PA stability question

[triple_core]

pa stability

Hi everyone,

I'm a beginner in PA design. In the interstage matching, I use simple high pass to transfer the impedance. After tunning, I find the performance of my PA can be much better. But the output of the interstage matching is higher than the input. I think the output must smaller than the input so I made the stability simulation. I find if I put a serial resistor to the base of third stage HBT, the stability factor is larger than 1. But why the output is larger than the input?

In the following is the waveform of the output of second stage and input of the third stage. Does it oscillate or not? Thanks.

 

[RF-OM]

linc2 circuit design

First of all based on the pictures input voltage swing is about 8V and output is just about 3.6V. So there is no amplification, there is big insertion loss.

Secondly, you cannot measure amplitude with oscilloscope in order to evaluate matching performance because you do not know the impedance in the points of measurement.

Series resistor at the input will help with stabilization, but will increase the noise. Probably (depends on your application) it is better to use resistor at the amplifier output.

 

[triple_core]

fast linc2

Thanks RF-OM.
The two pictures are in false sequence. The upper one(8V) is the output and lower one(3.6V) is input.
second the waveforms are simulated by ADS, not measurement. In ADS, the voltage swing of any wire can be obtained directly, the current swing can be obtained by insert a I_Probe.

how do you mean add resistor at amplifier output? My output of third stage amplifier is 50Ohm. I find from Cripps's book, adding a resistive attenuator of 1 to 3 dB in the interstage matching can stabilize the circuit. After insert this resistor, the K-factor of my output stage becomes larger than 1.

I do not know how to measure the stability factor of the second stage.

 

[RF-OM]

stability factor pa

Okay. I understand about the plots. Resistor at the output can be connected in series or in parallel with the load. Another words series or parallel with collector inductor (for BJT case). Good simulator do it automatically with "Stabilize" option. I did not use ADS for a while and do not remember have ADS this option or not and any way now another version already in use. I usually do it with LINC2 where it is easy and fast. Today I did highly linear CATV LNA design for 50 - 1050 MHz band and all design with analyzes, noise figure optimal match, stabilization and Monte Carlo took about 15 minutes. Finding good active element and corresponding S-parameter file with noise parameters was much longer than actual design. ADS is really balky.

 

[triple_core]

amplifier stability cripps

RF-OM, thank you again for your suggestion. But I'm designing class AB Power Amplifier, LINC2 is not suitable.

In my plot, both voltage and circuit of the output is larger than the input. Do you think it is possible? This is only in- and output of the passive interstage matching network.

I have done the EVM simulation. The result is quite good. If the EVM simulation passed, can I say there is no oscillation?

 

[RF-OM]

ads stability simulation

Sorry, I did not think that it may be such case. You are right, LINC2 is linear simulator and cannot help in this case. Definitely something is wrong if both voltage and current are higher. It means that your passive circuit has power gain! Please check the simulation setup and everything. Sometimes it may be wrong number into models or property lines for elements. Regarding the EVM I cannot tell you something for sure without checking all the design. There may be other reasons for good EVM results.

 

[khouly]

if u can upload ur ADS project or shcematic so we can check the setup

khouly