What if you did this with a SMPS ?

[MathWizard]

I was checking the allowable mini-TX power levels, and I read this.

RSS-210 – Licence-exempt Radio Apparatus (All Frequency Bands): Category I Equipment
Annex 2 – Devices Operating in Frequency Bands for Any Application
A2.1 Band 160–190 kHz

Devices using this band shall limit the total input power to the final radio frequency stage to 1.0 watt maximum, and the total length of transmission line, antenna and ground lead (if used) to 15 metres maximum. Example: A coaxial or twin-wire transmission line of L metres long has a wire length of 2L. If a loop antenna of N turns is used with this transmission line, compute the length of wire used by the N turns, and add the result to 2L. The total shall not exceed 30 metres.

Is that so you don't interfere with industrial sites or hospitals ? Or more like the power grid, with such low frequencies, and large lengths ?

So that's maybe 2-3x the frequency of a common computer SMPS. But so what if you instead of driving dozens of amps into a computer, you just put some load/big antenna on it. Would it do anything cool, or just make a pile of noise ? I'm just curious since they have such high power ratings, what would happen, to themselves.

[Kim Christensen]

The relatively short antenna length is just to reduce the range. The shorter the range, the more people in a given area can use the same frequency and not bother each other.  (Wave length at 160Khz is 1.8 Km so a 30m antenna is not optimal)

A computer SMPS produces a PWM squarewave, so yea, connecting an antenna to one of the transformer windings inside would spew a bunch of noise. That's why those power supplies have filters on the power connections: To prevent unintended RFI.

[T3sl4co1l]

Wiring to an SMPS secondary (before the rectification and filtering) would be only slightly better than a spark gap transmitter...

Depends on type.  All of them will have sharp waveforms (read: tons of harmonics).  Your antenna coupling network will require a strong lowpass or bandpass filter to deal with these.  None of them are crystal controlled, so you won't be able to keep a, say, 500Hz CW channel (or whatever they're allocated as, at frequencies down there) tuned in.  And many of them aren't fixed at all, where the switching frequency depends on load and component values rather than a fixed (e.g. RC timer) reference -- and wildly so, far more than you'll be able to hold a, say, NBFM (couple kHz) channel (if such are permitted in the band at all, which I'm guessing not..?).

So, in very, very broad strokes, yes, it's sort of how a high-efficiency (class D) transmitter works... But it's missing 110% all the subtle things to make an actual transmitter: there's no coupling network, it's not filtered, it's not modulated, it can't stay in its channel... take your pick!

Ed: incidentally, SMPS are over that frequency nowadays.  Probably the cheaper and older models are still under 150kHz (FCC Part 15 lower threshold, so they only have to worry about levels of harmonics), but the newer and more compact types are pushing 300kHz, even low MHz with GaN FETs.

Tim