Re: noise figure
NoiseFigure new
= sqrt ( ( NF1^2 - 1 ) + ( NF2^2 - 1 ) + ( NF3^2 - 1 ) + ... + 1 )
where all noise figures are ratios and referenced to a
single location.
A few points:
1 Definition:
Noise Figure is the ratio of increased noise in a system above the expected
level of Johnson noise. For our 50 ohm system that will give a voltage of
sqrt(4KTRBw) where K is Boltzmann's constant, T is temperature in absolute
Kelvin, R is the resistance of the system (for our case 25 ohms {50 ohms in
parallel with 50 ohms}), and Bw is the bandwidth of interest in Hertz. For
a Bw of 1 MHz that yields a noise floor of around 0.641uV, or -3.9dBuV
2 Relate noise figure to some place in your system.
For simplicity (and ease of using specs) relate to the "front end" Noise
figure specs relate to the front end of a system block. The noise is
amplified, or attenuated, along with the signal and therefore track
together. As you go from the output to the input of each system block, gain
subtracts from the NF and attenuation adds to the NF. Modify the Noise
Figure by each block you must go through to get to the input. For example,
through cable loss, add the few dB. For gain, subtract the gain.
3 Make up a list (Use Excell spread sheets) You will end up with
contributions from every block now referenced to the front end. then...
4 Remember that uncorrelated noise does not add, but adds as the square
root of the sum of squares. However! you must only take into consideration
the noise contribution from each "additional" noise source. You cannot keep
adding in the contribution from the 50 ohm source impedance. Therefore,
each Noise Figure ratio must be squared and then have 1 subtracted from it.
After combining all the contributions, you will add the 1 back. Simply take
the square root of that total sum and find the 20log value and you will have
your total system's new NoiseFigure.
For example, let's find the noise figure for a receiving system that
attaches to an antenna consisting of long cable, amp, cable, and spectrum
analyzer (SA).
Passive devices have 0 dB noise figure (they do not add any noise)
Antenna is passive, but converts volts per meter into volts in a 50 ohm
system. Since it does not add any noise, there is no difference between
minimizing the NF at the antenna port or at the field that it measures.
So let's find out the NF of our system at 200MHz
long cable 3dB attenuation
amp24dB 6 dB NF
cable -
SA 32dB
Note: You could have a perfect receiver that contributes no noise located
after the long cable and you would still have a 3 dB NF That's why
amplifiers are placed near the signal source.
Check your particular SA. It can have a Noise Figure from 26dB to 36dB
depending on its design. That means for a 1 MHz bandwidth you can only see
down to around -80dBm.
So let's move the amplifier out to the antenna and change the order of the
list:
cable -
amp gain 24dB with NF = 6 dB
long cable atten 3dB
SA with NF = 32dB
The list would show 6 dB NF at the input and (32-24+3=11) 11dB from the SA.
That is a ratio of 2 and a ratio of 3.55.
New Noise Figure is sqrt( (2^2-1) + (3.55^2-1) + 1 ) = 4 or 12dB
See how the SA still dominates?
To rewrite the equation:
NoiseFigure new
= sqrt ( ( NF1^2 - 1 ) + ( NF2^2 - 1 ) + ( NF3^2 - 1 ) + ... +
1 )
where all noise figures are ratios and referenced to a
single location.
- Robert -
Robert A. Macy, PEm...@california.com
408 286 3985 fx 408 297 9121
AJM International Electronics Consultants
619 North First St, San Jose, CA 95112
-Original Message-
From: KC CHAN [PDD]
To: emc-p...@majordomo.ieee.org
List-Post: emc-pstc@listserv.ieee.org
Date: Friday, December 21, 2001 5:46 PM
Subject: noise figure
>
>Hi all
>
>It may not be purely EMC question, actually it is RF related, but I am sure
the experts here can answer my questions.
>
>We all know that we need to have a pre-amp. that is as lower noise figure
as possible, but how low it is enough or how it is related to the noise
floor viewed by a receiver or spectrum analyzer.
>
>Thank you
>KC Chan
>
---
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.
Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/
To cancel your subscription, send mail to:
majord...@ieee.org
with the single line:
unsubscribe emc-pstc
For help, send mail to the list administrators:
Michael Garretson:pstc_ad...@garretson.org
Dave Healddavehe...@mediaone.net
For policy questions, send mail to:
Richard Nute: ri...@ieee.org
Re: noise figure
I may not have been as clear as desirable in the line where I say, "N = -67 dBuV/Hz." That is based on a noise figure numerically equal to one, or 0 dB. To give a numerical example, say it is a radiated measurement where you have determined that the limit is 40 dBuV/m (30 - 88 MHz), the antenna factor is 13 dB/m with a 3 dB pad at antenna to reduce mismatch, and cable loss is negligible. If we desire that the noise floor of the receiver/analyzer be 6 dB below the limit, then N (dBuV) = 40 dBuV/m - 13 dB/m - 3 dB - 6 dB = 18 dBuV With a perfect noise figure (0 dB) N = -67 dB uV/Hz + 10 * log (120,000 Hz) = -16 dBuV That means that your analyzer/receiver noise figure can be no greater than 18 dBuV - (-16 dBuV) = 34 dB. This is easily achievable with a receiver. You might be right on the edge with an HP8566 spectrum analyzer with no video filtering. The HP8566 has a 29 dB noise figure with full video filtering. -- >From: "Ken Javor" >To: "KC CHAN [PDD]" , emc-p...@majordomo.ieee.org >Subject: Re: noise figure >Date: Sat, Dec 22, 2001, 3:59 AM > > > Take the limit you are measuring to, and subtract the transducer factor and > any cable losses. That yields the receiver signal in dBuV which corresponds > to a spec level signal. Subtract off a factor for how much you want the > receiver noise floor to be below the spec level, typically 6 dB. Call this > level N. Then basic statistical thermodynamic theory says that N = kTBF, > where K is Boltzman's constant (the ideal gas law constant divided by > Avogadro's number), T is absolute temperature in Kelvin, B is bandwidth in > Hertz, and F is the noise figure. At room temperature, 25 degrees Celsius > or 298 Kelvin, N = -67 dBuV/Hz, with the bandwidth dependency going as 10 > log (BW) because the basic equation is for noise power. Now given your > bandwidth, say 120 kHz, you can solve for the necessary noise figure. If > your receiver/analyzer noise figure is higher than that needed, there is > another simple equation that will allow you to specify a preamp that will > allow the combination measurement system to perform as required. > > -- >>From: "KC CHAN [PDD]" >>To: >>Subject: noise figure >>Date: Fri, Dec 21, 2001, 6:42 PM >> > >> >> Hi all >> >> It may not be purely EMC question, actually it is RF related, but I am sure >> the experts here can answer my questions. >> >> We all know that we need to have a pre-amp. that is as lower noise figure >> as possible, but how low it is enough or how it is related to the noise >> floor viewed by a receiver or spectrum analyzer. >> >> Thank you >> KC Chan >> >> >> --- >> This message is from the IEEE EMC Society Product Safety >> Technical Committee emc-pstc discussion list. >> >> Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ >> >> To cancel your subscription, send mail to: >> majord...@ieee.org >> with the single line: >> unsubscribe emc-pstc >> >> For help, send mail to the list administrators: >> Michael Garretson:pstc_ad...@garretson.org >> Dave Healddavehe...@mediaone.net >> >> For policy questions, send mail to: >> Richard Nute: ri...@ieee.org >> Jim Bacher: j.bac...@ieee.org >> >> All emc-pstc postings are archived and searchable on the web at: >> No longer online until our new server is brought online and the old >> messages are imported into the new server. >> > > --- > This message is from the IEEE EMC Society Product Safety > Technical Committee emc-pstc discussion list. > > Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ > > To cancel your subscription, send mail to: > majord...@ieee.org > with the single line: > unsubscribe emc-pstc > > For help, send mail to the list administrators: > Michael Garretson:pstc_ad...@garretson.org > Dave Healddavehe...@mediaone.net > > For policy questions, send mail to: > Richard Nute: ri...@ieee.org > Jim Bacher: j.bac...@ieee.org > > All emc-pstc postings are archived and searchable on the web at: > No longer online until our new server is brought online and the old > messages are imported into the new server. > --- This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ To cancel yo
Re: noise figure
