Hi, Well firstly, the radio seems to be working and you appear to be getting near full power on 40 with the correct ALC showing. Pretty sure with some further careful alignment including L09 you should get 100 watts. Did you check all the other bands for full power too? Everything should be tested including RX for preselector tracking to TX. The works.
So all that being true then you could put the case lids back on and walk away!! What i am saying is, maybe you are looking too hard?, or did you want to take the thread into an educational direction rather than a repair.
Ok, some comments, rambling on, just generally putting it out there..
Your SA looks correct, but we already did this test quite some posts ago. Its basically showing 7Mhz fundamental, 2nd and 3rd harmonic. Thats because its before the finals which then filters further and results in a near pure 7. So thats correct.
The ALC is really easy. Some RF from the final grid bias is first decoupled to remove the -V, then rectified to create a negative control voltage related to the amplitude of the RF at the grid. Thats it. So as RF increases, the -V control line goes more negative. That signal is then fed to the NB/Proc board to reduce carrier level as needed. The whole thing forms a negative feedback loop to control the PA finals which keeps them in the linear range avoiding distortion and harmonics. So as you talk, the SSB speech envelope whos peak power is much higher than you see on the PO meter is the signal used by the ALC. The very idea of ALC is to remove final distortion.
I'm pretty sure you already know all that, we covered this very early on. It seems to be working as it should.
As you talk, with a low mic gain setting, the ALC should not kick back and stay at full scale. There is a separate preset somewhere to set full scale (FSD). As you increase the mic gain to higher levels and talk louder, the ALC should just start to kick back. At that point the audio mic gain is sufficient to activate some ALC. Go further and really drive it hard and the ALC will really kick back as it tries its best to keep the finals limited (clean). Without any ALC, the finals would distort and clip - a major problem for transmitters, its against the law to transmit like that and cause all sorts of problems with TVI etc.. ALC is our friend.
As said, L09 was simply out of tune. L09 forms part of a tank circuit in series with the preselect variable inductor and the whole thing in parallel with the capacitor trimmer board. Classed as a 'rejector' circuit or tuned tank. Its called a rejector because it rejects current at its resonant point. If there's no current at resonance, then there is high impedance and therefor high voltage across the tank. This is exactly how a crystal set radio works. Its just a rejector circuit then rectified by a diode. When tuned, the voltage across it is maximum. All other frequencies are minimum.
Series tuned circuits are called 'acceptor' circuits because they accept current when tuned.
Your L09 / preselect coil / Trimmer board combo is exactly that. A big, band select-able tank circuit. So when tuned, the current through it is minimum = impedance is high = voltage across it is maximum. in other words - peaked.
ok, Since L09 was NOT tuned correctly (the actual fault) current was flowing in the tank at 7Mhz and that leads to some pretty strange effects. Its not a resistive current path, its reactive. Things go out of phase and it goes complex. Some tuned circuits react differently to others especially where there are resistors involved (effecting Q) or other reactive components like L and C.. Remember that its not just a LC tank, there are other components in the circuit right back to the finals (remember pulling the finals had an effect?). What i suggested earlier on was that because of the mis-alignment of L09, you actually got a rise in 7Mhz amplitude but because of the complex phase nature of the tank, it was similar to a bad SWR where much of the power becomes reflected back. That then caused the problem with the ALC level.
I think if you REALLY want to see what happened in detail, you'd need to enter in the exact values from the circuit into something like a SPICE modelling simulator to get the result. So there are some assumptions here but on the whole, i'm pretty confident its correct.
EDIT: or set up a real tracking generator and sweep through the tuning range while viewing your SA and adjust L09.
Sorry that was a long post.