Prev: forward error correction capabilities?
Next: OMAP-L1x Processor Universal Serial Bus 2.0
From: Paul Wolf on 16 Jun 2010 05:53
Dear professionals,

I am a newbie to the topic of carrier recovery. I have to build up a
point-to-point wireless communication link consisting of a quadrature
modulator and demodulator. I think to apply -1/+1 PRBS signals on I and Q
inputs of the transmitter. So I will get 4QAM constellation
(45°/135°/-45°/-135°).

1. Is it possible to use a Costas loop to recover the carrier frequency and
mix the received signal with this carrier frequency down? Please provide me
with information about the appropriate Costas loop configuration for my
case.

2. And another question: will the Costas loop give me 4-PSK constellation
like (0°/ 90°/ 180°/ 270°) or the wanted one (45°/135°/-45°/-135°)
?

3. And the last question: which kind of AGC is easy to implement in analog
way? Maybe in conjunction with the PLL of the Costas loop?

Thanks a lot for Your help!
Paul
From: Jason on 16 Jun 2010 10:31
On Jun 16, 5:53 am, "Paul Wolf" <kstatnikov(a)n_o_s_p_a_m.gmail.com>
wrote:
> Dear professionals,
>
> I am a newbie to the topic of carrier recovery. I have to build up a
> point-to-point wireless communication link consisting of a quadrature
> modulator and demodulator. I think to apply -1/+1 PRBS signals on I and Q
> inputs of the transmitter. So I will get 4QAM constellation
> (45°/135°/-45°/-135°).
>
> 1. Is it possible to use a Costas loop to recover the carrier frequency and
> mix the received signal with this carrier frequency down? Please provide me
> with information about the appropriate Costas loop configuration for my
> case.
>
> 2. And another question: will the Costas loop give me 4-PSK constellation
> like (0°/ 90°/ 180°/ 270°) or the wanted one (45°/135°/-45°/-135°)
> ?
>
> 3. And the last question: which kind of AGC is easy to implement in analog
> way? Maybe in conjunction with the PLL of the Costas loop?
>
> Thanks a lot for Your help!
> Paul

4-QAM is just a rotated version of QPSK. Just treat the received QAM
symbols as if they were QPSK; the Costas loop will rotate the QAM
constellation onto the expected QPSK constellation automatically. You
still have to resolve the ambiguity as to which phase position the
loop locked in (there are 4 possibilities for the QPSK Costas loop),
but once that is cleared up you can make it work. You just need to
keep track of how the locked QPSK constellation relates to the QAM
constellation that you were expecting.

Jason
From: Tim Wescott on 16 Jun 2010 11:00
On 06/16/2010 02:53 AM, Paul Wolf wrote:
> Dear professionals,
>
> I am a newbie to the topic of carrier recovery. I have to build up a
> point-to-point wireless communication link consisting of a quadrature
> modulator and demodulator. I think to apply -1/+1 PRBS signals on I and Q
> inputs of the transmitter. So I will get 4QAM constellation
> (45°/135°/-45°/-135°).
>
> 1. Is it possible to use a Costas loop to recover the carrier frequency and
> mix the received signal with this carrier frequency down? Please provide me
> with information about the appropriate Costas loop configuration for my
> case.
>
> 2. And another question: will the Costas loop give me 4-PSK constellation
> like (0°/ 90°/ 180°/ 270°) or the wanted one (45°/135°/-45°/-135°)
> ?
>
> 3. And the last question: which kind of AGC is easy to implement in analog
> way? Maybe in conjunction with the PLL of the Costas loop?

Why do you need AGC, why can't you use one of the many standard methods,
and why do you think it needs to be implemented in conjunction with the PLL?

There are some wider system design concerns that can go into AGC design,
particularly if you have more than one gain stage that you can adjust.
Perhaps you should share more information about your receiver.

But at its simplest, AGC is an amplifier with a variable gain or a
variable attenuator stuck in the circuit somewhere; you monitor the
amplitude of your signal at some point downstream, and twiddle the
command to the variable gain/attenuation.

I think someone who'd been designing radios for as long as I've been
designing control loops could write an entire book on AGC, with six or
seven chapters devoted to theory and another six or seven devoted to all
the different ways it's ever been done, and why each method is good and bad.

--
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com
From: j on 16 Jun 2010 11:18
A comment re agc:

You really need to do a gain line up. Adding gain randomly can cause
you a lot of grief in terms of ISI. Too much gain in the wrong place
can cause sever distortion which will be signal dependent.

regards
From: Vladimir Vassilevsky on 16 Jun 2010 11:34


j wrote:

> A comment re agc:
>
> You really need to do a gain line up. Adding gain randomly can cause
> you a lot of grief in terms of ISI. Too much gain in the wrong place
> can cause sever distortion which will be signal dependent.

A comment re AGC:

1) There are only two types of AGC: backward or forward. Backward is
better suited for analog. The rest is the minor implementation issues.

2) You don't need an AGC. What you really need a pair of diodes. BAV99
would do.

VLV



 |  Next  |  Last
Pages: 1 2
Prev: forward error correction capabilities?
Next: OMAP-L1x Processor Universal Serial Bus 2.0