simulate the lna

Dear all,

I have designed a negative feedback LNA with 18dB power gain and 3dB NF. At the same time, I have constructed the gilbert cell down conversion mixer with -6dB power gain and 9dB NF. As I simulated LNA+mixer, it finds that the simulated NF is much higher than the calculation . If I removed the feedback path in the LNA and do the LNA+mixer simulation again, the simulated NF value is quite near to the hand calculation. Therefore, how could I fix this problem I have tried to add a common source buffer between the LNA and mixer. It still could not solve.

Thanks

wccheng

when u remove the feedback , what is the gain of the mixer alone

and also check , when u make a feedback the BW will be higher , but when u use no feedback the BW will be lower , this also will affect the noise


and what is the simulator do u use ?

khouly

Dear khouly,

I am using Cadence SpectreRF simulator to do the simulation. The mixer gain doesn't vary so much before and after feedback path is removed in LNA.

Thanks

wccheng

and when u simulate the LNA , and mixer with the feedback , what is the gain is it 12 db this mean that LNA give the 18 , and the mixer is -6 , or the LNA gain is droped

khouly

Dear khouly,

The simulated gain of LNA(with feedback)+mixer is similar to the summation of standalone LNA(with feedback) gain and mixer. However, the NF is large different between the simulation and hand calculation.

thanks

wccheng

wccheng
in spectre you have a very useful tool that shows you the noise figure contributors (noise summary)
Using this infos you can better understand the main noise contributors in the two cases and so understand root causes to correct them.

You can also probe noise at LNA output with mixer load and LO running and see if there noise figure is higher than 3 dB.

As you are talking about power gain, are you sure about the right input/output impedances of both circuits and test benches used for NF simulations?

I hope it can help.
Mazz

Dear Mazz,

Very thanks for your helping. Could you tell me how could I set in order to allow me to see the noise contribution of each elements?

thanks

wccheng

You don't have to set anything.
Just do the simulation and then go to menu print results-> noise summary. Write down in the form the info you want to be printed and see the results.
Mazz

Dear Mazz,

Thanks for your information. Another thing I want to ask that the k-factor simulation. I know how could I get the k-factor simulated value. But I could not find out where could I get the delta value? Could anyone tell me how to get delta value?

Thanks

wccheng

well, another cause for the difference between your hand calculations and the simulation results may be the output impedance of your lna which is lowered by the feedback. if you remove the feedback, both your mixer input and lna output have considerably large impedances and your calculations become then again valid. i suppose you use the Friis equation to calculate the overall noise figure, don't you?

wccheng
what do you mean with delta value?

Mazz

When u do the whole simulation, the output of the lna will be much smaller than 50Ohm, which is used to do the mixer alone noise simu. from Lee's book, we know the source impedance will affect NF.