Please edit and leave in parts of the message you are referring to in your reply. In my case
when I dont see trailing message content I dont bother to read or reply as I cant be bothered
to figure it out who anyone is talking to.
I'm not used to this mailing list stuff, I tried to treat it like reddit, sorry.
Back to the topic: I feel so stupid. The Spurs *do* change when I tune.
On 20m, the 3rd harmonic of the VFO (59 MHz * 3 = 177 MHz) is mixed with the 4th harmonic of the 1st IF (45 MHz * 4 = 180 MHz), resulting in spurs 3 MHz (180 - 177 = 3) left and right of the signal when I tune to 14.0 MHz. When I tune to 14.05 MHz, the spurs are 2.85 MHz from the carrier, when I tune to 14.10 MHz they are 2.7 MHz away, and so on.
On 15m, the 2nd harmonic of the VFO (66 MHz *2 = 132 MHz) is mixed with the 3rd harmonic of the 1st IF (45 MHz * 3 = 135 MHz), resulting in spurs 3 MHz (135 - 132 = 3) left and right of the signal when I tune to 21.0 MHz. When I tune to 21.05 Mhz, the spurs are 2.90 MHz from the carrier, when I tune to 21.1 MHz they're at 2.80 MHz, and so on.?
These mixing products are generated in the final mixer. This also explains why the products aren't there in CW mode, since the IF isn't used at all in CW mode, and the frequency is directly generated by the VFO. Maybe add a note to the block diagram, currently it says the VFO tunes from 45-75 MHz, which only is the case in SSB mode.
I modded the PA to RD16HHF1s following DF3FY's instructions, checked the output afterwards, but except for the known harmonics issues it was clean. I replaced the LPF relays with name brand parts to fix that, checked the harmonics afterwards (now at -45dB, nice), and this weird signal still wasn't there. Then I took everything apart to put the bitx into a metal case, and this signal appeared, now I can't get it to go away.
The problem is that I don't remember if I tested everything using SSB mode with a 1khz signal from my computer (tune button in WSJT-X is convenient for this, also for finding the sweet spot for NF gain and so on), or if I got lazy at some point and used CW mode, or at which stage I used which.
I don't know how the unmodified design prevents these signals from getting strong enough to matter, do the IRF510s just not have enough gain at these high frequencies? BPFs at the output would completely remove them, of course.
On 15m the signal was weak enough to be within legal limits (around -45dB), but still very noticeable. On 10m, I saw some spurs at similar distance to the carrier, but they were at like -60dB, so nothing to worry about at all.
DD2TH
