OFDM FFT 802.11a/g QUESTION ~~~~~~~~~~~~~~~~~~~~

Hi,

    I'm somewhat of a newbie to the whole
orthogonal frequency division multiplexed
world, having come from the GSM, DCS, PCS,
and broadcast band power amplifier worlds,
as well as some 2-18 GHz DLVA work.

    It's very interesting that they use
53 carriers: 48 data, 4 pilots, and one
nulled carrier.

    My question is that it seems that
after our product does the A/D conversion on
the baseband signals, that we are left with 24
positive frequencies and 24 "negative" ones.

    I understand that the negative
frequencies are something like the
imaginary portions of the FFT, or
something like that.

    But if all the carriers are carrying
different data, then the negative
frequencies are not just a mirror image
of the positive carriers, so it's not
like a dual-sideband signal at all, where
you can get rid of one side and still
retain all the info.

    And how do they down-convert in 802.11,
and extract "negative" frequencies?  

    Thanks for your help.

S.

On 1 May 2005 23:43:41 -0700, radio913@aol.com wrote:

>Hi,
>
>    I'm somewhat of a newbie to the whole
>orthogonal frequency division multiplexed
>world, having come from the GSM, DCS, PCS,
>and broadcast band power amplifier worlds,
>as well as some 2-18 GHz DLVA work.
>
>    It's very interesting that they use
>53 carriers: 48 data, 4 pilots, and one
>nulled carrier.
>
>    My question is that it seems that
>after our product does the A/D conversion on
>the baseband signals, that we are left with 24
>positive frequencies and 24 "negative" ones.
>
>    I understand that the negative
>frequencies are something like the
>imaginary portions of the FFT, or
>something like that.
>
>    But if all the carriers are carrying
>different data, then the negative
>frequencies are not just a mirror image
>of the positive carriers, so it's not
>like a dual-sideband signal at all, where
>you can get rid of one side and still
>retain all the info.
>
>    And how do they down-convert in 802.11,
>and extract "negative" frequencies?  

It's exactly the same process as you've likely done in
GSM/PCS/whatever; a complex mix to baseband provides I and Q (aka real
and imaginary) channels that are orthogonal and remove any ambiguity
between positive and negative frequencies.   The only difference with
OFDM is that you take a DFT  (usually via an FFT) of the baseband
signal rather than processing it in the time domain.


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Hi,


    I'm somewhat of a newbie to the whole
orthogonal frequency division multiplexed
world, having come from the GSM, DCS, PCS,
and broadcast band power amplifier worlds,
as well as some 2-18 GHz DLVA work.


    It's very interesting that they use
53 carriers: 48 data, 4 pilots, and one
nulled carrier.


    My question is that it seems that
after our product does the A/D conversion on
the baseband signals, that we are left with 24
positive frequencies and 24 "negative" ones.


    I understand that the negative
frequencies are something like the
imaginary portions of the FFT, or
something like that.


    But if all the carriers are carrying
different data, then the negative
frequencies are not just a mirror image
of the positive carriers, so it's not
like a dual-sideband signal at all, where
you can get rid of one side and still
retain all the info.


    And how do they down-convert in 802.11,
and extract "negative" frequencies?


    Thanks for your help. 


S.

On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:

...

>    My question is that it seems that
>after our product does the A/D conversion on
>the baseband signals, that we are left with 24
>positive frequencies and 24 "negative" ones.
>
>
>    I understand that the negative
>frequencies are something like the
>imaginary portions of the FFT, or
>something like that.
>
>
>    But if all the carriers are carrying
>different data, then the negative
>frequencies are not just a mirror image
>of the positive carriers, so it's not
>like a dual-sideband signal at all, where
>you can get rid of one side and still
>retain all the info.
>
>
>    And how do they down-convert in 802.11,
>and extract "negative" frequencies?

Complex values.   Most implementations use direct-conversion to
baseband with a quadrature mix and two ADCs, one for the I channel and
one for the Q.   The FFT/IFFT is complex, so the output subcarriers
will be complex and therefore not necessarily symmetric about zero Hz.




Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Eric Jacobsen wrote:
> On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:
>
> ...
>
> >    My question is that it seems that
> >after our product does the A/D conversion on
> >the baseband signals, that we are left with 24
> >positive frequencies and 24 "negative" ones.
> >
> >
> >    I understand that the negative
> >frequencies are something like the
> >imaginary portions of the FFT, or
> >something like that.
> >
> >
> >    But if all the carriers are carrying
> >different data, then the negative
> >frequencies are not just a mirror image
> >of the positive carriers, so it's not
> >like a dual-sideband signal at all, where
> >you can get rid of one side and still
> >retain all the info.
> >
> >
> >    And how do they down-convert in 802.11,
> >and extract "negative" frequencies?
>
> Complex values.   Most implementations use direct-conversion to
> baseband with a quadrature mix and two ADCs, one for the I channel
and
> one for the Q.   The FFT/IFFT is complex, so the output subcarriers
> will be complex and therefore not necessarily symmetric about zero
Hz.
>

    I understand the addition of the 90 degree out
of phase of the quadrature component.  What i don't
get is what a "negative" frequency looks like in the
baseband.  Like what it would look like in the time-domain
on an oscilloscope.


