Bill, (et al)
As you may know, RFI is specified as either conducted (over wires, etc.) or radiated (via electromagnetic waves). Either way, you suffered an RFI problem albeit slightly different than the norm associated with line/load variations (mentioned earlier) as I have witnessed over the years with various products having the design aspects of the products like that which you have.
I must point out that Bill's product problem has an inherent design difficulty very difficult to overcome vastly different from the design difficulty difficult to overcome inherent in the very different topology of the solor
regulator. Both generate RFI yet both have very different sources of radiation.
Regardless, do not assume that ANY product sold in the US adheres to ANY standards regarding emission or conduction of RFI. Somewhat sadly it is incumbent upon the purchaser to verify the claimed adherence to various standards, including the "minimum" of U/L
safety ones (which do nothing to force manufacturers to minimize RF radiation).
For at least thirty years there has been an FCC class "B" standard which, in some industrial US applications must be adhered to, YET does not quite qualify to pass the somewhat equivalent "CE" standards which MUST be adhered to in order to be legally accepted into the European markets. In general the "CE" standards are more stringent. Yes, there are various CE standards and one must be astude as to which one(s) affect/effect products and applications.
Mastervolt is one European manufacturer attempting to make
marketing inroads into this country with relatively good product quality that meets such standards. One obvious "problem" is that such products require more engineering and production costs than those not so "quiet" in both the "RF" world as well as presenting the
power grid with a unit representing a resistive load (technically referred to as a load not generating unacceptable harmonic components to the generator). There are others, like Xantrex (in some specific cases). So far, I know of no products affordable as consumer solar regualators which provide necessary peak power point
tracking. Perhaps others will point out a viable source for this application area.
There is a cost, of course. Instead of paying around 8 dollars per 10 Amps of
charging current one will pay around 20 dollars per 10 Amps for a CE rated charger. But, guess what? You probably will not have an RFI problem, although not necessarily so, especially in the case of HF receivers which are particularly sensitive to a "noise floor" caused by ANY existence of switching
noise in an immediate area.
The good news is that although for several years various circuit topologies have been identified as being ones which lend themselves to relatively "low" generation of RFI it has only been in the last 5 years or so that the cost of implementing those topologies has come down, largely due to the lower cost of solid-state switching, inductor, and
offshore manufacturing (like it or not). In addition, there has been what I call a "fall-out" of designs from industry which make their way into the marine market to give us a benefit in this regard. I have observed for decades that the marine
electronics products lags the state of the art industry products by 12 to 20 years. That lag is diminishing for several reasons. One of the main reasons has historically been that the marine market is relatively miniscule compared to the world industrial market for manufacturers, including their ability to attract and retain really good engineers. Maybe I'm getting too far off the original topic, I apologise.