Posted by
Don Wilhelm-3 on
Jul 25, 2006; 6:03am
URL: http://elecraft.85.s1.nabble.com/Funny-power-measurements-tp392406p392411.html
I did fail to point out that by doing these measurements, you will only be
verifying that the K2 firmware is capable of controlling the power to within
10% by its ALC control algorithms.
As long as the power is kept below the maximunm that the transmit chain will
produce, the Power Control ALC will keep the output level within those 10%
bounds, and in the process compensate for any weak points in the transmit
chain (such as a mistuned bandpass filter). I believe the K2 is unique in
the way it controls the power output, and the method used will compensate
for gain variations through the transmit chain as long as it is operating
below the maximum gain limits.
So Fran, your measurements will really only be testing the ability of the
firmware to track the requested power setting. If your goal is to test the
gain variation through a band (i.e. checking the bandpass filter bandwidth),
you will have to either operate the K2 at max power or disable the Power
control ALC and use a fixed amount of BFO injection for your test.
73,
Don W3FPR
> -----Original Message-----
>
> Fran,
>
> You are correct in leaving the KAT2 out of the picture for that
> measurement - unless the KAT2 is in a bypass state. The addition of extra
> tuning elements (the tuner L and C values) can create a tuning
> peak that is
> unaccounted for in the relationship that you are trying to
> verify. Too many
> variables cause un-needed confusion.
>
> If you have a 10X probe for the 'scope you can make accurate RF Voltage
> measurements. Check the probe and 'scope calibration first, most 'scopes
> have a built-in calibrator that can allow you to check that a nice clean
> square wave is displayed (if not adjust the compensating cap in
> the probe),
> and also check that the vertical deflection is correct based on
> the internal
> calibrator.
>
> Apply the 10X probe at the dummy load location rather than at the signal
> generator (K2) end for the better accuracy.
>
> Be aware of the frequencies that you will be testing and the specified
> frequency limits of your 'scope and probe. With my 100 mHz 'scope and my
> 150 MHz rated probes, I find the 'scope displayed peak to peak voltage is
> correct through 15 meters and it drops off noticably at 10
> meters. This is
> a frequency factor of approximately 5 (100 MHz to 20 MHz) where I
> can depend
> on the displayed voltage being accurate. Check your equipment
> before simply
> believing that it is telling you what you are seeing -
> instruments sometimes
> 'tell falsehoods' - know your equipment, and trust it only to the extent
> that it has proven itself.
>
> Once you know that you have good RF Voltage measurements, it is a simple
> matter to compute the power. The standard method of first
> converting to RMS
> voltage and then calculating will work, but you can also use formula
> reduction to do it all in one 'fell swoop' - the reduced formula for power
> from resistance and peak to peak voltage is: Power = (Vp-p)squared/8R, so
> for a 50 ohm load, it is a easy matter on most calculators to obtain the
> square of the peak to peak voltage and divide by 400. A 40 volt peak to
> peak RF Voltage will be 4 watts of power into a 50 ohm load. The
> derivation
> of this relationship is an "exercise left to the student" <G>.
>
> You may want to do calculation of the potential error values too.
> You should
> be aware that the potential percentage error of your power
> calculation will
> be proportional to 2 times your ability to correctly read the peak to peak
> voltage and also directly proportional to the tolerance of your
> dummy load.
> The net of all that is to say you will have to read the RF voltage
> carefully. A reading of 40 volts p-p will be 4 watts, but if your reading
> is in error by 1 volt (2.5%), the error (at that power level) will be 0.2
> watts (or 5%) - the percentages will hold true for all power levels, the
> actual error value will have to be computed for any particular level.
>
> All that is not to say that such measurements are bad - indeed quite the
> opposite, for I consider this method to be the best available for the
> equipment I have. It is just valuable to be able to quantify how far off
> the mark your readings and results can be due to whatever variables are
> beyond our control or abilities - that is why it is important to
> understand
> the limits of whatever equipment may be used in the measurements - we then
> know whether any 'problems' are real or are the result of measurement
> tolerances.
>
> If your goal is to have less than 10% power output variation across each
> band, I would expect that you could achieve that for all bands
> except for 10
> meters (10 meters is quite broad). The K2's power controlling
> ALC will also
> be a factor, and can contribute to a raw variation of up to 10%
> in addition
> to your potential measurement error. You can minimize the K2 power ALC
> error by taking many readings at the same frequency (I would say averaging
> 10 readings should be sufficient). If you do not average, and
> you read the
> RF voltage to within 1 volt, you could state that a 15% variation in
> readings would indicate that the 10% spec is met.
>
> 73,
> Don W3FPR
>
>
> > -----Original Message-----
>
> >
> > I am using a 50 ohm load that is intended to be used to convert test
> > equipment from hi-impedence input to 50 ohms. It is labled 50 ohms and
> > measures 50 ohms with my DMM. It should be purely resistive
> and I do not
> > have any equipment to prove that it is.
> >
> > The need to load it properly using the KAT2 was only because I
> wanted the
> > KAT2 in-circuit for improved forward power measurement within the K2. I
> > knew that it would need to load things to cancel itself out.
> >
> > Based upon what Ron, Don and Jack have said, I guess that I
> should remove
> > the KAT2 and RF probe and try something similar to what I tried
> > last time I
> > attempted to understand how well this radio tracks against the published
> > specifications.
> >
> > Would it be reasonable to just place a short piece of coax
> > between the load
> > and an O'Scope, measure the peak-to-peak voltage, divide by
> > 2*SQRT(2) and go
> > from there?
> >
> > Would this require knowing the coax velocity factor and dealing
> > with prioper
> > lengths of coax based upon wavelength (half or quarter) at the different
> > frequencies?
> >
> > I ask these questions because last time I tried using my Scope to
> > make these
> > measurements, the measurements were obviously incorrect because they
> > indicated that my rig was way more than 100% efficient.
> >
> > What I expect is that I can make apropriate measurements at 3
> > points on each
> > band and see that I get no more that 10% variation across said
> > band and from
> > band-to-band. I expect that I have a problem on 10 Meters because I am
> > getting a HiCur indication at 3 amps when set for 10 watts, so something
> > needs changing there. However, until I have trustable
> measurements on the
> > other bands, I expect that untangling the issues will be impossible.
> >
> > Help from the experts is needed on how to make such measurements without
> > fancy RF power meters. Obviously a purely resistive 50 ohm load is the
> > first requirement, and I believe that I have that.
> >
> > 73,
> > Fran
> >
>
_______________________________________________
Elecraft mailing list
Post to:
[hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
http://mailman.qth.net/mailman/listinfo/elecraft
Help:
http://mailman.qth.net/subscribers.htmElecraft web page:
http://www.elecraft.com