Take the limit you are measuring to, and subtract the transducer factor and any cable losses. That yields the receiver signal in dBuV which corresponds to a spec level signal. Subtract off a factor for how much you want the receiver noise floor to be below the spec level, typically 6 dB. Call this level N. Then basic statistical thermodynamic theory says that N = kTBF, where K is Boltzman's constant (the ideal gas law constant divided by Avogadro's number), T is absolute temperature in Kelvin, B is bandwidth in Hertz, and F is the noise figure. At room temperature, 25 degrees Celsius or 298 Kelvin, N = -67 dBuV/Hz, with the bandwidth dependency going as 10 log (BW) because the basic equation is for noise power. Now given your bandwidth, say 120 kHz, you can solve for the necessary noise figure. If your receiver/analyzer noise figure is higher than that needed, there is another simple equation that will allow you to specify a preamp that will allow the combination measurement system to perform as required. -- >From: "KC CHAN [PDD]" >To: >Subject: noise figure >Date: Fri, Dec 21, 2001, 6:42 PM > > > Hi all > > It may not be purely EMC question, actually it is RF related, but I am sure > the experts here can answer my questions. > > We all know that we need to have a pre-amp. that is as lower noise figure > as possible, but how low it is enough or how it is related to the noise > floor viewed by a receiver or spectrum analyzer. > > Thank you > KC Chan > > > --- > This message is from the IEEE EMC Society Product Safety > Technical Committee emc-pstc discussion list. > > Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ > > To cancel your subscription, send mail to: > majord...@ieee.org > with the single line: > unsubscribe emc-pstc > > For help, send mail to the list administrators: > Michael Garretson:pstc_ad...@garretson.org > Dave Healddavehe...@mediaone.net > > For policy questions, send mail to: > Richard Nute: ri...@ieee.org > Jim Bacher: j.bac...@ieee.org > > All emc-pstc postings are archived and searchable on the web at: > No longer online until our new server is brought online and the old > messages are imported into the new server. > --- This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ To cancel your subscription, send mail to: majord...@ieee.org with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Michael Garretson:pstc_ad...@garretson.org Dave Healddavehe...@mediaone.net For policy questions, send mail to: Richard Nute: ri...@ieee.org Jim Bacher: j.bac...@ieee.org All emc-pstc postings are archived and searchable on the web at: No longer online until our new server is brought online and the old messages are imported into the new server.
Re: Noise Figure
>I am trying to determine the maximum noise figure that I can have for my 1-16.5 GHz EMC receiver system which includes the horn, mixer for above 18 GHz, preamp, cable and spectrum analyzer. I know the following: * Signal level received by the horn * Gain through the system * Noise figure from the horn to the spectrum analyzer * Noise floor of the spectrum analyzer >I understand that noise figure is the ratio of the s/n of the input to the s/n of the output of the system, so if I need an output s/n of 6 dB and have a NF of 14 dB, I need a s/n input to the horn of 20 dB. So it appears that I have one missing piece of information - the input noise to the horn. >Now I am lost. Suggestions? Not sure how to proceed, so I always go back to basics. Pretend we have a system of antenna, cable, amp, and spectrum analyzer. The antenna is passive and has a NF of zero The cable has a loss of 2 dB. The amp has a gain of 24 and a NF of 1.8 dB (pretty good amp) The spectrum analyzer has a NF of around 32 dB (You can verify that from calculating the noise floor, minimum reception levels, etc.) The noise power from Johnson noise in a 50 ohm system is 4KT*50*BW, where K is 1.38022E-23, T is 277C, and let BW be a variable for later. Therefore, the Johnson noise voltage, enoise = 8.7444E-10, or -181dBV, per rt Hz. You said emc receiver so we should use dBuV, but I'm more familiar with dBm, so I'll use that til we're done. enoise is -168dBm per rt Hz. The emc spectrum analyzer will have a receive bandwidth of 120KHz so the Johnson noise floor is around -117dBm. Since the spectrum analyzer has a noise figure of around 32 dB, you get roughly a receiver that can "see" down to -85dBm, (-117 +32 = -85dBm) Actually noise bw is wider than the receive bw, so you won't be able to see down to -85dBm. Because noise power does not add linearly, you have to reference all signals at a single point to rms the noise contributions. I usually work backward to the input leaving the bw effect til last. So we have, a table of noise contributions: spectrum analyzer: 32-24+2 = 10 ! boy does that spectrum analyzer dominate, eh? amp: 1.8+2 = 3.8, good argument for amp then cable, too. rms all values: sqrt(1.55^2+3.16^2) = 3.52, or the system NF is 10.9 dB. You can see the spectrum analyzer dominate. That means the minimum signal out of the antenna that you can see will be -117 + 11, or -106 dBm. Note that's better than the straight spectrum analyzer, but still could be better. Massaging the technique should get you to the answer you were looking for. - Robert - - This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to majord...@ieee.org with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to ed.pr...@cubic.com, j...@gwmail.monarch.com, ri...@sdd.hp.com, or roger.volgst...@compaq.com (the list administrators).
