FPGA Central

Hi,

I made a complete DSSS demodulator with DBPSK and DQPSK demodulation.

In the case of DQPSK demodulation I've got the same PN_code on I and Q
branchs.

What is the advandatage to put different PN_code on I and Q ?

And what is the advantage to make a complex spreading like this :
At spreading (modulation) : (I+jQ).(PN_I+jPN_Q)
So I signal becomes : I.PN_I-Q.PN_Q and Q signal becomes Q.PN_I+I.PN_Q

I'm looking for best performance.

I've got these sensitivity on a 2.4 GHz radio with lenght 11 of the
PN_code
and a 44 MHz radio band :
-92 dBm at 2 Mbps (theory says -94 dBm) in DBPSK at 1E-5 BER
-88 dBm at 4 Mbps (theory says - 91 dBm) in DQPSK at 1E-5 BER

Thanks...

[email protected] wrote:
> Hi,
>
> I made a complete DSSS demodulator with DBPSK and DQPSK demodulation.
>
> In the case of DQPSK demodulation I've got the same PN_code on I and Q
> branchs.
>
> What is the advandatage to put different PN_code on I and Q ?
>
> And what is the advantage to make a complex spreading like this :
> At spreading (modulation) : (I+jQ).(PN_I+jPN_Q)
> So I signal becomes : I.PN_I-Q.PN_Q and Q signal becomes Q.PN_I+I.PN_Q
>
> I'm looking for best performance.
>
> I've got these sensitivity on a 2.4 GHz radio with lenght 11 of the
> PN_code
> and a 44 MHz radio band :
> -92 dBm at 2 Mbps (theory says -94 dBm) in DBPSK at 1E-5 BER
> -88 dBm at 4 Mbps (theory says - 91 dBm) in DQPSK at 1E-5 BER
>
> Thanks...

usually you judge modulation and coding performance against a signal to
noise ratio such as C/N or Es/No or Eb/No. The performance measured
against input power in dBm will be a function of some of the receiver
hardware perfromance such as gain and noise figure, not just the
modulation and coding.

Mark

We can calculate the sensitivity if we know the (S/N) at the input of
the demodulator and the noise floor and the noise figuire of the
receiver like this :

Sensitivty=Noisefloor + NoiseFigure-Gp+(S/N)demod
Gp : process gain for DSSS

and for DBPSK a BER of 1E-5 is achieved with a S/N of 10.5dB
and DQPSK with S/N of 13.5 dB

ok

so are you achiving that performance at those S/N at the demod input
and does that agree with theory?

Mark

we have 2 dB less performance for sequence of lenght 11
but with long sequence (63) whe have 5dB less than theory
we think that our radio is less performant when we approach the noise
floor....

I try to seperate the perfromance of the "radio" from the perfromance
of the "demod".

The performance of the "radio" is a function of noise figure, phase
noise etc.

If you measure the S/N at the INPUT TO THE DEMOD (not the radio) and
the BER out of the demod and conpare to theory, that will give you the
performance of the demod and the coding.

If the S/N at the input to the demod is not what it should be, then you
need to improve the performance of the "radio" , i.e. improve the
noise figure.

Mark

So I repost my question :

I made a complete DSSS demodulator with DBPSK and DQPSK demodulation.

In the case of DQPSK demodulation I've got the same PN_code on I and Q
branchs.

What is the advandatage to put different PN_code on I and Q ?

And what is the advantage to do a complex spreading like this :
At spreading (modulation) : (I+jQ).(PN_I+jPN_Q)
So I signal becomes : I.PN_I-Q.PN_Q and Q signal becomes Q.PN_I+I.PN_Q