[SOLVED] Question about Cascaded Amps

[youngguns21]

I am putting together some amps in a chain with a couple of SP2T switches and have a question.

LNA->SP2T -> A1 -> A2 -> SP2T -> out
...........|...........................|
............------->A3 -------->

Each of my amps are MMIC's that were put into their own fixture to check out individually.

The SP2T's are broad band pin diode modules. And everything by themselves look pretty good as far as S11, S21, S22 are concerned.

So why is it that I am getting slopes of 7dB at the output?
And why do the SP2T's make the gain have incredible ripple?
And why do the SP2T's make S22 look a lot more awful then when I look at the pieces 1 at a time?

I'd love some enlightenment.

 

[jiripolivka]

There are two kinds of SPDTs: one is simply a PIN device replacing a mechanical relay contact. When "open", such device presents a high reflective coefficient. Due to such mismatch, it is often not suitable in applications like yours.

The other SPDT switch is "matched", instead of "open" state it introduces a matched 50-Ohm termination. Such device I think is preferable for your application.

Low-noise amplifiers like MMICs have often their S22 not too good, some also S11. Then any imperfect match in connected components like a SPDT switch can cause even an oscillation.

In your case, I would suggest to sacrifice 3-6 dB of gain and use 3-6 dB attenuation pads in SPDT switch lines. I hope this will improve your system response.
There are also MMIC "switched-gain" amplifiers with a structure similar to your circuit, in which MMIC designers solved the mismatch problem on chip.

 

[youngguns21]

For the switches, S11 and S22 VSWR (for insertion loss state) is under 1.3. When the switch is in isolation, I am not concerned. I am testing the amp chain one at a time, so nothing is connected to the switch path that is "off".

 

[tony_lth]

VSWR=1.3 has big difference with 50R load, so when two amp cascade, there should be some ripples. This is standing wave functions.
You can just try to add 10dB PAD between amps, or between switch and amp, you should see better ripples.

 

[youngguns21]

Thanks for all the help.
I put some pads between the amps and added an amp just after the 1st switch. I had no idea that the amps would be that sensitive to a mismatch, even small ones. The pads help out a lot!!!