Help me with stabilizing an LNA

Hi all,

I am designing a LNA. In ADS I get the stability factor 1.004 and the Stability measurement 0.391 but the stability circles (both source and load) include the whole Smith Chart!
The Stability factor and measurement say that LNA is stable, but the stability circles conflict with this.
Under these conditions, can I trust that the amplifier is stable and go for the fabrication?
If no, how can I stabilize this? Should I add a resistor to the circuit? This leads to higher noise figures which is undesirable.
I should mention that the transistor was unstable at first, then I put a TL in source. After that I got these results.
And the final point: if I add a shunt resistor in drain, the whole Smith chart would be in the stable region and the stability factor would be about 3 and the stability measurement about 0.7 then. But as I said, this definitely leads to worse noise figures..

If anyone has any idea, I would be glad to hear that.

Thank you very much,

Cheers

in ADS data display make sure that the smith chart max value is 1 , instaed of autoscale

khouly

Hi

"I am designing a LNA. In ADS I get the stability factor 1.004 and the Stability measurement 0.391 but the stability circles (both source and load) include the whole Smith Chart! "

What is conflicting in that? It means that the whole Smith chart i.e. all passive terminations will result in a stable amplifier!

flyhigh

Thanks khouly, flyhigh:

I tried that scaling factor and as I said before the whole smith chart is in the unstable region. When the magnitude of S11 is less than a unity(in my case), the region within the stable circuits are unstable regions. (in my project, the whole smith chart is in unstable region) but when I look at stability factor and stability measurement, I find it stable.

Am I wrong?

I think when the stability circle contain the centre of unit circle, then their intersection is stable; when the stability circle don't contain the centre of unit circle, then their intersection is not stable!

Hi

"When the magnitude of S11 is less than a unity(in my case), the region within the stable circuits are unstable regions"

I think this is where you are wrong, it is stable inside the circle.

S11>1 indicates the unstable operation (phisically, the reflected power is larger that incident power, this can lead to oscillation)

Usually, 50 Ohm point (center of the Smith chart) is considered an indicator for the stability (if S11<1). If a stability circle contain this point, the apmlifier is stable inside the circle and unstable outside.

flyhigh

Thank you all for your participation in this topic

Dear flyhigh: Your expression of S11>1 and ringing and oscillation is completely right. Also your last expression about the center of the smith chart that is stable for the S11<1 case. But I think here's a misunderstanding about the stability circles. We are talking about the typical stability circles (which you can find in Gonzales and ADS) and as I said before, and as Gonzales's saying, for the case (S11<1) that is very typical for FETs, the stable region is outside the stability circuits.
I have just solve that problem with inserting a shunt resistor at drain. But another question is how to measure the noise figure of the LNA in ADS? There are some parameters such as: NFmin, nf, nf(1), nf(2). The NFmin has a very good approximation with the real minimum noise figure, I heared that. But nf and nf(1) and nf(2) aren't accurate. Has anyone encountered with this problem?