Well, I was thinking it'll be a faulty input attenuator because that's the first thing 'in the line of fire', but it seems that I have a blown First Mixer.
The 1000MHz IF output is only -80dBm with a LO input of 1054MHz @ 10dBm and RF input of 54MHz @ -40dBm (from the front panel Cal output that I've confirmed as good).
The LO is behaving OK and ranges from 10 to 14dBm over the 1000-2000MHz tuning range. You can set the Dispersion control to 'CW' which is like a zero span mode, which turns off the LO sweep.
I haven't been able to identify the mixer component, but the manual says 'the first mixer is of the double balanced type, using four matched Schottky-barrier diodes configured as a ring modulator'.
Which means it's something like this;
If it is, I'd expect to see the schottkys with a multimeter diode test in either polarity between IF and RF, and IF and LO. However this one is just showing a short between all inputs and output, but this could be by design depending on the internals. I've lifted the resistors from the LO and RF pins so there are no external paths to ground.
So I'll need to identify the mixer to confirm the fault, and then find a replacement, either the same component or a like-for-like with similar drive level, conversion loss and frequency ranges. LO input range is 1000-2000MHz, RF input is up to 1000MHz, IF output is 1000MHz, drive level is around 10dBm, and that's about all I can determine for now. Any replacement will need to be a close match to the original.
Maybe someone can recognize the mixer component from the attached photos?
It's looks similar to this one from Pulsar, even the pinouts are the same (but not the specs);
https://www.pulsarmicrowave.com/product/mixer/X2L-10-301