Question about specification of isolation for switch

I'm a bit confused about isolation specification for switch.
A simple sketch below is a 3-port switch (SP2T).
RF0 is signal input port.
RF1 and RF2 are output ports and switched through control logic.

---> RF1
RF0
---> RF2

If the isolation RF0 --> RF2 and the isolation RF1 --> RF2 are both specified as 30dB when it throws to RF1 path (Assume RF0 --> RF1 path is lossless),
what the RF2 port power would be when a 30dBm CW signal inputs to RF0?

Thanks a lot.

Any switch is specified by two parameters: insertion loss when the contact is closed, and isolation when the contact is open.

To your question: if FR0-RF2 isolation is specified as 30 dB, then applying 30 dBm to any of those ports will result in 0 dBm in another if the contact between them is open. If you close the contact, the 30 dBm signal will appear as 30 dBm (or ~29 dBm due to insertion loss) on the other port.

To improve isolation, you can use two such switches cascaded (then it will be ~60 dB), or, use a "terminated" switch in which the contact when open between two connector ports, will be contacting a termination (50 Ohms typ.), so it will not be radiating through open-contact capacitance.

Also, when selecting a switch, be sure it can handle a high power like 30 dBm. Small-size switches and semiconductor ones (PIN or FET) are designed for a lower power than 30 dBm.

Thanks.
I'm confused by the spec of the isolation RF1 --> RF2 = 30dB.
To me it looks the RF2 port will be coupled by the power present at the RF1 port too.
So when the power inputs to the RFC is 30dBm and the contact between RFC & RF1 is close, we will see the RF1 port = 30dBm and RF2 port = 0dBm (assume insertion loss is 0dB).
Will the power 30dBm at RF1 port couple to RF2 port too (isolation RF1 --> RF2 = 30dB)?
So the RF2 port will be copuled by both RFC and RF1; hence the power at RF2 port will be the combination of 0dBm (coupled by RFC) and 0dBm (coupled by RF1) = 3dBm ?

Thanks.

Isolation simply tells you how much of a reduction to the input power you see on the unconnected port, it's not really coupled... more so, it's a leakage phenomenon. Inside a switch, you have elements that may physically disconnect from one output port, and then connect to the other output port. There are still materials (plastic, fiberglass, etc) that touch both output 1 and output 2, so you can never keep all of the input signal from sneaking out the "disconnected" output port. Some small amount of energy will always leak around in a real part.

Assume the isolation from RF1 to RF2 is 30 dB, and the insertion loss from RFC to both ports is 0.1 dB (when each is selected). If you input a signal at +30 dBm into RF1, then +29.9 dBm will come out of the common port, RFC. Then, 0 dBm will come out of the disconnected port, RF2.

For this part, you have two flavors of isolation: 1) RFC to the disconnected port, and 2) connected port to disconnected port (RF1 to RF2). The spec you gave is the latter, RF1 to RF2. The spec on the part may call out both... some do, some don't.

BTW, you can't have a situation where the insertion loss from RFC to RF1/2 is 0 dB, and have anything but infinite isolation. Power in must equal power out. If you have no loss on your thru path, you need to magically create energy to put some of it out of the disconnected port, so you could measure an isolation value. If you have 0.1 dB of insertion loss with 1W input, that's 22.76 mW lost somewhere in the switch. With 30 dB of isolation, the power out the disconnected port is 1 mW, so 21.76 mW is left to heat (or is reflected at the input due to imperfect return loss, etc). Be careful to track little things like this... oversimplifications will almost always come back to bite you when you least expect it.

Very important for some switches (especially the FET type) is the termination of the second path when first path try to meet spec for isolation or insertion loss.