Understanding BPSK Spectrum: Differences from DSBSC Signal Generation

[Firefox123]

I have heard that the specturm for a BPSK signal looks like similar to the spectrum for DSBSC, but when I generate a BPSK signal and look at it on the spec an it does not look like DSBSC.

I tried generating it with a Sig Gen using an internal square wave for the modulating signal and I also tried using an external square wave and then phase modulating that. I am using a phase deviation of 180 degrees, but the carrier does not appear to be very suppressed...


Thoughts?

 

[chroot]

BPSK is a digital modulation scheme, where the frequency is constant and only the phase is keyed (of course, this leads to some spectral energy spread around the "carrier").

It should look nothing at all like DSBSC, which is an analog modulation scheme that purposefully intends to supress the carrier.

Why did you expect them to be similar?

- Warren

 

[Firefox123]

BPSK is a digital modulation scheme, where the frequency is constant and only the phase is keyed (of course, this leads to some spectral energy spread around the "carrier").

Right...I am familiar with what BPSK is...

It should look nothing at all like DSBSC, which is an analog modulation scheme that purposefully intends to supress the carrier.

I am also familiar with DSBSC...and I do agree that I would not necessarily expect them to look similar...

Why did you expect them to be similar?

- Warren

Unfortunately I can't go into much detail. Someone told me they should look similar and the results of the application do suggest a DSBSC type spectrum. Perhaps I heard him incorrectly and he wasn't using BPSK. Are there any other digital modulation schemes that would give a DSBSC type spectrum?

 

[Firefox123]

BPSK is a digital modulation scheme, where the frequency is constant and only the phase is keyed (of course, this leads to some spectral energy spread around the "carrier").

It should look nothing at all like DSBSC, which is an analog modulation scheme that purposefully intends to supress the carrier.

Why did you expect them to be similar?

- Warren

According to this website:

http://www.ece.eps.hw.ac.uk/Modules/B33cl2/Cl-9-03r.pdf

a BPSK signal does have a suppressed carrier. This is not the first time I have read this on a website or in a textbook. To quote the website...

If the binary signal represented by the 0,1 square wave is converted into a ±1 or polar waveform, the modulated signal will show phase reversals as the signal changes level, as shown in Fig 9.3. The signal now has zero dc level and the carrier will not be present in the spectrum of the modulated signal, the spectrum will be that of a suppressed-carrier DSB signal.

This is usually called binary phase-shift keying2 or "binary PSK", and is one of the most efficient forms of data modulation.

I have read this several times now, yet when I generated a BPSK signal it did not appear to have a suppressed carrier.

I not sure where the disconnect is between what I am reading and what I saw in the lab. Anyone have any ideas?

 

[mdelisio]

According to this website:

http://www.ece.eps.hw.ac.uk/Modules/B33cl2/Cl-9-03r.pdf

a BPSK signal does have a suppressed carrier. This is not the first time I have read this on a website or in a textbook. To quote the website...

I have read this several times now, yet when I generated a BPSK signal it did not appear to have a suppressed carrier.

I not sure where the disconnect is between what I am reading and what I saw in the lab. Anyone have any ideas?

Don't confuse "suppressed carrier" with a spectrum that has no power spectral density at your carrier frequency.

Technically speaking, a non suppressed carrier will have a "delta function" at the carrier frequency. Practically, what this will mean in the lab is that your power reading on your spectrum analyzer will stay the same regardless of your resolution bandwidth setting. This will be true no matter how tightly you zoom in on the spectrum.

A BPSK waveform with a random bitstream will have a power spectral density that looks a lot like a sinc function, centered on the carrier frequency. Note that this is a suppressed carrier, because there is no "delta function" at the carrier frequency. That is, as the resolution bandwidth decreases, the absolute power will decrease proportionally.

In your pdf, the author is putting in a sequence of 1, 0, 1, 0, 1, 0, ... which is hardly a "random" bitstream. When you crank through the math here, you get nothing at the carrier frequency.

In this sense, I would agree that BPSK is similar to DSB-SC. On the other hand, on-off keying (OOK), would not be a suppressed carrier technique.

As far as the specifics of your question, you will want to make sure that when you feed your modulator with a "1", you get one phase deviation, and when you feed it with a "0" you get 180 degrees different. Sometimes the modulator will want the input to be bipolar (1 = 5V, 0 = -5V, for example) and sometimes the modulator will want the input to be unipolar (1 = 5V, 0 = 0V, say)

Hope this helps.

PS A good text to look at for an introduction is Lathi or Haykin.

 

[Firefox123]

Don't confuse "suppressed carrier" with a spectrum that has no power spectral density at your carrier frequency.

Agreed...

Technically speaking, a non suppressed carrier will have a "delta function" at the carrier frequency. Practically, what this will mean in the lab is that your power reading on your spectrum analyzer will stay the same regardless of your resolution bandwidth setting. This will be true no matter how tightly you zoom in on the spectrum.

Your comment about the res bandwidth is making me take a second look at suppressed carrier and unless I'm messing up the math...I believe you are correct here...

A BPSK waveform with a random bitstream will have a power spectral density that looks a lot like a sinc function, centered on the carrier frequency. Note that this is a suppressed carrier, because there is no "delta function" at the carrier frequency. That is, as the resolution bandwidth decreases, the absolute power will decrease proportionally.

Again I think you are correct...

In your pdf, the author is putting in a sequence of 1, 0, 1, 0, 1, 0, ... which is hardly a "random" bitstream. When you crank through the math here, you get nothing at the carrier frequency.

In this sense, I would agree that BPSK is similar to DSB-SC. On the other hand, on-off keying (OOK), would not be a suppressed carrier technique.

Ok...good. This is the type of "BPSK" I was referring to...not a random sequence.

I've worked quite a bit with OOK so I'm pretty familiar with its spectrum, but haven't worked with BPSK as much...

As far as the specifics of your question, you will want to make sure that when you feed your modulator with a "1", you get one phase deviation, and when you feed it with a "0" you get 180 degrees different. Sometimes the modulator will want the input to be bipolar (1 = 5V, 0 = -5V, for example) and sometimes the modulator will want the input to be unipolar (1 = 5V, 0 = 0V, say)

Hmmmm...I might have forgot to take that into account, which I think would explain the spectrum I saw. It definitely was not suppressed carrier but had a somewhat flattened spectrum with plenty of power in the carrier.

Hope this helps.

It does. Thanks.