[Elecraft] Dummy Load recommendation

[CUTTER DAVID]

Don
Thanks for those further notes.  I'm also thinking of the best we can do with the most primitive equipment.  Amateurs have been making RF probes forever and now we have cheap digital multimeters with >1Mohm input impedance it should be even easier to get a good estimate of power using the method you outline but without a scope, which itself has to be expensively calibrated and you still have to interpret an analogue display.  
I wonder just how good we could make and calibrate an RF probe. 
David G3UNA

> On 31 July 2019 at 16:28 Don Wilhelm <donwilh at embarqmail.com> wrote:
> 
> 
> David,
> 
> Your question about "what is good enough" is an informative one.
> When I was doing repair/alignment work for others, it was important to 
> me to have tools that had traceable calibration info.  I had some good 
> 50 ohm dummy loads that were good to 1 GHz, and an LP-100A that was 
> calibrated to NIST standards (which I used as a sanity check for other 
> power measurement instruments).  Having those calibrated tools were 
> important to maintain my customer's trust.
> 
> In amateur practice, such care is not necessarily needed, but the 
> K3/KX2/KX3 TX gain calibration does require a good 50 ohm load up to 50 
> MHz.  That is easily checked by sweeping it with an antenna analyzer.
> 
> As for the power measurement, once you have a good 50 ohm dummy load, it 
> is easy to use a 'scope (with a 10X probe) connected across that dummy 
> load (use a TEE adapter to gain access to the coax center conductor). 
> Then you can measure the RF voltage (peak to peak shown on the 'scope), 
> and use the formula Vp-p squared then divided by 400 to quickly obtain 
> an accurate calculation of the power.  Nothing really expensive except 
> for the 'scope (which can be borrowed in many cases).
> 
> So what you need first is a good 50 ohm dummy load, and secondly, access 
> to a 'scope having reasonable calibration for the vertical deflection at 
> the frequency of interest.
> 
> With that combination, you can check the error in other in-line power 
> measurement equipment.
> 
> In most cases a 10% or 20% error is not going to make a big difference 
> in on-the-air performance - anywhere in the ballpark is usually OK.
> 
> You are correct that we have become complacent because of digital 
> displays.  Take a 4 digit display in an instrument having accuracy of 
> 10% - the difference between a display of 10.41 and 9.37 or 11.45 is 
> meaningless given the 10% accuracy of the instrument itself.  You should 
> round that reading to 10 volts.  Analog readouts automatically do that 
> compensation because you cannot read the meter scale with 4 digit 
> accuracy.  Even my 12 inch slide rule in college was limited to 3 
> significant digits.
> 
> There is a difference between doing it "good enough" and with greater 
> accuracy depends on the purposes of your measurements.
> 
> 73,
> Don W3FPR
> 
> On 7/23/2019 11:38 AM, CUTTER DAVID via Elecraft wrote:
> > I agree with all that has been said and suspect a lot of responders on this list are or were engineers who are used to using precision equipment.  I think it can be misleading to look at some instruments built into a transceiver for instance, but it never seemed to matter, until Elecraft came along.  I did my calibration with a Bird power meter and a Bird 500W load but I have no idea when they were last calibrated, so, what value do I put on the measurements?  Well, it's good enough for me. But calibration is a special case and I'd never heard of amateurs calibrating for power before Elecraft.  Accurate power meters are costly to buy and maintain if you are in business, but for me now I'm retired it doesn't matter so much because I'm not doing it for a paying customer.  I think we are misled by the easy precision of the digital multimeter which for a few £ or $ is ten times as good as what we previously thought were good analogue instruments.  In short, I think technology can cause us to bite our finger nails unnecessarily.  10% or 20% power error is hardly noticeable to the guy on the other side of the world.