S.

<radio913@aol.com> wrote in message
news:1115423351.853877.150650@o13g2000cwo.googlegroups.com...
>
> Eric Jacobsen wrote:
> > On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:
> >
> > ...
> >
> > >    My question is that it seems that
> > >after our product does the A/D conversion on
> > >the baseband signals, that we are left with 24
> > >positive frequencies and 24 "negative" ones.
> > >
> > >
> > >    I understand that the negative
> > >frequencies are something like the
> > >imaginary portions of the FFT, or
> > >something like that.
> > >
> > >
> > >    But if all the carriers are carrying
> > >different data, then the negative
> > >frequencies are not just a mirror image
> > >of the positive carriers, so it's not
> > >like a dual-sideband signal at all, where
> > >you can get rid of one side and still
> > >retain all the info.
> > >
> > >
> > >    And how do they down-convert in 802.11,
> > >and extract "negative" frequencies?
> >
> > Complex values.   Most implementations use direct-conversion to
> > baseband with a quadrature mix and two ADCs, one for the I channel
> and
> > one for the Q.   The FFT/IFFT is complex, so the output subcarriers
> > will be complex and therefore not necessarily symmetric about zero
> Hz.
> >
>
>     I understand the addition of the 90 degree out
> of phase of the quadrature component.  What i don't
> get is what a "negative" frequency looks like in the
> baseband.  Like what it would look like in the time-domain
> on an oscilloscope.

The signals just run backwards on the scope...
Sorry, couldn't resist the wise-crack on a Friday eve.

I hope you aren't starting another thread that starts an argument over real
and complex signals  - we've been over several rounds of  "real signals can
be sent over a wire, complex is just math and cannot be sent over wires".

>
>
> S.
>

On Fri, 6 May 2005 17:38:43 -0700, "Bhaskar Thiagarajan"
<bhaskart@deja.com> wrote:

><radio913@aol.com> wrote in message
>news:1115423351.853877.150650@o13g2000cwo.googlegroups.com...
>>
>> Eric Jacobsen wrote:
>> > On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:
>> >
>> > ...
>> >
>> > >    My question is that it seems that
>> > >after our product does the A/D conversion on
>> > >the baseband signals, that we are left with 24
>> > >positive frequencies and 24 "negative" ones.
>> > >
>> > >
>> > >    I understand that the negative
>> > >frequencies are something like the
>> > >imaginary portions of the FFT, or
>> > >something like that.
>> > >
>> > >
>> > >    But if all the carriers are carrying
>> > >different data, then the negative
>> > >frequencies are not just a mirror image
>> > >of the positive carriers, so it's not
>> > >like a dual-sideband signal at all, where
>> > >you can get rid of one side and still
>> > >retain all the info.
>> > >
>> > >
>> > >    And how do they down-convert in 802.11,
>> > >and extract "negative" frequencies?
>> >
>> > Complex values.   Most implementations use direct-conversion to
>> > baseband with a quadrature mix and two ADCs, one for the I channel
>> and
>> > one for the Q.   The FFT/IFFT is complex, so the output subcarriers
>> > will be complex and therefore not necessarily symmetric about zero
>> Hz.
>> >
>>
>>     I understand the addition of the 90 degree out
>> of phase of the quadrature component.  What i don't
>> get is what a "negative" frequency looks like in the
>> baseband.  Like what it would look like in the time-domain
>> on an oscilloscope.
>
>The signals just run backwards on the scope...
>Sorry, couldn't resist the wise-crack on a Friday eve.
>
>I hope you aren't starting another thread that starts an argument over real
>and complex signals  - we've been over several rounds of  "real signals can
>be sent over a wire, complex is just math and cannot be sent over wires".

Ha.  But complex *CAN* be sent over wires
if you have two wires.

See Ya',
[-Rick-]

"Bhaskar Thiagarajan" <bhaskart@deja.com> writes:

