Topband
[Top] [All Lists]

Topband: Mixer Noise and SDR radios

To: <Topband@contesting.com>
Subject: Topband: Mixer Noise and SDR radios
From: "Tom Rauch" <w8ji@contesting.com>
Reply-to: Tom Rauch <w8ji@contesting.com>
Date: Tue, 13 Feb 2007 09:10:24 -0500
List-post: <mailto:topband@contesting.com>
I wouldn't want anyone to get the impression mixer noise is 
less or non-existent on a SDR!!

All mixers generate noise, and the typical mixer (once we 
move away from 6BE6 converter tubes) is so far below noise 
floor on 160 the mixer makes virtually no contribution to 
noise floor at all.

In diode or other passive mixers, noise figure is almost 
always established by conversion loss and the noise figure 
of the stage following the mixer. A typical passive diode 
mixer and 5dB NF stage following the mixer makes noise from 
our antennas firmly set the noise floor of the receiving 
system. If not, we only need add a 5dB noise figure amp of 
enough gain and a screwdriver blade antenna would set the 
noise floor at most locations.

We heard reports (or excuses) that one contester who showed 
up regularly in "packet pileups" was using a SD Radio to 
monitor the band, so W8XR brought his Flexradio SDR1000 and 
Firewire sound system down from Ohio. W8XR set up away from 
single operator N5OT (no cheating allowed) and tapped into 
my rearmost receiving antenna. This antenna is 3000 feet 
away from the transmitting antenna field and allows duplex 
operation at frequency spacings as close as a few kilohertz 
with FT-1000 MKV transmitters and receivers.

Our goal was to see if there was any substance to claims a 
pileup could be spotted on an SDR frequency display.

We found the SDR useless in the presence of the transmitter 
despite the Yaesu's and Drakes being very useful. The 
reasons are simple to understand.

1.) There isn't any magic, the SDR is nothing more than a 
mixer with a system noise figure higher than the noise 
figure of our standard receivers. While our other receivers 
made -130 to
-135dBm sensitivity at 250Hz bandwidth without preamp, the 
SDR was about -106dBm without preamp and -135.9 with preamp.

2.) We measured a best average 20kHz IM3 DR with preamp off 
of 85dB. This is actually worse than the Yaesu performance, 
and far worse then the heavily modified Drake receivers.

3.)  We measured a best average 2kHz IM3 DR with preamp on 
of 84.6dB. The IM3 DR was very non-symmetrical above and 
below the test oscillators, indicating distortion in 
multiple stages. The upper IM DR was only 76.2dB, the lower 
was 93.1 dB.  This is ballpark for other communications 
receivers, but not as good as modified conventional systems 
or a stock Ten Tec Orion.

4.) 20kHz blocking dynamic range was 98.7dB, again being 
best with the preamp on. This is because the sensitivity and 
noise figure was the limiting factor without the preamp.

5.) The display had odd artifacts when subjected to very 
strong signals. It generated a large number of strong 
spurious "pips" from our transmitter that weren't really 
there. If we clicked on them there was no signal.

Our conclusion was in a contest environment the SDR at its 
present state of development needs a conventional mixer and 
roofing filter. Since it has no narrow pre-mixer selectivity 
all signals from the BC band up impinge on the mixer and all 
stages following the mixer and rob later stages of dynamic 
range. While a normal receiver only has to process filtered 
signals through multiple stages, the SDR depends on later 
stages processing and sorting hundreds of strong signals.

We found the display totally useless for spotting "packet 
pileups". The receiver was totally destroyed at any spacing 
by our own transmitter while the Yaesu and Drakes were 
largely unaffected at a few kHz or more spacing.

This isn't to say there were some very positive aspects of 
the SDR. As an example on an almost empty band the operator 
could "click and jump" to signals to see if they were worked 
before. No more tuning across empty frequencies. On an 
almost empty range of frequencies it is very easy to spot a 
new signal popping up and check out who it is.  The SDR 
shined when the band wasn't loaded with strong signals, but 
was not very useful on a crowded band. It totally fell apart 
when the local transmitter came on at any signal spacing.

The difference in strong signal density and local noise 
floor almost certainly the key to the differences in 
opinions. Under conditions of low strong signal density and 
high local noise floor, the SDR shines.  Decrease local 
noise floor and pack in the multiple strong signals (even at 
very wide spacing) and the picture changes.

IMO the best combination is exactly what Ten Tec does, a 
narrow roofing filter with a good DSP system.

73 Tom 


_______________________________________________
Topband mailing list
Topband@contesting.com
http://lists.contesting.com/mailman/listinfo/topband

<Prev in Thread] Current Thread [Next in Thread>
  • Topband: Mixer Noise and SDR radios, Tom Rauch <=