Unwanted Tone - DC Power Woes

[MicMaven]

I would expect a radio receiver to have really good power filtering, though.

Agreed !... And yet, this is not the radio's speaker - it is a smallish, plastic box speaker accessory sold long ago by Radio Shack. Shoot, it even has a microphone input to use as a ... oh boy, perish the thought ... are you sitting down? . . . a 1 watt public address speaker! Yeah ... sure. But considering he likes how it sounds and it works for him, I plod along as a good trooper with a positive outlook.

You could partially eliminate other devices from that problem by powering the speaking from the buck converter, and checking to see if the noise is present with nothing else connected to the speaker, just the power.

Yes. I believe we tried that with poor results on the main supply, but not the separate 12 v supply I took with me as an ersatz test reference. I will, however, try to double check that when I can. We may have overlooked a material combination.

Then plug headphones into the receiver and see if the noise is present on the receiver output with nothing else connected, just the buck converter powering the speaking sharing the same main power supply.

Audio from the rig was always clear - before and after connecting the speaker.

If the noise is present in the speaker with nothing else connected, then the noise is from the power supply into the small amplifier, typically you would address that with additional bypass capacitance.

OK - that sounds like what Mr. @ruffrecords recommends - and I have ordered a capacitor to test it. Do you agree a 1000 uF capacitor is a good start?

Antenna input doesn't seem very likely unless the receiver is tuned to a frequency which is a harmonic of, or close to a harmonic of 150kHz.

Right. We did all we could to minimize RF interference - either to the speaker or from the speaker to the rig, both with an antenna and on a dummy load.

If the noise is not present in the speaker alone, nor in the headphone output of the receiver, you can start making combinations to see when the noise shows up:
Very tedious I know, but if you want to find exactly where the noise is being injected you have to be systematic and look at all the combinations.

Right. We tried as many combinations as we could think of, including running the amplified speaker on other power supplies and mixing an matching other inputs in the system. Fortunately, there are relatively few interconnects in this mash up. (See PDF layout image in OP.) I will, however, revisit and try your full list when I return, to assure we left no stone unturned. It is possible we missed one or more combinations. Unfortunately, the logistics preclude my popping over whenever I like to test stuff. It may be next week before I can revisit his shack. Darn it.

Likely the most straightforward approach would be to just waste the power of a linear regulator with a large enough heatsink to handle dropping the 5-ish volts from the main supply down to 9V.

Funny you mention using a linear regulator. That mirrors Mr. @Brian Roth's suggestion, and I have ordered parts to test that possibility. As requires a heat sink, I contemplate mounting the regulator chip to the inside of a metal project box as the heat sink and to shield it from stray RFI/EMI in the shack. Nothing like re-engineering a solution as ya go along, eh?

Yeah! Great stuff. Thanks for the detailed reply! James