[Amps] MOSFET amp filtering - was: auto-tune

[Catherine James]

Jim Brown wrote:
> Cathy James wrote:

>> I hear a lot of recommendations here for antennas that are flat across the band and very close to SWR of 1:1 to keep solid-state amps happy.  That is completely unrealistic on the low bands.
>> I have to re-tune my 160 meter dipole for even small excursions of a few tens of kHz.  A cage dipole would be broader-banded, but that is difficult to build and install, especially for a 270 foot long antenna that hangs in the trees.  I certainly couldn't put one up in the treetops where my current wire dipole lives, as it would get all tangled in the branches.
 
> Excellent post, Catherine. You have hit the nail squarely on the head.  However -- even a dipole at 100 ft is "low" for 160M, so it's both 
 inefficient and radiates more at high angles than low.

It all depends what you are trying to do.  For solid regional coverage, a low dipole gives excellent NVIS coverage.  For DX, it is less than ideal.
 
> A FAR better choice for something suspended between trees is a Tee vertical, where a flat top wire provides top loading for the vertical 
 section, which does the radiation. Because antenna current splits equally left and right into the top wires, radiation from the top cancels, and you end up with a nice vertical radiator and a nice low angle of radiation.

That's great for DX, but lousy for regional coverage.  And here in northern New England, our ground conductivity is awful, which makes verticals challenging.  No matter how much you improve the near field with an excellent radial system, the far field will still have lousy conductivity and that will push the radiation angle higher.

> AND -- there's another point that virtually EVERYONE who has commented in this thread seem ignorant of.  ANY distortion mechanism produces BOTH harmonics and INTERMOD. On SSB, it shows up as splatter, and on CW it  shows up as clicks.

I'm still in the process of learning about intermod and how to minimize it .

> the rise and fall of ANY rectangular wave consists of an infinite number of harmonics that excite IM.

Yes, in Fourier analysis terms, the definition of a square wave is essentially the sum of a sine wave and all of its odd harmonics.
 
> Yes, we can filter the harmonics, but we can't filter the IM.

That's why I'm trying to understand better how good design suppresses IMD. It's easy to say "you need a linear response", but real world systems will never be perfectly linear, and there's often a serious tradeoff between linearity and efficiency.
 
>> 10 doesn't appear to have opened at all for the contest last weekend, at least not here in New England.
 
> That depends entirely on when you were in the shack. I worked 44 states with 1500W and a 3-el SteppIR. The only states I missed were DE (N3DXX was active, but not in the few hours the double-hop sporadic-E path was open between us), WY, MO, IA, NE, SD, and ND. I also missed the VE provinces east of VE3 and NT. I chose to only work  CW.  I also worked 20 countries, but no EU, AS, or AF. DX was double-hop to the Caribbean and trans-equatorial to SA and OC.
 
Clearly you did much better than the New England ops I've spoke to!  I'm not sure how much of that is regional (e.g., how long did New England have good paths to anywhere outside of NE?) vs. time of day.

> I now need only VT and SC to finish it QRP. I've been at that for four years.

Well, I'm in Vermont, so we can try some time with a sked.  My RF noise is pretty low here, so you may get lucky!

73,
Cathy
N5WVR