1 kW RF power LDMOS destruction test by bad SWR

On a sunny day (Thu, 18 Sep 2014 09:07:54 -0700 (PDT)) it happened Lasse Langwadt Christensen wrote in :

triodes..

Bottom up!

I was curious about that too. At the end of the day it can't be anything other than a line stretcher with a short at the far end. Maybe some sort of roller inductor but with a bifilar helix instead of a single winding?

-- john, KE5FX

Yeah, so the brass box doesn't need to handle 1200W dissipation.

And of course, if you reflect the power in phase, no one cares, the load as seen by the transistor simply opens up. I do it all day, every day: it's called a switching power supply. Reactive loads are perfectly fine, as long as you have the diodes to cycle that energy back into the supply.

One would perfectly reasonably expect the supply current to rise and fall with phase, which is the observed effect. Whether the transistor is oscillating or doing other untoward things is impossible to tell (it's also undefined what the brass box's response is at other frequencies, especially those the test circuit may be prone to oscillation at).

A resonant tank would seem to do a good job. To go all the way around, it would have to have multiple resonances (not so good..), or be something like a variable transmission line as John suggested. With a sufficiently high Q, either option should do a good job spinning circles. (A transmission line, of course -- if long enough -- can go around an arbitrary number of times, while the lumped circuit can only go around N times for 2N total poles. Or something like that.)

Tim

Well, all I know about it is that it's running at a single test frequency of 225MHz and that it has to handle 1200 watts, which is about 250 Vrms into 50 ohms. A roller inductor with a sliding contact will produce enough arcing to destroy any plating in short order.

Basically, the box is providing a 0 to 360 degree phase shift without any sliding contacts, which also eliminates the slotted line idea. I tried Googling for various mutations of "variable RF phase shifter" and found that most high power mechanical phase shifters are just sliding coaxial lines. For example: However, at 225 Mhz, to go through 360 degrees of phase shift, the line would need to move 1 electrical wavelength about 0.9 meters. I don't see that happening inside the mystery brass box, especially with a single know adjustment.

Another possibility is a series resonant tuned circuit, with the knob adjusting the variable capacitor. This can go through +/-90 degrees as it goes from inductive to capacitive. Couple (somehow) two or more of those in series and we have +/- 180 degrees. The odd location of the brass box adjustment knob at the corner might indicate a gearbox or pulley system to drive all the caps.

As for dissipation, the reactive devices do not dissipate any heat. However, both the capacitors and resistors have resistive components (as part of the Q) which will produce some heat. The rest of the power is reflected back to the RF output stage.

On a sunny day (Thu, 18 Sep 2014 17:16:58 -0400) it happened "Tom Miller" wrote in :

Right, first with DC generators, then with 3 phase AC, all against the 'hum'. This is why I pointed out to J.L. "Transistors any time".