Here we are, after some days. Ludwig, you were right, too much current was drained, around 1.3A when TX at 12 V (different values when Vin=8V, but same result, too much current drained). I start saying the problem was solved. I'm going to explain how.
1: current at 12 V was around 1.3A for those bands;
2: solder joints checked, like all the others of the rig; board cleaned and all, no visible shorts between pads (with microscope).
3: Q507 was working correctly (it broke after I had solved, anyway)
4: power was always being monitored on external wattmeter, only on 40 m around 1 W (less than diagnostics, but not much difference);
5: T501 and T507 desoldered, checked, scraped again, resoldered.
Problem was still there. Everything before PA was ok, so I started looking on the LPF section. Almost fortunately I noticed that on the diodes D513 and D518 (QMX+) there was the same wave as output when transmitting on 40m. That was a contradiction because the voltage must be 0 V on the diode of the LPF referred to the band in use, which in my case was the 40m, so LPF2. Instead, I could read 0 V also on D518 that is on LPF3, which should have been to around 11 V, not 0V. Also, on that diode I could see the sine wave at 7 MHz which should have never been passing through that filter. At that point, I understood that the problem was there and started working with my friends Alfredo IK4POF and Walter IK5ZWU. We started from mosfets Q511 and Q517. Q517 was in short. I removed it and measured Q517 which was NOT FAULTY as initially thought: the worst had yet to come. At that point we decided to exclude LPF3 so D513 and D518 were removed and the radio tested: working perfectly on all the bands except 20/17/15, the ones related to the LPF3, obviously. It was more than clear: LPF3 section had a short or something like that, between the two diodes. Checked soldering problems, checked capacitors, axial inductors, toroids: every single component was OK when removed from the PCB, so the problem was on PCB and it was around the joint D517-L515-C531. Indeed we found UNEXPECTED 180 ohm to ground path and this kept the two diodes conducting: this resulted that the filter LPF3 ?was parallel connected to the LPF2 which I had selected (because I was on 40m). We tried to to figure the reason out and thought it could have been a kind of a problem on the internal layers, maybe due to my mistake on soldering. I don't know exactly. Anyway, we decided to bypass the faulty joint and make an "aerial link". So all done and resoldered, it has started working nice and with much relief from my side (and Alfredo and Walter's too, who helped me in veeeery nice remote meetings :-D ). Such a great experience but we were successful, finally - and I learnt a lot. Power output reports after playing a little bit with the inductors:
160m: 5.5 W
80m: 4 W
60m: 4 W
40m: 3.8 W
30m: 5.5 W
20m: 4 W
17m: 2.8 W
15m: 4 W
12m: 2.5 W
11m: 4 W
10m: 3 W
6m: 2.5 W
Thanks everybody again, I hope sharing will be helpful for similar situations. Ciao!
