I've built two power meters with the AD8307 and AD8310 (in this long thread). They work well but were built and tested with low power signals which didn't cause any problems. Power levels up to only 10 mW. The previous one meant for ham bands only would read higher powers - but that was tested after it was put inside a shielded case. The use of that with higher powers meant that it needed an external attenuator attached in-line outside the box.
This project is being tested with high power emitters within a foot of the project electronics. The AD8307 is extremely sensitive and even though the power is fed to the coupler and from the coupled outputs to the boards with coax, the boards are not shielded and the input to the AD8307 can pick up RF easily due to the unshielded input traces.
I added several quick & dirty RF shields, and ferrite chokes. They are lame at the moment but I need to prove that shielding is the next problem. I put the HT inside some metal window screen, and grounded that. I made a small enclosure from cardboard, and put metal tape around it. I put the two AD8307 boards inside it, and grounded it. I put two ferrite chokes on the wiring to the AD8307 boards, and one choke on the power supply wiring.
Next I removed the 50 ohm load from the output of the Putnam, connected a 40 dB attenuator, and connected my 437B to it. Now when I transmit my development system and the 437B will measure power at the same time. This will prove to me that they are or are not in agreement.
I've measured the Baofeng HT many times using my hp 437B power meter. It's not a labratory-grade RF source, and so is probably going to vary a few tenths of a watt.
The results are as follows:
hp 437B vs development system (watts)
VHF low power
1.6 vs 1.7
VHF high power
4.5 vs 4.3
UHF low power
1.3 vs 1.3
UHF high power
3.6 vs 3.7
So now I'm confident that the cal constants are good, the attenuation values for the ports are good, and the system can read correct powers. But the shielding is not good enough because I tried another transceiver - a Yaesu FTM-7250D - which can transmit up to ~50 W. I don't have any shielding that I can put the radio inside. So any power I used higher than the low power setting, 25 or 50 W, caused improper readings on the development system (too high), whereas at the same time the hp 437B gave correct power numbers. The same thing happens when using the Baofeng HT on UHF if I do not enclose the little radio in the window screening.
So I think I want to work on a better temporary shield box for the development system. I do not want to commit to buying any new case, an Arduino Due for faster processing, or further programming work regarding a larger OLED display with on-screen graphics for the user, until I am sure it can reliably measure power from a more powerful transceiver.