can we stabilize this transistor?

[ayhz2002]

I would like to know how to stabilize it without using resistors.
the graph shows stability circles for load and source. The circles going from left are for source and from right are for load.


ayhz2002

 

[flatulent]

keep load and source

Keep the load and source impedance at all frequencies in the middle and bottom right area of the chart (capacitive reactance) This can probably be done with a series capacitor and a shunt capacitor in the circuits for source and load. The best would be to have a 50 ohm source and load at all frequencies.

Can you post the S parameter file also?

 

[ayhz2002]

this is the file

 

[flatulent]

excessive gain at low frequencies

I have put the s parameter file in Ad Lab Plus and played with series and shunt feedback. There is no way to stabilize the transistor at lower frequencies because the gain is so high. Unless you can use resistors in your design, such as in shunt with the input and output, you are stuck with figuring out an input and output network that stays in the center and bottom right end of the smith chart.

Is this a school project where you cannot change anything or a project where you can change the transistor bias point or even use another transistor?

The gain going up so much at low frequencies is a very poor situation with microwave transistors which is why they are made with low gains at low frequencies.

 

[ayhz2002]

thanks for the work you did.
It is a school project. I will talk to the professor about this issue. The best matching network I used gave me instability between 100MHZ and 400 MHZ, and other regions were stable. The problem I have to make sure it is stable in the entire frequency band (100 Mhz to 6 Ghz). I have used resistors just to test the stability, and I ended up overstabilizing the transistor where I have lost the entire transistor gain (S21 was almost 0.08).

ayhz2002

 

[BigBoss]

Dont't you think to apply a ' serial or parallel ' feedback to the transistor ?
Generally this technique reduces S21 and provides more stability..

Or ' balanced configuration' might be a solution ..another way to..

Good Luck

 

[flatulent]

i did them

The ad lab plus has series and parallel feedback. I tried many combinations and could not get unconditional stability at all frequencies at the same time.


Putting 20 ohms in series with the input cured the problem, but the problem statement was that no resistors are allowed. This reduced the gain by about 2 dB.

The only way without resistors is to have an input and output network that presents the transistor with impedances that are in the stable region at all frequencies.

 

[ayhz2002]

I just talked to the professor and he told me to design first at 5.5 GHz and then put the matching network and simulate again and see if this helps. If not, I have to keep iterating untill I find the right values for reflection coefficients.

I tried it already and did not work, but I will keep trying anyway.

 

[flatulent]

network topology

Keep in mind that the transistor needs to see a capacitive reactance at low frequencies. Using a high pass network topology with a capacitor on the transistor side will be the most likely network to work.