NF measurement of direct-convertsion mixer

[TWang]

y factor direct conversion

Hi everyone:

Does anyone know how to measure the NF of a direct-conversion mixer ?
Because it's a frequency converting device, it's difficult to find a noise source to cover both RF(5GHz) and IF(10KHz) bands, maybe we need two noise sources. One is for calibration of the instrument, another one is used as the DUT input. Is it correct ?

 

[Mazz]

y factor direct conversion receiver ic

I don't think it is correct.
The problem is that you need to know the ENR of your noise source both at 5GHz and 10 KHz. Usually RF noise sources have ENR specified from 10 MHz.

Being a mixer I don't expect it's noise figure to be low. So you can use the Gain Method as explained in :**broken link removed**

I hope it can help.
Mazz

 

[flyhigh]

direct conversion receiver and ssb or dsb nf

Hi,

No, one doesn't have to know the noise figure of the noise source at IF frequency! The noise is "folded" from RF to IF and noise source should be calibrated at RF.
Good old HP noise figure meter had a specific mode (1.9) for measuring frequency translating devices. Modern instrument have much more features, e.g. a choice of SSB or DSB noise measurements. As you have direct conversion receiver, you should use DSB mode to measure your noise figure.

flyhigh

 

[Mazz]

noise figure direct conversion measurement

I'm sorry, flyhigh but I have to disagree with you.

In the good HP 8970B Noise Figure Meter 10 MHz - 1600 MHz (with options it can go higher in frequency) the 1.9 special mode is different respect to other modes exactly because the calibration is divided into two steps: the first one measures the NF of the instrument receiver at RF freqs (using the ENR at RF) and the second one (transparent to the user) measures the NF of the instrument receiver at IF freqs (using the ENR at IF). So the minimum IF is 10 MHz and is too high for the direct conversion mixer output freqs. This is exactly the reason that doesn't allow the use of this instrument.
This problem can be solved using a spectrum analyzer (for example a R&S with FS-K3), but you have to know the ENR fo noise source also at IF (10 KHz) and this is not usually possible.
The simplest solution is the gain method.
Mazz

 

[flyhigh]

Hi Mazz

You are probably right about 10kHz frequency, I never used such a low IF in my measrements. Perhaps the instrument can extrapolate to such a low frequency? Perhaps not due to 1/f noise. Anyway, one learn something every day! Thanks!

flyhigh

 

[TWang]

Hi Mazz

Thanks for your answer. BTW, I found a noise source product from noisecom, which has the ENR as low as 10KHz (10KHz ~ 1GHz). If I use this one to calibrate the NF of the instrument first at IF frequency, and than, use another one (Agilent 346C 10MHz~26.5GHz) to be the noise input of the direct-conversion reciever, maybe i can measure the 1/f noise figure by using Y method. Is this correct ? Has anyone done this before? Thanks.

 

[Mazz]

I have done a brief search and I can tell you that is possible. (This is surprising me...but as flyhigh said, one learn something every day!)
Go at the end of page 13 of this application note: **broken link removed**
Mazz

p.s.: remember that your receiver (i.e. the noise figure meter) MUST have it's bandwidth as low as 10 kHz.

 

[TWang]

Thanks Mazz!
BTW, you mentioned about the Gain Method. Is that the typical way pepole use to measure the NF of a direct-conversion reciever ?

Regards,

Tao

 

[Mazz]

I cannot answer to you.
I think it depends on:
-the circuit you are measuring
-the equipment you have
-the effort you want to spend on it
-the accuracy needed.

The way you are following is for sure the most accurate.

Mazz

 

[njfl]

Thanks Mazz!
BTW, you mentioned about the Gain Method. Is that the typical way pepole use to measure the NF of a direct-conversion reciever ?

Regards,

Tao

I have done some studies on measuring NF of direct conversion receivers and found that both, the Y-Factor and Gain methods are equally suitible and accurate. The Gain method is a lot simpler to implement also.