> <radio913@aol.com> wrote in message
> news:1115423351.853877.150650@o13g2000cwo.googlegroups.com...
>>
>> Eric Jacobsen wrote:
>> > On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:
>> >
>> > ...
>> >
>> > >    My question is that it seems that
>> > >after our product does the A/D conversion on
>> > >the baseband signals, that we are left with 24
>> > >positive frequencies and 24 "negative" ones.
>> > >
>> > >
>> > >    I understand that the negative
>> > >frequencies are something like the
>> > >imaginary portions of the FFT, or
>> > >something like that.
>> > >
>> > >
>> > >    But if all the carriers are carrying
>> > >different data, then the negative
>> > >frequencies are not just a mirror image
>> > >of the positive carriers, so it's not
>> > >like a dual-sideband signal at all, where
>> > >you can get rid of one side and still
>> > >retain all the info.
>> > >
>> > >
>> > >    And how do they down-convert in 802.11,
>> > >and extract "negative" frequencies?
>> >
>> > Complex values.   Most implementations use direct-conversion to
>> > baseband with a quadrature mix and two ADCs, one for the I channel
>> and
>> > one for the Q.   The FFT/IFFT is complex, so the output subcarriers
>> > will be complex and therefore not necessarily symmetric about zero
>> Hz.
>> >
>>
>>     I understand the addition of the 90 degree out
>> of phase of the quadrature component.  What i don't
>> get is what a "negative" frequency looks like in the
>> baseband.  Like what it would look like in the time-domain
>> on an oscilloscope.
>
> The signals just run backwards on the scope...
> Sorry, couldn't resist the wise-crack on a Friday eve.

Actually that's pretty close to the truth. In order to see a complex
signal on a scope, you'd use x-y mode, and in that mode a positive
frequency will move counterclockwise while a negative frquency moves
clockwise.

> I hope you aren't starting another thread that starts an argument over real
> and complex signals  - we've been over several rounds of  "real signals can
> be sent over a wire, complex is just math and cannot be sent over wires".

Then all those orthogonal signalling schemes (all PSKs, e.g.) are
being transmitted "illegally"???

Yeah, OK, so we use modulation to get two signals onto one wire. Still,
the point is that it isn't "just math."
-- 
%  Randy Yates                  % "...the answer lies within your soul
%% Fuquay-Varina, NC            %       'cause no one knows which side
%%% 919-577-9882                %                   the coin will fall."
%%%% <yates@ieee.org>           %  'Big Wheels', *Out of the Blue*, ELO
http://home.earthlink.net/~yatescr

Randy Yates wrote:

   ...

> Yeah, OK, so we use modulation to get two signals onto one wire. Still,
> the point is that it isn't "just math."

I can multiplex two signal of any type if I modulate. Call me back when 
you can send a baseband complex signal on a single wire in the same 
direction. (Hybrids allow two baseband signals to be transmitted on the 
same wire if they travel in opposite directions.)

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Randy Yates wrote:
> "Bhaskar Thiagarajan" <bhaskart@deja.com> writes:
>
> > <radio913@aol.com> wrote in message
> > news:1115423351.853877.150650@o13g2000cwo.googlegroups.com...
> >>
> >> Eric Jacobsen wrote:
> >> > On 4 May 2005 22:37:57 -0700, radio913@aol.com wrote:
> >> >
> >> > ...
> >> >
> >> > >    My question is that it seems that
> >> > >after our product does the A/D conversion on
> >> > >the baseband signals, that we are left with 24
> >> > >positive frequencies and 24 "negative" ones.
> >> > >
> >> > >
> >> > >    I understand that the negative
> >> > >frequencies are something like the
> >> > >imaginary portions of the FFT, or
> >> > >something like that.
> >> > >
> >> > >
> >> > >    But if all the carriers are carrying
> >> > >different data, then the negative
> >> > >frequencies are not just a mirror image
> >> > >of the positive carriers, so it's not
> >> > >like a dual-sideband signal at all, where
> >> > >you can get rid of one side and still
> >> > >retain all the info.
> >> > >
> >> > >
> >> > >    And how do they down-convert in 802.11,
> >> > >and extract "negative" frequencies?
> >> >
> >> > Complex values.   Most implementations use direct-conversion to
> >> > baseband with a quadrature mix and two ADCs, one for the I
channel
> >> and
> >> > one for the Q.   The FFT/IFFT is complex, so the output
subcarriers
> >> > will be complex and therefore not necessarily symmetric about
zero
> >> Hz.
> >> >
> >>
> >>     I understand the addition of the 90 degree out
> >> of phase of the quadrature component.  What i don't
> >> get is what a "negative" frequency looks like in the
> >> baseband.  Like what it would look like in the time-domain
> >> on an oscilloscope.
> >
> > The signals just run backwards on the scope...
> > Sorry, couldn't resist the wise-crack on a Friday eve.
>
> Actually that's pretty close to the truth. In order to see a complex
> signal on a scope, you'd use x-y mode, and in that mode a positive
> frequency will move counterclockwise while a negative frquency moves
> clockwise.
>

     I was thinking it might actually be something like
this, is the phasor moving clock-wise or counter-clock-wise?



> > I hope you aren't starting another thread that starts an argument
over real
> > and complex signals  - we've been over several rounds of  "real
signals can
> > be sent over a wire, complex is just math and cannot be sent over
wires".
>
> Then all those orthogonal signalling schemes (all PSKs, e.g.) are
> being transmitted "illegally"???
>
> Yeah, OK, so we use modulation to get two signals onto one wire.
Still,
> the point is that it isn't "just math."


     Sigh...  I don't think anyone here's really
gonna help me.....


S