Thanks :)
Lars

Franklin, we need those other specifications to be able to recommend a model.
frequency band, acceptable loss in each path, acceptable isolation between output ports, power handling.

I have a native French RF-engineer as colleague, I'm sure he can give it a try.

I think lowpass filters are the most economical way to handle this. You have to analyze the variety of DUTs and see that you get a filter for each flavour, taking in account not only that the lower freq is within the passband byt also that the higher freq gets attenuated enough.
There are tunable filters, quite expensive things. A bandpassfilter covering the low freq of all DUTs could be possible.
Another approach could be a diplexer if one is to be found that suit all types of DUTs. This would give the advantage of possibility to proper terminate the unwanted signal into 50 ohm, like this one: https://www.markimicrowave.com/Assets/datasheets/DPX-0508.pdf

A limiter will add loss to the path when input signal is high enough. So yes, both signals will be attenuated. If no suitable filter is at hand, you can McGyver one;
make a 1/4 stub that kills the 8 GHz signal, A SMA T and a short piece of semirigid to be cut and grinded to suitable length, You need a network analyzer to get it right, or at least a spectrum analyzer + white noise source.

Some loss will occur at 4 GHz also, so the stub filter needs to be characterized. Also, your test object might respond poorly to total reflection of the 8 GHz signal, pad with 10 dB in front of the filter. Is it for measurements or for usage ? Lowpass filters are reasonable priced at mini-circuits.

I guess your shorter cable, in combination with height above ground plane and LISN termination is a more or less perfect radiator for 120 MHz. How is the shielding of the cable terminated at the DUT's end and at the LISN ? Is the shield open at the LISN, the shield itself may become a resonator.
And, as Tom Fagan wrote, a pigtail grounding of the shield may act as a good injection point, especially if connected to signal GND on a PCB inside DUT.
The shielding shall be a extension of the faradays cage that encloses the DUT, connected to the shell of the shielded connector.

I find it often a lot easier to filter the RF away within the DUT than try to prevent the cable from radiating.
120 MHz is not that hard to kill with a CM + DM filter. Got a current probe ? Make a breakout cable and measure the RF current in the cables at the DUTs end, measure DM and CM and try to figure out what filter type you need.

Thanks :) found a HP K920B on Ebay for reasonable money

Thanks Glyn :)
It makes sense that both E and H component stays within the same boundary, and if the silver is thick enough it should be fine.
Yes, I found this document covering the potential problem with too thin gold over nickel:
https://www.simberian.com/AppNotes/ModelingConductorLoss_2016_01.pdf
I guess it also applies to waveguides and cavities for mw.

There is a paper with an interesting title. but sadly it's locked in at researchgate
Sadly. much good research ends up there, unvaliable for most people:
https://www.researchgate.net/publication/286480769_Dielectric_Properties_of_Polymer_at_Microwave_Frequencies

24 GHz microstrip bandpass filter on output is mentioned here:
https://electronics.stackexchange.com/questions/440004/has-anyone-designed-a-low-noise-amplifier-in-the-ism-band-24-ghz-using-the-dis

This slideshow illuminates printed circuit board properties in a PIM context:
https://www.slideshare.net/manfredhuschka/the-influence-of-rf-substrate-materials-on-passive-intermodulation-pim

Jos, I've never seen a written source with all needed information cooked down.
As this doc linked below describes, the way of working is, follow guidelines created by eariler mistakes, and hope for the best ;)
The theroy is easy; avoid all parts and connections that can behave nonlinear at high RF power. High contact pressure, silver against silver, no magnetic materials, no oxides and so on.
Some parts are easily understandable, like that red copper oxide has semiconducting properties and was actually used in recifiers. The "rusty bolt effect" is also well known in medium-wave transmitter installations. But there seems to be other, not so good known mechanisms as well, more subtile.

The doc is in english, but the file name is in swedish.
It is published by the Swedish space authority, and the file name is ( translated): characterization of passiv intermodulation in space communication antennas.

https://www.rymdstyrelsen.se/contentassets/de067a79466749efa22b953340e47293/10.-karakterisering-av-passiv-intermodulation-i-rymdkommunikationsantenner.pdf

OK, Roader, take a look at this: https://www.microwaves101.com/encyclopedias/directional-couplers
A directional coupler, think of it in this way; think of a road, going north-south, with traffic in both directions. That road is the main path of the coupler, the signal source is in the north and the load in in south.
Then, add a coupled path; a measurement device that registrate the traffic only in north-south direction. Thats your forward coupled path. You can also add a coupled path registrating the traffic south - north. That would be the reverse coupled path.

Roader, the network analyzer har directional couplers on the ports, making it possible for the analyzer to separate forward and reflected power when measuring s11 and s22.. For a scalar analyzer (that only can measure magnitude); for instance, if forward power out on port 1 is 0 dBm, and the reflected wave is -10 dBm, the return loss can then be calculated to 10 dB, or, in linear fashion, gamma is 0.1.
For a network analyzer, also the phase between the stimuli ( forward power) and the recieved (reflected power) can be measured. Then, not only the magnitud can be measured but the reflection in complex form, as magnitud + angle, or real+im component.
The same goes of course for transmission measurements, s21 and 12, but that is easier to understand.
There are a lot of good tutorials on network analyzers, explaining there way of working.
With some math skills in complex numbers the secrets will be revealed for you in no time :)