[Elecraft] Power Measurement Question

[Don Wilhelm]

Bob,

The answer is - "It all depends ..." -- read on if you want "the rest of 
the story".

I constantly rely on the RF voltage produced across a known accurate 50 
ohm resistor for my power measurement calculations, so that is the 
method I depend on.

The W1 is quite accurate, but there is one "gotcha" - the power display 
is granular.  The 4 watt level is indicated when the 4 LED *just* 
illuminates, and there is no additional indication on the LEDs until you 
get up to 5 watts.
Yes, the W1 software provided by AB3AP's software can indicate the 4 to 
5 watt level in increments down to .01 watt,  but there is no assurance 
of exactness at that level of refinement.  The W1 manual claims the 
normal (manual method) for calibration results in an accuracy of 0.5 dB, 
and I have no reason to disbelieve it - BUT at the 4 watt level,  that 
0.5 dB translates to a potential difference of up to 0.488 watts. - One 
must conclude from that fact that at 4 watts indicated by the W1, the 
error can be almost a half watt.  So be careful about relying on the 
number of significant digits that can be displayed unless your 
calibration technique is adequate to justify the number of digits that 
you see.  Note well that at 4 watts indicated by the W1, the 0.5 dB 
accuracy says that the actual power can be anywhere between 3.52 watts 
and 4.48 watts (note that I dropped the last significant digit from the 
0.488 watt = 0.5 dB value).  If you have not calibrated your W1 to 
greater precision than that 0.5 dB accuracy, you will just have to 
accept that fact.  This is in no way to be construed as a deficiency of 
the W1, for it is more accurate than the majority of wattmeters commonly 
available to the amateur community.

The DL1 accuracy does depend on the exact value of the resistors used 
for the dummy load.   The DL1 states a power measurement accuracy of 
10%, so again, the result is about the same - at the 4 watt level, an 
error of up to 0.4 watts is possible - so if it indicates 4 watts, it is 
somewhere between 3.6 watts and 4.4 watts.

Actually an error of 0.5 dB or 10% is very good in the world of power 
measurement where many (even most) commonly available amateur grade 
wattmeters can only claim 20% of full scale.  That means that on a 100 
watt scale, the error can be as much as 20 watts!  Even the well 
respected Bird meters are speced for 5% (but only after calibration) of 
full scale - and that means with a 100 watt slug, a recently calibrated 
Bird can have an error of 5 watts.  How many Bird wattmeters in the 
hands of amateurs have just come out of a calibration lab and have the 
sticker to prove it?  Both the Elecraft W1 and DL1 will provide better 
accuracy than a recently calibrated Bird wattmeter, but like everything 
else, they have their limitations (based on the specs).  All I am asking 
is that these limitations be understood and that you have some idea 
about what that means.  "Trust no tool until it has proven its worth".

I personally use 1% precision resistors for dummy loads when testing 
power accuracy, and even at that, I will claim no accuracy better than 
5% just due to the granularity in my methods of determining the RF 
voltage across these precision resistors.  All sources of error must be 
taken into account.  RF power measurement is no trivial matter when 
accuracy is desired - zero error is unattainable with any measurement 
method that I know about.

73,
Don W3FPR

k7hbg @dslextreme.com wrote:
> Hi Everybody;
>
>  Don just reminded me to check the math in the DL1 "Manual" and update it if
> necessary. Good Idea!
>  I also got out the W1 and checked my K2 power output Vs the W1 and my
> DVM voltage reading at the DL1 test points. They all read about the
> same *within
> 10% but*
> which reading would be the most accurate? I suspect it would be the DL1 but
> it would
> be nice to know for sure.
>  Ok RF Engineers, all thoughts are welcome!
>
> Best Regards, Bob K7HBG
>   
>