Hi Erik
I have a few comments/suggestions.
I do not know how you test setup was, but there is a need to ensure the IIP3 level is not influenced by cross modulation between the two signal generators. The power combiner from the two generator ought to have as high isolation as possible between the two input ports, so the test signal from one signal source is not feed directly to the ?second signal source. Next it is common practice that the level from the two signal source are identical and thus the two sidebands are more equal, too to facility the calculation as half power of side tone + the test signal level.
I used some Mini Circuit power combiners with isolation and I also did build a few as described by N2PK Paul for his N2PK VNA, going down to fairly low frequencies.

I used the reflection bridge PCB seen om the link ?? and I did purchase these separately as an add on although not listed as a separate item. where I used the SMD type of T1-T6 transformers. Inspiration came from

In addition a video from National Instrument is quite good to watch ??

In this video you get a method how to figure out if it is the Signal Generators are reducing the IIP3 measurements or not, by adding attenuation in front of the SA. You have enough dynamic range to use some 10dB attenuation or maybe more in front of the SA.

Kind regards

Kurt

?

?

Kurt,
Can you confirm I'm doing the IIP3 measurement as it should be done?