Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

mixer downconversion noise output level dB/Hz

yefj

Advanced Member level 5
Advanced Member level 5
Joined
Sep 12, 2019
Messages
1,609
Helped
1
Reputation
2
Reaction score
5
Trophy points
38
Activity points
9,714
Hello, I have a mixer shown in the link below.The way I calculate the noise coming out of the mixer is as following:
-174+NF-conversion loss=-174+0.5-6=-179.5dB/Hz so far I am correct?
also the
Mixer Noise Figure typically measures within +0.5 dB of conversion loss for IF frequencies greater than 5 MHz.
However I was told that we add another 3dB because we convert noise from RF twice.
If its true, Is there some visualisation I could use to see why we add another 3dB?
Thanks.

1739178640270.png
 
The output of the mixer is F.LO +/- F.RF. You have to account for the image frequency.
 
Noise O/P will be -174dBm/Hz + NF at LSB -Loss at LSB + -174dBm/Hz + NF at USB -Loss at USB.
Don't forget to convert to power (W) before adding. It is only necessary to calculate noise at both frequencies if loss or NF are different.
 
Hello G4BCH,LSB is lower side band USB upper side band.I am doing downconvertion from 10GHz to 0.5Ghz using 9.5GHz LO signal.
How can I see the NF at low and high side bands in the datasheet shown below?
"Noise O/P will be -174dBm/Hz + NF at LSB -Loss at LSB + -174dBm/Hz + NF at USB -Loss at USB."
 
Since the upper and lower sideband are so close, in this case the noise figure and loss for the will be the same for both USB & LSB, 6 to 7.5dB. Without seeing the plots I can't say exactly what it would be, probably nearer 6dB, but I were using this in a system I'd use 7.5dB to be on the safe side.
 
Hello G4BCH , could you give me an example where we do need to look in the data sheet and find the noise in upper and lower side bands.
Is there some data sheet you could reccomend ?
Thanks .
 
If there is a big difference between upper & lower sideband frequencies and the data sheet sows the NF varies over that range. Usually you don't need to account for it.
 
Hello, Tony , Is there some sample datsheet I could use its plots to see the noise math of DSB mixer?
Thanks.
--- Updated ---

If there is a big difference between upper & lower sideband frequencies and the data sheet sows the NF varies over that range. Usually you don't need to account for it.
Could you give practical eaxmple from a datasheet where I could look at its plots and calculte the noise?
Thanks.
 
Manufacturers don't give plots that would make that very easy to just read off the conversion loss. You can get an estimate from some data sheets. Take for example the minicircuits SYM-63LH https://www.minicircuits.com/pages/s-params/SYM-63LH+_GRAPHS.pdf ( a randomly picked example) the conversion loss over frequency is only plotted for an IF of 10MHz ( easy to measure). The shape of the loss plot will be similar for different IFs. If you were to have an IF of 500MHz and an LO of 2000Hz then the loss at 1500MHz would be about 7dB and the loss at 2500 would be about 9dB.
A similar situation can occur if you have an image rejection filter or are using an image rejection mixer and the rejection at the image frequency is not very high. In those situations the added image noise has to be taken into account. This I am told is particularly important in radio astronomy where all noise sources have to be accounted for.
 
From the graphs in that particular mixer, which was just chosen at random, it is not easy or may be not possible to determine the loss under all conditions. Usually the conversion loss plot is close enough for most IFs. If in doubt the safest way is to measure the performance under the conditions that it is being used.
To determine the output noise, as I said above you have to treat signal and image signal paths separately and then combine them. If this is used as one of you other posts where you showed an RF amplifier you will also have to account for the gain and NF of that at both signal and image frequencies.
 
Hello G4BCH,my signal is noise, suppose my LO=9.5Ghz and RF is 10Ghz.
could you please visualize where is the "image"?
Thanks.
"treat signal and image signal paths separately and then combine them"
 
Yes, So in a mixer we are downconverting twice the noise.two questions:
1.what if the LO frequency will not be so close to the RF frequency ,what do I need to do in such case?could you give an example?
2.I am driving the LO input of a mixer with the follwoing amplifier.I know that LO power influences the conversion loss, is there something else?
Is the LO SNR signal makes a contribution to the output noise?
do i need to look somewehere in the datasheet of this amplifier to see its contribution to the output noise of a mixer?


1739783408314.png
 
1.what if the LO frequency will not be so close to the RF frequency ,what do I need to do in such case?could you give an example?
There will always be an image response, it comes from the mathematics of the mixing process. Multiply 2 sine signals together.
Your only option is to use a filter to reduce the image signal; only of use if you have an amplifier in front of the mixer, or to use an image rejection mixer.
2.I am driving the LO input of a mixer with the follwoing amplifier.I know that LO power influences the conversion loss, is there something else?
Yes it can have an effect on the noise at output of the mixer (see blow), if the LO is large enough the level does not affect the conversion loss significantly.
Is the LO SNR signal makes a contribution to the output noise?
In this case maybe. the amplifier is high gain and wideband. The noise at the mixer input is:
Thermal noise (-174dBm/Hz) + amplifier gain + amplifier noise figure - mixer LO-RF isolation
in this case -174 + 21 +7.5 + 35 = -180.5dBm/Hz best case
or -174 + 21 +7.5 + 55 = -170.5dBm/Hz worst case
do i need to look somewehere in the datasheet of this amplifier to see its contribution to the output noise of a mixer?
See above.

If you have and RF amplifier I would use a bandpass filter at the mixer input that will give sufficient rejection at the image frequency. You know what your system requirements are so it should not be hard to work out.
I would also add a filter in the LO path to reduce the AM noise at the image frequency. Whether this a really needed depends on whether the RF amplifier has sufficient gain such that the input noise swamps the LO noise. Again you know what you have in your system and what the requirements are..
 
Hello G4BCH, why our noise is dB/Hz
Why do all the calculations in this unit until we multiply by the BW.
I am used to signal by power dBm units.
But in noise case we just say -174+NF+IL+3
And it’s all dB/Hz
The /Hz is confusing because we act as if it’s regular power input will gain etc.. but we just stick the /Hz .
Is there some intuition regarding this ?
Thanks .
 

LaTeX Commands Quick-Menu:

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top