the carrier frequency (fc) with nulls on either side at ±1/T. Then there are two more smaller lobes with two more nulls at ±2/T. The lobes and nulls repeat with that pattern with the lobes getting smaller in amplitude at the higher frequencies.
When you filter the binary signal first to reduce the harmonics, you get one big fat lobe centered on the carrier with nulls at ±1/T as before. The other lobes are greatly reduced putting most of the signal power in the main lobe. The bandwidth of the main lobe is 2/T. That means you can just about send a 1 Mb/s square wave through 1 MHz wide channel with BPSK. Not bad. We say the spectral efficiency is about 1 bit per Hz.

That is my view. Hope it helps.

Lou Frenzel
User avatar
CeaSaR
Posts: 1949
Joined: Sat Nov 08, 2003 1:01 am
Location: Phoenixville, PA USA
Contact:

Re: Transmitter Modulators

Post by CeaSaR »

Thank you for joining in, Mr. Frenzel.

From your description, it seems that most of what we, "the collective participants",
have found "out there" in regards to any type of spectrum plot has been found to
be severely lacking. Most have been either broad frequency vs. time or plain old plots
of a generic waveform, again, in the time domain.

Also, from what you describe, it sounds almost like all the filtering brings it very close
to an ideal transmission mode where (almost) all of the energy is put forth in the main
carrier, with greatly reduced sidebands, reducing quickly to insignificant levels.

I know that your post will be better understood by the 2 main posters, Robert Reed (OP)
and MrAl, and also Bob Scott.

Again, thank you for contributing to the "coffee clutch". Hope to see more from you, here
in the forum, in the future.

CeaSaR
Hey, what do I know?
Robert Reed
Posts: 2277
Joined: Wed Nov 24, 2004 1:01 am
Location: ASHTABULA,OHIO
Contact:

Re: Transmitter Modulators

Post by Robert Reed »

Hello Lou Frenzel
Thanks for picking up on our discussion. I have read and enjoyed many of your articles in the past by the way. From what you describe pretty much agrees with my understanding and experience with angle modulation. But i have a few questions that I hope you have the time to answer. If Bessel functions in regards to the modulation index apply here, as deviation (or phase angle) are increased , carrier power starts to migrate into the sidebands and at a modulation index of 2.405 the carrier actually disappears (Bessel Null) and all the transmitted power is in the sidebands. My point being that it would seem that these relative side band/carrier amplitudes would be changing to some degree with various data rates -true?
Also in the article, it mentioned the need for a Linear final amplifier which indicates the need to preserve some sort of AM modulation fidelity.This was confusing to me as all of the angle modulated transmitters I have worked on deliberately clamp the out going carrier amplitude and also limit (post IF limiters) the captured signal in the receiver.
I know that any explanations here are definitely NOT one liners and this can be a very profound subject, so I won't feel slighted in the least if you prefer not to go into great detail on this.Thanks again for chiming in :grin: .
User avatar
MrAl
Posts: 3862
Joined: Fri Jan 11, 2002 1:01 am
Location: NewJersey
Contact:

Re: Transmitter Modulators

Post by MrAl »

Hello again,


There is definitely even harmonic as well as odd harmonic content as it shows up quite clearly doing a Fourier.

But also interesting is that it looks like (still doing some spice stuff) when the binary code repeats (such as a double 0 or double 1 being sent) the average level of the transmitted signal (assuming much of the harmonic content has been filtered out with a band pass filter) goes up for the second and remaining bits, until a bit that is opposite is sent and then the level falls but for only one cycle. This is interesting because we would be able to look at the signal and assuming we were sync'd we could tell what bits were coming at us.
For example, sending four bits 1111 (after a previous 1 was also sent) would result in four cycles with peaks of say 2. But after that say we send another 1, then we would get another peak at 2, but if we instead sent a zero after that stream of 4 ones we'd see a peak at only 1 (roughly). So if we see peaks like this:
22221
that would mean four 1's and a zero was sent. Peaks like this:
222211222 (peaks)
would mean we sent four 1's then a zero, then another 1 and then three more 1's like this:
111101111 (binary)
But if we got this instead:
222212222
that would mean four 1's followed by a zero followed by four more zeros.
So the rule seems to be every time we see a peak of 2 that means the previous bit is being repeated, and every time we see a level of 1 that means the bit is opposite to the previous bit.
I dont know what significance this has but i thought it was interesting.

Of course the phase at the end of the cycle period changes too, so if we send a 1 and we see 0 phase shift at the end of the cycle and then at the end of the next cycle we see 180 degree phase shift that means a 1 followed by a zero was sent.
LEDs vs Bulbs, LEDs are winning.
Post Reply

Who is online

Users browsing this forum: No registered users and 35 guests