Spectrum Analyzer weird results from LNA

Hi.I have designed an LNA for S-band between 2.2-2.4 GHz bands. The thing is when i test it i see this weird stuff on Spectrum analyzer like some 50 MHz signal is getting mixed in.I couldnt really work out a reason. can you give me a clue why it might be doing this ? You can see my marker is at 2.2 GHz but it doesn't give the wanted amplification. If it worked it would have amplified to about -20 dBm.

It probably oscillates..

It seems that Signal Analyzer shows between 2 and 2.4Ghz and your marker is @2.2GHz.

Your signal is shifted to 2.3GHz; that is the difference between simulation and measurement. You may have to include tolerances for components in your design.

Your LNA takes 200mA from a 12V power supply (2.4W). Like a power amplifier.
Even if is a multi-stage LNA (looks like it has 4 stages) this DC current at 12V seems too much.
In case that I don't misinterpret the posted picture, very rarely an RF system needs an LNA with 4 stages, which can be pretty unstable if is not properly shielded, decoupled, biased, etc.

Last stage is actually a bfp780 a dirivng amplifier to increase P1db. Also I checked the oscillations again and it shows oscillations 864 mhz and it's multiples.First two stages is from macom maal 011078.And the 3rd stage is bfp 640esd. Also power supply supplies two LDO evaluation boards from TI TPS47A01 http://www.ti.com/lit/ds/symlink/tps7a47.pdf. First two stages are fed with 3V 50 mA each. According to my AWR simulations 3rd stage uses 33.6mA and last stage must use 84.1 mA( I used non-linear models from infineon's own AWR library). So I actually might now be getting enough current to last two stages.

1- Do u have a way of being sure they were matched properly? I don't see any mmbx or ufl connectors.

2- Have you done load-pull analysis?

It sure does look like one of the stages is oscillating at 600 MHz. Stick you thumb all over the circuit, and see if it magically goes away.

I assume you tested the signal generator all by itself and did not see any +/- 600 MHz sidebands, right?

This can easily happen if the bias lines have odd resonances.....like the capacitor to ground terminal having too high an inductance

Yes sometimes when i stick my thumb or multimeter over the 3rd stage it goes away. But solved the problem by replacing the capacitors at rf out side with other values i had, now works fine. Thank you biff44.

well maybe. Depending on the LOAD the amplifier sees in your application, the oscillation might come back. Keep your eye on it.