Resistor for high power amplifier design

[jayce3390]

Hi,

I am designing a 30W Rf hybrid microstrip power amplifier in S frequency band.

I encounter stability problem that's why I need to implement serie resistor on the RF path to improve the power amplifier stability.

I am looking for a 10-20 ohms chip resistor able to support 1 to 4W power handling range.

Can somebody give me manufacturers names?

Thank you very much.

 

[khouly]

this resistor will decrease the gain of ur amplifier.
for resistor manufacturers check uniohm.

i think u need to check the matching for stability

Khouly

 

[jayce3390]

Thank you for your answer.

Yes of course the resistor makes the gain & PAE worse but unfortunately I have to add this resistor in the matching network to make the Rollet stability factor upper than 1.

I tried to modify my input matching network in order to improve the global stability but I don't find solution.

:|

 

[khouly]

u need to study the source of instabilty , u need to plot load and source stability circiles and check which one can be modfied ,
even u can mismatch the load to get ur amplifier stable

khouly

 

[VSWR]

It's a waste of power to add this resistor in the RF path to make the amplifier stable. In many cases there may be another solution. WinRAR and upload your project/simulation files here so we can take a look.

I encounter stability problem that's why I need to implement serie resistor on the RF path to improve the power amplifier stability.

I am looking for a 10-20 ohms chip resistor able to support 1 to 4W power handling range.

 

[khouly]

i agree with VSWR

Khouly

 

[RF-OM]

I agree with VSWR. Usually this resistor is below 5 Ohm. Do you have S-parameter file for your transistor?

 

[jayce3390]

RF-OM,

Yes I have S parameters files of my transistor from 0.5 GHz to 10GHz.

You can find it in attachment.

VSWR & Khouly, I will upload my project as soon as possible.

thank you very much

 

[biff44]

http://www.usmicrowaves.com/res/power_beo/custom_beo_microwave_power_chip_resistors.htm

 

[RF-OM]

I am on the business trip now. When return back to home I will check stability analysis for you.

 

[RF-OM]

I check your S2P file today for stability analysis with LINC2 circle utilities. It looks like your amplifier should be stable in 2-4 GHz range (S-band), but be unstable on higher frequencies closer to 10 GHz. Please see attached jpg file with this simulation. May be you can use filter to limit the bandwidth?

 

[jayce3390]

Hi,

Thank you for your help RF-OM, My device seems to be stable at 2 GHz, but We can't be sure that it will be stable at 2.1GHz for example?

I thought that the the device should be stable at all frequencies, not only in the frequency band... is it true?

thx

 

[RF-OM]

Of course it is true. But often you may artificially cut off the bandwidth in order to eliminate unstable region and make your circuit stable. There are different ways to do it. For example stabs at the input or output, some kind of filtering and so on. If your transistor is unstable and you do not want to change the bandwidth or cannot do this, you need to chose another transistor.

 

[BigBoss]

Frequency selective negative feedback will improve the overall stability.Adding an extra pole with RC or RL circuits on the frequency where the stability decreases, this will help you to pull up the stability factor more than one..
Note that, the stability circles tell me that amplifier is NOT unconditionally stable..

Take care. !!

 

[RF-OM]

It is absolutely right. I checked stability circles and analysis clearly shows the problems. But without knowing all design details it is hard to give exact recommendations. There is not one frequency problem, it is rather relatively broadband and frequencies are high enough to make it complicated to use lamp elements. RC circuits will almost for sure make things worse due to parasitics. I think that there must be solution that limits the bandwidth for at least 5 or 6 GHz, definitely not more (if I remember the stability analysis right). Layout may also contribute a lot, especially parasitic inductance in emitter or source connection. Without seeing the design it is hard to make right conclusion. Probably these details were included in file provided by the author of the question, but I was not able to open it and check.

 

[jayce3390]

This is not a wideband application, only 1,8% around ~2 GHz, I added a serie resistor to the RF path of the input matching network that improved the Rollet stability factor but makes the PAE decreasing by 10 pts.

I think it is not a good idea to add resistive element at the output matching in terms of PAE, I think it is worse.
Is it possible to use shunt resistor instead of series ?



BigBoss you advise me to use a circuit between the output and the input pads of my device, RC series for example?

 

[jayce3390]

I can't find the source of unstability in my design. The used device shows K factor <1 full analysis band.

Is it possible to design with this device?