Noise floor raising over the receive chain

[deepaksingh2112]

Dear All,

I am working with a receiver (2 stage down conversion) for Surveillance Radar. My receiver chain is having one 30 dB programmable attenuator after LNA stage and before 1st stage down conversion.

Complete receive chain gain with the full attenuation and without any attenuation is 25 dB & 55 dB respectively. My receiver Bandwidth is 1.4 MHz. I am getting Noise floor of -52 dBm in both the cases i.e. with 55 dB gain and with 25 dB gain.

My receiver dynamic range is -20 dBm to -110 dBm. So while feeding pulsed signal(1us) below -95 dBm at LNA, IF2 signal power (-40 dBm or below),signal is getting burried in noise floor and spectrum is not able to trigger the same.

Can any one tell me
1. whether the Noise floor can be reduced or Not?
2. Why noise floor is maintained at same level with 55 & 25 dB gain?
3. Why noise floor is around 80 dBm in case of CW pulse but -52 dBm in case of pulsed signal?

Thanks for your suggestion.

 

[chuckey]

The noise floor is set by the NF of the LNA, once its added its noise further amplification will not change the NF. (1. and 2.)
3. With a CW, you have in any given time, 100% noise + 100% signal, with a pulsed setup, you have 100% noise + duty cycle X signal. ?
Frank

 

[deepaksingh2112]

Dear Chuckey,

Can you plz explain further.

As per your explanation, Noise floor in case of pulsed signal will be less as compared to continuous signal. But in my case I am getting high noise floor w.r.t CW signal.

Mentioned Power level in my first comment is wrong. Observed noise floors are as follows.

Noise floor in case of CW signal is : -80 dBm
Noise floor in case of pulsed signal is : -52 dBm

As per your statement, Noise floor will not be affected by further gain after LNA.

But as much as I have studied, Noise figure only will be controlled by NF of LNA, But noise floor will be affected by receiver chain gain.

Kindly clarify the same.

 

[vfone]

If the LNA gain is high enough (>15dB) and there are no switched attenuators in front of it, might be an AGC setting problem when you go to a pulsed signal.
Check the AGC transfer function (attack time, decay, knee, etc.).

 

[deepaksingh2112]

Dear vfone,

I have used one AGC attenuator after 2nd stage down conversion. Part No is : HMC624LP4E

Kindly clarify what to mean to say in above statement, whether the selected AGC is OK or not.

 

[biff44]

Observed noise floors are as follows.

Noise floor in case of CW signal is : -80 dBm
Noise floor in case of pulsed signal is : -52 dBm
.

measured where, at the output of the receiver chain, at the input? How about a detailed block diagram showing the components?

 

[mtwieg]

The noise floor is set by the NF of the LNA, once its added its noise further amplification will not change the NF. (1. and 2.)

This is only a rough assumption. If you have a 15dB LNA and follow it with a 30dB attenuator, then the overall NF is definitely going to be much higher than the LNA NF (by around 15dB).

If your receive chain is well designed, changing the attenuation/gain should change the noise floor. That's how you know your noise is dominated by the source, not the intermediate stages in the chain.

As for the pulsed vs CW issue I'm not quite understanding how you're testing this.

 

[volker@muehlhaus]

Also, do we possibly have a misunderstanding? Noise figure vs. noise floor (as measured with spectrum analyzer)?

 

[vfone]

If the receiver have the SAME SYSTEM GAIN for both, CW operation and pulsed mode operation, there should be no difference in the system noise floor (or system noise figure) when the receiver receives continuous wave or a pulsed wave signal.
Any decrease of the system gain due to AGC (whatever stage) it will increase the total noise figure of the receiver.
As I mentioned before, check the AGC transfer function which might affect your noise figure measurement in pulsed mode if is not set correctly.

 

[ge]

Noise Figure, Noise Floor, SNR
Pick your noise, lol.

 

[shashy.br]

Hi Deepak its me shashank,

What are the seetting of the spectrum analyzer that u have used??

Have u tried processing the signal using pulse compression technique?