[Amps] crossmodulation in PA ?

[Tom Rauch]

> At our small contest station LN8W / LA8W we have a strange 
> problem..
> It appears that we have crossmodulation taking place in 
> the power amplifier.
> When running 1 KW to a 20m yagi at 90 feet, I detect cross 
> modulation
> products from BC-stations appearing on a range of 
> frequencies. I have
> excluded
> unwanted mixing in poor connections in the 
> antenna,feedline or the immediate
> surroundings like tower and nearby antennas or metal 
> constructions. Field
> measure-

Peter,

I have considerable experience with this both at HF and VHF.

First the amplifiers are linear amps, and as such make very 
poor mixers. Think of the PA like an AM transmitter plate 
modulated by any "audio" coming back up the feedline. In 
this case though the "audio" is really RF, so the sidebands 
are many MHz away.  Class C or other hard switched PA's have 
a fairly good conversion efficiency, but for the same 
reasons we cannot plate modulate a linear stage a good 
linear amplifier does not make a good mixer.

Second, while connectors and cables CAN cause a problem in 
exceptionally clean carefully built sites (like Cell sites) 
or extremely high power sites like TV or FM transmitters 
where a 100 kW carrier might beat against another 100kW 
signal in the SAME feedline and antenna, you are only mixing 
millivolt signals with a kW or so. It is beyond reason to 
think anything but a horrible joint would cause mixing and 
re-radiation of a SWBC signal of a few hundred millivolts or 
less.

Now there is some possibility the diodes in a directional 
coupler or power meter could cause problems, but it isn't 
very likely. The loose coupling required to prevent damaging 
the diodes goes double for signals returned to the line, so 
a .1 volt SWBC signal would suffer the same attenuation 
passing twice through the coupling plus the mixing 
attenuation or loss in the diodes. It would be very easy to 
sort this out as a problem by varying transmitter power. If 
you varied transmitter power 10dB and saw very little change 
in spurious signals it would imply the mixing could be in a 
detetector diode. The reason the level wouldn't change much 
is the diodes are in and out of conduction  If you varied 
the power ten dB and saw a large change, something close to 
or well in excess of 10 dB, it would imply an external 
non-linear device.

I'd try a pad on the receiver ahead of any diodes or other 
obviously non-linear devices. If you add a 10dB pad in front 
of the receiver (or spectrum analyzer) and the mixing drops 
significantly more than 10 dB, then look at the receiver. 
Most of the time it is there. The most common cause of a 
problem like you have is the receiver (and spectrum analyzer 
since it is a receiver), although it can be in other devices 
around the site.

73 Tom