flicker noise in upconversion mixer

[pugongying]

upconversion + noise

It seems that it is similar to phase noise in vco
but I haven't found any reference about flicker noise in upconversion mixer
can you recommend some good material?
thanks very much

 

[amr87]

flicker noise mixer

I have the same problem here. I am designing a direct upconversion Mixer using the Gilber-cell scheme and the flicker noise is really an issue.

I think that it is a must in active mixers and it depends on whether your modulation scheme is immune to such noise near the DC (I am talking about direct upconversion mixers).

There is a good paper "Noise in RF-CMOS mixers" A simple Physical Model",A.Abidi and H.Darabi.
It is about noise mechanisms in down-conversion mixers. But you can also understand the dominant noise sources in the up-conversion counterpart.

By the way, had you found any material regarding flicker noise in up-conversion mixers?

 

[biff44]

upconversion flicker noise

CMOS does have flicker noise! Maybe you would have better luck finding papers on CMOS oscillator design and how they combat flicker noise. Then you can try the same thing for mixers. If you have the room, one approach is to have an analog control circuit with a lot of gain (70 dB or so) and clean up the audio flicker noise. Once it is cleaned up at low frequency, it is no longer there to upconver to the microwave band.

How the device is biased up directly affects the amount of flicker noise. In a fet, if the depletion region sweeps into an area where the doping is changing, or near a metalization region, you can sweep charges into trapped states, where they linger and eventually randomly get released and cause a flicker "pulse" of current. So the more linearly the device is biased, the less flicker noise.

There was a thread a couple weeks ago on crystal oscillators that discussed some methods.

Rich

 

[demian1977]

direct up conversion mixer

Fliker noise gives problem only if you ara working below 1MHz, if tou are working 100MHz or above, it's not a problem. You should use low fliker noise shottky diode if the frequency is very low, and you need low phase noise (radar application for exemple)

 

[biff44]

noise in rf cmos mixer a simple physical model

I would tend to disagree with that. You can upconvert a 10 Ghz signal with a 30 Ghz LO, and still have phase noise 100 Hz from the carrier on the output that is due to 100 Hz audio flicker noise in the device!

Do the math: (sin Fif*t) * (sin Flo*t) * (sin Fflicker*t)=?
You will see all the low frequency flicker component get included at all output frequencies.