Why these two decoupling capacitors de-sens the RX?

Hi, All,
I designed a GSM module with NXP PA.
But if decoupling capacitors, C733 and C707, are mounted.
The RX is desensed more than 7dB.
Without these two capacitors, the rx sens is -106dBm.
But with these, the rx sens iw worse than -100dBm.
If changing C733/C707 to 100nF and 100pF, the rx sens is also desensed.
Why?
Thanks.
Best,
Tony Liu

it might be a software error (you can not assume these analysis programs do not have glitches and convergence problems!). Try putting a 0.1 ohm resistor in series with the 4.7 uF capacitors, and see if the problem suddenly "goes away".

there also it a possiblity that that series FET is oscillating with the 4.7 uF caps there....put a voltage probe on the FET and see what it is doing. look at low frequencies.

If the circuit is working fine without those decoupling capacitors, I think is fine. You have a lot of other decoupling capacitors on the RF PA supply path.
Most probably there is a PCB layout issue, with poor ground design.

in fact it's no so difficult to check this.If you set the SA with "zero span", you can see the eventual oscillation that may come from supply circuit without modulation.This will-even a small amplitude fluctuation- disturb the BER in other words the sensitivity.But I do personally not think that is a decoupling capacitor related problem on the supply path because there are many caps on that path.
However a MOS related oscillation cannot be avoided by these capacitors.

The fact that a couple bypass capacitors are having an impact despite far removed from the RF circuitry should be alarming. My guess is that there is some interaction with components not shown on the schematic.

How are you measuring sensitivity?

In any RF circuit (and not only RF) if you have poor ground layout design and if you place randomly decoupling capacitors, for sure there will be situations where the circuit will produce unwanted spurs.
So if you are in that situation, a quick fix (luckily could be a permanent fix) is to remove the decoupling capacitors placed in the wrong place.

Most probably a self oscillation. Check with the oscilloscope probe.
Also, during the time the tip of the probe touches the PCB, check for sensitivity.
Sometimes it happens that the spurius components of the probe stop oscillation, creating confusion about the result.



Most probably a self oscillation. Check with the oscilloscope probe.
Also, during the time the tip of the probe touches the PCB, check for sensitivity.
Sometimes it happens that the spurius components of the probe stop oscillation, creating confusion about the result.