Why Gilbert noise figure is more at IF frequency than at RF frequency?

[kapil86]

Respected sir/mam

I have designed Gilbert mixer, When i do noise figure(NF) simulation at IF (2 MHz) and RF (2.4 GHz) Frequencies, i found that NF at IF is more than (4dB) to RF frequency. What is the region for that. someone help me please.

Thank you

 

[tony_lth]

Re: Gilbert Noise figure

4dB is Not bad for passive mixer, you can compare it with Hittite or Mini-Circuits mixers.

 

[kapil86]

Re: Gilbert Noise figure

sir, i am talking about active mixer not passive and i want to know the region, why noise figure is more at IF frequency than RF frequency.
Thank you

 

[tony_lth]

Re: Gilbert Noise figure

could you give more details about your test results?
the region is dependent on your requirements, you can refer to other vendors.

 

[RCinFLA]

Re: Gilbert Noise figure

I assume you are feeding input at bottom device. Noise figure will be dominated by the noise figure of that device.

4 db is a good number although you will likely get worse performance with real L.O. source sideband noise.

 

[Mazz]

Re: Gilbert Noise figure

kapil86

let's start from the basics: a mixer is a device that up/downconvert a signal. You should have an input LO signal, and input RF signal and an output that in you case is RF+LO and RF-LO.
I'm assuming that you have RF=2400MHz, LO=2398MHz and IF=2MHz. Could you please confirm?

There is also an unwanted RF signal feedthrough to IF port.

You are saying that RF noise figure at 2400MHz (so calculated at RF leakage) is 4 dB better than the IF noise figure (so, for example, RF NF=6dB, IF NF=10 dB).

If all said is true, you're doing some mistake, as your RF gain should be very low, so you're not interested at all to RF NF.

What is your IF gain? Is it reasonable and comparable with your calculation?

Mazz