RF Noise Bridge for antenna tuning?

An early electronics project for me was to construct an RF noise bridge. This remarkably simple device, combined with a receiver, let me measure resistive and reactive impedance by adjusting for a null in the noise on the receiver.

I'm putting together a compact station for portable operation and it occurred to me that a noise bridge might be a good way to adjust an antenna tuner used to match an end fed wire in a tree. Sometimes it's hard to find the resonant point.

I tried this with a random wire, a Z match and a QMX transceiver (with the recent SSB software). It was not a good way to tune up, partly because of the QMX's excellent AGC.

To see what's going on I hooked up an SDR as the receiver, so I can see the noise over a broad bandwidth:

The antenna is an off centre fed dipole and the tuner is a simple L match. The dip can be seen above up the top end of 40m after tuning. A traditional antenna analyser confirms the tuning:

So it does work but is difficult compared to other options such as minimising an LED on the Z-match or using a NanoVNA. 

I wonder if I'm not doing this correctly. I'm not seeing what I would characterise as a "deep null".

The Palomar bridge is a nice bit of gear and I rather like the warning on the back of the manual:

They have an article about using the noise bridge for antenna tuning. Oh, and this story is another example of Betteridge's law of headlines.