BJT Gilbert Cell with Emitter-degeneration

[tomrett]

Hello forum,

im trying to design a Gilbert cell mixer and im stuck on one requirement.

Background: the RF differential voltage swing should above 900mVpp (currently RF=991.2mVpp). This swing must remain the same. IFm and IFp are equal to 20mVpp while the IF signal frequency is set to 1MHz with a DC voltage level of 1.6V. LOp is equal to a DC voltage of 2.7V while LOm is equal to a DC voltage of 2.4V. The power supply voltage VCC is equal to 3.3V.

Im now interested to achieve the following requirement: achieve linear operation for IF differential input voltage swing of 800mVpp. I know i have to use emitter-degeneration, but im not sure how do achieve this, while keeping the RF differential voltage swing above 900mVpp.

What i have tried so far: I tried using a resistor (84.5 Ohm) at the Emitter to increase the voltage drop, but this somehow resulted in a lower voltage drop at IFm_d (i can upload a photo of my derivation, if required).

Has anyone any suggestions? Im fairly new to this topic and i would be happy to receive any helpful input on this.
Thanks!

 

[BradtheRad]

In schematics of a simple bjt amplifier often a capacitor is installed in the emitter leg, besides an emitter resistor. The purpose is to allow DC operation of the transistor, while giving high gain to the AC component. I believe this is often done in rf circuits though I don't know what is the custom in Gilbert cells.

 

[D.A.(Tony)Stewart]

Have you looked at commercial IC's.
Falstad has this ideal model. https://tinyurl.com/22xn4mju
You can resize the plot pane and change as you like then save or export.

 

[aliarifat794]

Are you simulating the circuit?

 

[D.A.(Tony)Stewart]

The operation of your circuit has differential input signals across Vbe pairs which is very nonlinear.
Whereas the circuit I showed has 4 independent matched current sinks with lots of headroom for the current sources to allow 4Vpp on either input.

The output has a differential gain of 4 to max rail to rail 29Vpp output using CMOS R2R.
It might not be what you need though.