James Post subject: Inquiry of noise figure Unread postPosted:
Sat Mar 12, 2005 10:36 am Dear Friends, I am trying to
improve the noise figure my system. There is like this.
a) Current BPF component : 5.15 ~ 5.825GHz (700MHz Bandwidth)
b) Will be changed BPF component : 5.725 ~ 5.825GHz (100MHz Bandwidth)
If I changed the a) to b), noise figure will be improved up to 5dB.
Is that right? Assuming same input signal comes. I appreciate
any help or comment or feedback, Many thanks :roll:
Top Kirt Blattenberger Post subject: Here's
a start Unread postPosted: Sat Mar 12, 2005 6:38 pm Offline
Site Admin User avatar Joined: Sun Aug 03, 2003 2:02 pm
Posts: 308 Location: Erie, PA Greetings James: Changing
the filter bandwidth will not affect the noise figure calculation, only
the insertion loss (gain) of the filter will. However, reducing the
bandwidth by a factor of 7 as you propose will reduce the noise power
by 10*log(7)=8.45 dB. That in turn will improve your signal-to-noise
ratio by 8.45 dB, since the noise power is dependent on the system bandwidth.
So, the minimum detectable signal will be 8.45 dB lower. That translates
to a range increase factor of 10^(8.45/20)=2.65. These calculations
hold for ideal environment and assume no other factors in the system
equation change. It also assumes that the relative bandwidth changes
are the noise power bandwidths of the filters and not just the 3 dB
bandwidths, and that the final bandwidth at the detector is set by your
new filter. At 100 MHz, it likely is not. The benefit you will
most likely realize is a reduction in the out-of-channel interference
both from direct signals and from intermodulation products created by
those out-of-band signals that end up in-band. For all the formulas
you need to do the calculations yourself, please go to this page and
click on the links of interest. https://www.rfcafe.com/references/electrical.htm
On this page is a simple online cascade calculator: https://www.rfcafe.com/references/spread
... alc-ss.htm Also, you can download RF Workbench (shareware
by RF Cafe) that includes calculations for determining filter noise
bandwidth: https://www.rfcafe.com/business/software ... kbench.htm
- Kirt Blattenberger :smt024 Top Profile
Guest Post subject: Unread postPosted: Sat Mar 12, 2005 8:01
pm Dear Kirt: Thank for your careful answer regarding my
question. Let me confirm a few things as follows: 1. I think
the sensitvity will be 8.45dB better if I change 5.15 ~ 5.825GHz (700MHz
Bandwidth) to 5.725 ~ 5.825GHz (100MHz Bandwidth). Is that right? Is
sensitivy related with system bandwidth? 2. Based upon your reply,
could you explain more details regarding range increase means? Does
it cover more distance? So, the minimum detectable signal will
be 8.45 dB lower. That translates to a range increase factor of 10^(8.45/20)=2.65.
Best Regards, James Top Kirt Blattenberger
Post subject: Unread postPosted: Sat Mar 12, 2005 9:43 pm
Offline Site Admin User avatar Joined: Sun Aug 03, 2003
2:02 pm Posts: 308 Location: Erie, PA Greetings James:
1. Your sensitivity will be improved if the changed BW represents
the final BW that your detector or smapling system will see. That is,
if somewhere down the receive chain, there is, say, a 50 kHz bandwith
filter prior to detection, then having reduced the other filter from
700 MHz down to 100 MHz will have no measureable effect. As mentioned,
the benefit you might experience would be a reduction of interfering
signals that get translated inband. That still counts as a signal-to-noise
improvement, but in a different way. 2. Signal strength in power
falls off at a rate of 20 log(2) dB, which is approximately 6.02 dB
for every doubling of the distance. So, every 6.02 dB of improved sensitivity
theoretically results in being able to receive a signal twice as far
away. I say theoretically because in the real world, multipath and fading
due to barriers can make the change something other than 6.02 dB.
- Kirt Blattenberger :smt024
Posted 11/12/2012
|