Microstrip implementation of feedback amplifier
时间：04-07
整理：3721RD
点击：
hey there! i am designing a 2-stage feedback low noise amplifer for 100Mhz-6Ghz with negative RC feedback on every stage and LC ladder in-/output matching. main specifications: 20-24 dB gain, NF<3dB, S11,S22<-10 dB. the design with ideal lupled elements is complete and shows promising results of 24 dB gain, NF<3dB and S11,S22<-12dB throughout the whole band maintaining stability up to 30 Ghz.
Now it's time to convert the design to reality using microstrip line interconnections. I intend to use the Rogers RO5880 substrate (h=20mils, Er=2.2) for it's very low tanD=0.0009.....AND AM LOST!
Although (I think) I am aware of the common rules of microstrip line implementation, the microstrip interconnections mess up EVERYTHING! I do not want to replace the lumped elements with microstrip lines, just make microstrip line connections between them instead of the ideal wires in ADS!
THE FEEDBACK PATH IS THE MAIN PROBLEM! IT destroys the characteristics for good....now i've read somewhere it must not be to long, or it will cause oscillation...can somebody please explain why?....I need it to be at least 15mm long to get from drain, around the source grounding of the transistor, to the gate...also source grounding with vias introduces unwanted inductive parasitics and causes oscillation in the upper band...is there any way to prevent that?
any help on these problems would be greatly appreciated, so please help if you know a way or two...thanks very much in advance!
Now it's time to convert the design to reality using microstrip line interconnections. I intend to use the Rogers RO5880 substrate (h=20mils, Er=2.2) for it's very low tanD=0.0009.....AND AM LOST!
Although (I think) I am aware of the common rules of microstrip line implementation, the microstrip interconnections mess up EVERYTHING! I do not want to replace the lumped elements with microstrip lines, just make microstrip line connections between them instead of the ideal wires in ADS!
THE FEEDBACK PATH IS THE MAIN PROBLEM! IT destroys the characteristics for good....now i've read somewhere it must not be to long, or it will cause oscillation...can somebody please explain why?....I need it to be at least 15mm long to get from drain, around the source grounding of the transistor, to the gate...also source grounding with vias introduces unwanted inductive parasitics and causes oscillation in the upper band...is there any way to prevent that?
any help on these problems would be greatly appreciated, so please help if you know a way or two...thanks very much in advance!