[Elecraft] K3 or SDR-5000?

[Joe Subich, W4TV]

> According to the current FLEX website, they are now claiming 
> a 14 MHz IMD3 at 2 kHz spacing of >100 dB using ARRL measurement 
> methods. This puts it 6 dB (or more) better than the K3 at 2 kHz 
> spacing, assuming identical measurement methods. I think it's 
> arguable whether even a 6 dB difference in IMD3 is meaningful 
> when both are well above 90 dB, but there it is.  Comments on 
> this thought would be interesting.

I seriously doubt that Flex can achieve IMD3 performance > 100 dB
in any except a carefully controlled TWO SIGNAL case.  With any 
receiver using DSP and a broadband front end all signals in the 
front end filter are applied to the analog to digital converter 
at the same time.  In order to avoid IMD the ADC must be able to 
handle (and properly digitize) the absolute peak voltage of all 
signals present.  That peak is more than the scalar sum of the 
average power of each signal - it is the vector sum of the peak 
power (peak voltage) of all signals including the instantaneous 
peak voltage of things like static crashes/etc.  

While the Flex-5000 may be able to handle two strong signals in 
isolation (and with their architecture 2 KHz or 100 KHz does not 
make a difference), what happens in the real world on 80 meters 
or 160 meters when the band is full of S9+20 dB "local" signals 
plus static crashes during a contest?  I suspect the presence of 
the multiple signals will eat up the dynamic range in a hurry.  

The ARRL IMD testing methodology is seriously flawed when testing 
receivers with wideband front ends because it assumes 100% of 
the voltage that causes IMD comes from two "clean" and well isolated
sources.  In the real world when the "window" through which the 
signals that generate IMD can be viewed will accommodate more than 
two signals, ALL signals will contribute to the IMD.  Understand 
that the two signals in an IMD3 performance test do not need to be 
the same strength - it is the combined peak voltage that drives 
the stages being tested (preamplifier, first mixer, post mixer 
amplifier, IF stages and DSP ADC) into non-linearity.  It does 
not matter if that non-linearity is due to shifting the bias in 
a mixer or overflowing the ADC - both effects will show up as 
IMD.    

With an increasing number of DSP based receivers, we need a test 
(rating) system that attempt to measure the "total (peak) receive 
signal" necessary to generate a IMD products which exceed the 
receiver noise floor.  This is different than either two tone IMD3 
measurements at 2/5/10/20/100 KHz or a blocking measurement and is 
the only way to absolutely compare receivers with different design 
philosophies and different IMD "windows." 

73, 

   ... Joe, W4TV 
   

 
receive power" and front end bandwidth.