A question about RF NF Analyzer Architecture.

hp-8970

I read some articles about NF measurement.
The front-end Architecture of RF Noise Figure Analyzer is Filter+Amplifier+Up-converts mixer;
Why choose whis Architecture?
Why not Amplifier+filter+down-converts mixer?
A Low Noise Amplifier will reduce the Noise Figure of the measurement system which will improve the measurment accuracy.
And the equipments deal with low frequency signals is much cheaper and easy to produce.
Just like The Architecture of MW Noise Figure Analyzer:Amplifier+YIG filter+down-converts mixer.

Thank you!

rf nf

I guess it's related to the filter which is used to remove the image frequency to make SSB measurement.
Down-converts mixer need a very expensive YIG filter.
It's that right? Can anyone give me some useful information?
thanks.

what is nf analyzer?

The systems I have looked into had a broad band noise source applied to the system under test and a wide band true RMS voltmeter measuring the output.

The output noise power was measured with and without the noise source powered.

hp8970

The front-end filter is a low pass filter and it is necessary to reject all the frequencies out of the frequency band of the instrumentation (for example the image frequency of the noise signal to analyze).

The conversion type (of the mixer) is "upper" because otherwise, if w(IF) < w (RF) (that means down conversion mixer), you may have, due to the non ideal isolation between RF and IF ports of the mixer, a signal at the output of the mixer and believe it is due to a noise signal at the input (while it is due to the RF signal that come at the output of the mixer because the isolation isn't ideal).

Usually this is the architecture of the noise figure meter (see HP8970B architecture). The NFM is normally used with a test set to be able to performe noise measurements at millemeter wave frequencies. In this case the architecture may be:

isolator - LNA - YIG filter - MIXER - IF LNA ---> NFM

hp 8970 noise figure meter

Great Thanks to flatulent and marcomdd!

Yes, HP 8970 series NF Meter is a RF Noise Figure Meter(<3GHz).
And it needs test set(for example 8971C) to extend the measurement frequency.
The front-end Architecture of HP 8970 NFM (not the test set) is Filter+Amplifier+Up-converts mixer.
The first mixer in RF Noise Figure Meter(<3GHz) up-converts to an IF of around 4GHz.

But a Microwave Noise Figure Analyzers (Meter) also cover the RF portion of the spectrum (<3GHz) and in the RF region they have the same architecture as the RF Noise Figure Meter.

Agilent says Above the 3GHz point the architecture change into Amplifier+YIG filter+down-converts mixer. That's because up-converts need an IF and LO of unacceptably high frequencies. So the first (front-end) conversion must be a down conversion.

I can understand the front-end Architecture of a RF Noise Figure Analyzers (Meter).
But I can't understand the front-end Architecture of a Microwave Noise Figure Analyzers (Meter).

The difference is:
The front-end Architecture of RF Noise Figure Analyzer: Filter+Amplifier+Up-converts mixer;
The front-end Architecture of MW Noise Figure Analyzer: Amplifier+YIG filter+down-converts mixer.

8970 noise figure meter

htforever said:

Great Thanks to flatulent and marcomdd!

Yes, HP 8970 series NF Meter is a RF Noise Figure Meter(<3GHz).
And it needs test set(for example 8971C) to extend the measurement frequency.

Click to expand...

The freq range is 10 MHz - 1.6 GHz.

The front-end Architecture of HP 8970 NFM (not the test set) is Filter+Amplifier+Up-converts mixer.
The first mixer in RF Noise Figure Meter(<3GHz) up-converts to an IF of around 4GHz.

Click to expand...

It upconverts 2050 MHz with a YIG LO..

But a Microwave Noise Figure Analyzers (Meter) also cover the RF portion of the spectrum (<3GHz) and in the RF region they have the same architecture as the RF Noise Figure Meter.

@gilent says Above the 3GHz point the architecture change into Amplifier+YIG filter+down-converts mixer. That's because up-converts need an IF and LO of unacceptably high frequencies. So the first (front-end) conversion must be a down conversion.

Click to expand...

The architecture change because the output FREQUENCY of the test set (Amplifier+YIG filter+down-converts mixer) must to be in the frequency range of the NFM (RF Noise Figure Analyzers) so you necessary need a down conversion.

I can understand the front-end Architecture of a RF Noise Figure Analyzers (Meter).
But I can't understand the front-end Architecture of a Microwave Noise Figure Analyzers (Meter).

The difference is:
The front-end Architecture of RF Noise Figure Analyzer: Filter+Amplifier+Up-converts mixer;
The front-end Architecture of MW Noise Figure Analyzer: Amplifier+YIG filter+down-converts mixer.

Click to expand...

hp8970

generally speaking for a Noise Figure Meter (NFM) as the NF decrease, the measurement accuracy rise.
Lots of pubblication well describe the NFM effect on measurement; look for Agilent AN 57-1, AN 57-2, also visit www.maury.com.
But other factors should be taken in account. A mixer front end is cheaper and has less drift than Amplifier.
A mixer front-end is best suitable for DSB (double side band) measurement.
A LNA fron-end, especialy if followed by an YIG tuned band pass filter, is best suitable for SSB (single side band) measurement.
Agilent and Maury have different opinions about SSB vs DSB measurement.
Agilent say that SSB measurement is the best method.
Maury say that DSB measuremen is good enough.
If you use an economical DSB mixer measurement method, your NMF may have an intrinsic NF as low as 6 dB up to 26.5 GHz. Note that values aren't far from the NF you can get from good but expensive LNA placed at the input of SSB NFM.

I encourage You to carrefully read Agilent AN 57-1 and AN 57-2

hp 8971c

I do have read 57-1,57-2,57-3 from Agilent.
Making SSB or DSB measurement depends on the DUT property.
DSB measurement result is an average of USB and LSB. So for most DUT, it's not allowed and have to make a SSB measurement(need a RF filter in front of mixer).
If the DUT is designed to accept input signals on both sidebands(for example radiometer receivers), then a DSB measurement should be made.

There are two ways to make Microwave Noise Figure measurement:
1, Using RF Noise Figure Analyzer(Meter) +External Mixer Module(RF Filter,Mixer,LO, IF Filter and maybe LO filter and LNA).

2,Using a Microwave Noise Figure Analyzer(Meter,For example NFA series from Agilent).

Now I guess:
For a RF NFM, the IF frequency range is 10MHz~several GHz, if there is only one front-end Architecture in place(this can reduce cost and complexity), We have to use a up-converts mixer as the lowest RF frequency is only 10MHz.

For a MW NFM, in order to make measurement in the frequency range of 10MHz~26.5GHz, in the low frequency range, we can use the same front-end Architecture as a RF NFM(using a up-converts mixer). In the high frequency range, we have to use down-converts mixer in order to reduce the needs of high frequency LO and another expensive instrument. So a MW NFM has two kinds of front-end Architecture(one for RF NF measurement,another for MW NF measurement), it switchs at certain frequency point, and maybe 3GHz is just the frequency point.


For a RF NFM, because the RF filter has very small loss(maybe 1dB more or less),so we can put the LNA after the filter.
For a MW NFM, because the YIG filter has very very big loss,so we have to put the LNA before the filter in order to reduce the NF of the measurement system(F2).

Good !

A tell me something about the importance of an isolator in front of the receiver system !?

:lol: :?: :twisted:

p.s.

did you receive the papers I sent you yesterday ?

And about a thermometer to accurate know the Phys. Temperature ?