[Amps] IMD and spurious measurements

[Manfred Mornhinweg]

Hi Gary, Karl-Arne, Steve, Bill, Ron, Peter, etc,

As promised, here is the follow-up of my strange spurs measurements.

Indeed it would be quite strange to see DX broadcast signals, that come in at 
about 1 millivolt on my antenna, be mixed by a transmitter and amplified to a 
few volts. And those were in fact the actual levels! So I checked each part of 
the setup.

Mixer saturation was not involved. The dynamic range of a diode mixer with 
strong local oscillator drive, followed by a 16 bit computer sound card, is 
close to 80dB, and it's limited by the sound card rather than the mixer. While 
doing the measurements I was staying 15 to 20dB below the saturation point. 
Further attenuating the input to my mixer, so that the main signals were 40dB 
clear of saturation, didn't change the ratio between the spurs and the main 
signals at all.

I didn't have a directional coupler, so I built one. Taking my test signal off 
the coax cable, via the coupler, instead of capturing it with a tiny antenna, 
made a major difference: The strange spurs were pushed down nearly to the noise 
floor, with only the stronger ones being visible. So it's clear that I was 
getting who-knows-what noise into my mixer, via the small pickup antenna.

I then returned to a small pickup antenna, but a balanced one: A small loop, 
connected through a common mode choke to the mixer. This produced almost the 
same results as using the directional coupler. So I'm pretty sure that when 
using the monopole pickup antenna, I'm getting noise into the mixer because the 
other pole of the antenna is the computer wiring.

Anyway I couldn't make sense of the spurs, and I don't know where they come 
from. Their frequency relationships don't seem related to line frequency, power 
supply switching frequencies, nor any other I can recognize. And I live in the 
woods, far away from any neighbors, let alone cities, so there aren't any strong 
local signals here, except for those I generate myself.

Another interesting note about the directional coupler: It didn't matter whether 
I took my sample from the forward or the reflected signal port. Other than the 
amplitude of the whole thing, nothing changed. The ratio between spurs and main 
signals remained the same. This probably means that the spurs are not caused by 
anything from the outside world mixing in my mixer. And it also means that a 
simple nondirectional pickoff from the coax line should be as good as the 
directional coupler, for this kind of test.

Next I inserted a 1:1 audio transformer between my mixer and the computer sound 
input. The spurs got slightly smaller, about 1 to 2dB further down from the main 
tones. The improvement isn't a lot, but since there was some improvement, I left 
the transformer in place.

Next I inserted a 1:1 RF transformer, wound on a small ferrite toroid, between 
the DDS signal generator and the mixer. The DDS is powered and controlled 
through an USB interface, so it does have a potential for coupling lots of 
noise. At this point my mixer was galvanically connected to the coax line only, 
with both the oscillator input and audio output galvanically insulated. It seems 
to have improved the situation another little bit, but only marginally, so it 
looks like USB noise getting through the DDS wasn't much of a problem.

I can't conveniently run my computer from batteries to break any ground loops, 
because it's a desktop computer.

At this point the strongest of the strange spurs were about 65 to 70dB below the 
main tones, close to the dynamic range limit of my spectrum analysis setup, so I 
won't worry any more about them!

I can now conveniently measure the IMD of my transceiver and amplifier, with a 
valid dynamic range of greater than 60dB. That should be enough for all ham 
purposes.

About Steve's funny way to tell me to get on the air, well, trampling on others 
isn't my style! :-)  But seriously, I tend to enhoy more playing with radios and 
electronics, than talking to "typical" hams. I do sometimes engage in 
long-winded rag chewing, if any interesting matter comes up, and I happen to 
find someone interesting to talk with. But that's preciously rare these days! 
Most QSOs revolve around the weather, and what brand and model of radio the guy 
bought. When it does get technical, it's at the level of which of his 25 mikes 
sounds brighter, duller, more or less boxy, and so on. Those tests might be 
interesting for mike collectors, but they tend to bore me to the point of 
signing off.

And DXing? That's the dullest thing I know in ham radio! Invariably most DXers 
just want to exchange callsign, a fixed 5-9 report, maybe tell me the name and 
location, and then it's "thanks for a new prefix, bye bye!" And after three such 
contacts, at most, somebody puts me in a DX cluster, and the next thing that 
happens is a pile-up of guys calling, crying, shouting, without even making sure 
they can actually hear me. So I quietly slip away.

Compared to that, it's still far more fun to spend my time measuring the IMD and 
  efficiency of an amplifier at different bias settings, locking my transceiver 
to GPS, or playing any other of the games that involve a soldering iron and some 
instruments, rather than a mike.

Manfred


========================
Visit my hobby homepage!
http://ludens.cl
========================