[Elecraft] K3 Observations Digital Modes and SSB

[Kok Chen]

On Feb 2, 2011, at 2/2    9:47 AM, goldtr8 at charter.net wrote:

> What is a good IMD report number and good SN number from a  
> receiveing station and what are the numbers that state you are  
> messing up the bands.

The transmit IMD that you see bandied about are 3rd order IMD numbers.

If you take a look at a PSK31 signal that is idling (i.e., sending the  
Idle Varicode), you should see two carriers that are space 31.25 Hz  
apart.

Third order IMD will therefore create an extra pair of sidebands.  One  
that is 31.25 Hz away from the lower carrier and one that is 31.25 Hz  
away from the upper carrier.

Like CW and RTTY keyclicks, these are not useful and only uses more  
spectrum that you need to pass information.

The reason the idling PSK31 signal looks like two carriers idle signal  
is that it in not just phase modulated.  There is also a cosinusoidal  
amplitude component.  I had attempted to describe it a little when I  
wrote the PSK31 section in Chapter 16 of the 2010 ARRL Handbook.

(For old timers, think "double sideband suppressed carrier" signals.   
There too, you have the same thing -- two carriers on the RF spectrum  
when you transmit a single audio tone DSB signal.)

The existence of 3rd order IMD will just take up more spectrum space,  
and if they are loud enough, will bother everybody trying to copy the  
signal that is next to yours.

If you hard clip a PSK31 signal (as in using severe amounts of ALC),  
the cosinusoidal waveshape becomes a square wave.  Severe clipping  
will cause around -11 to -13 dB IMD.

But wait!  If you do PSK31 for long enough, you probably have seen  
signals with even worse IMD.

Careful look at those signals will show that they were not just  
clipping but there are overshoots in the square wave envelopes!  I  
have seen IMD as bad as -9 dBc.

I would prefer that the station transmits with 10 times more power but  
staying inside a PSK31 passband, than spreading their energy over the  
entire PSK sub-band!

If the SNR of a station is much worse than 25 dB, and his IMD is -25  
dBc, you will not visually see any extra "rails" on his signal.   
However, with sufficient integration, you can measure the IMD of a  
stable signal even below its SNR.  In practice however, I don't trust  
those measurements.  So, if someone tells you that they can barely  
copy you, I won't trust their IMD report.  Very often, they are just  
reporting the noise amplitude where the IMD component would have  
been.  Sometimes, it is their receiving chain that is producing the IMD!

Personally, I consider -25 dBc IMD to be barely acceptable.  It is  
something like transmitting an extra 1/2 watt signal on the sides of a  
100 watt signal.

Most stations that are properly adjusted will run between -25 dBc and  
-35 dBc IMD.  You can find some stations with obviously experienced  
PSK31 operators who do much better than that.  They are often both  
loud and clean.

My own K3 will not do any better than -35 dB transmit IMD.

In fact, this K3 produces around -30 dBc IMD if you use PSK-D at 5  
watts; so you are much better off using DATA-A from a good software  
program.  -30 dBc IMD is starting to fall into the "mediocre" range.

With my FT-1000MP at 5 watts, I have gotten -40 dBc IMD reports.  But  
you need to be loud at the receiving end for them to properly measure  
IMD that is that low -- the few times I got numbers that high were  
from a local station and with KH6 -- we have a pipeline from Portland  
to KH6 :-).   A signal has to be really loud to reliably measure -40  
dBc IMD.

> I am trying to understand the best way to set up my radio, and when  
> I look at the tables in the link, it tells me that keeping ACL at a  
> maximum of 4 with maybe an ocassional ticke to 5 was the key to a  
> clean signal with a K3 and output power did not matter.

The IMD numbers may be the same, but remember that you are referencing  
the IMD number to the carrier power.

Transmitting a -20 dBc IMD signal will produce a 1 watt image when you  
run 100 watts, while it produces a 0.1 watt image when you run 10 watts.

The reason IMD is important is that a 100 watt station with -40 dBc  
IMD will produce an extra intermodulation image that is only as strong  
as a 10 watt station with -30 dBc IMD.

The other way of looking at it is that a station running 100 watts  
with -40 dBc IMD uses up no more spectrum (no matter how far in  
terrestrial distance you are from him) than if he reduced his power to  
10 watts but his IMD has degraded to -30 dBc.

73
Chen, W7AY