HFU HF Underground

Technical Topics => The RF Workbench => Topic started by: Zazzle on September 25, 2017, 2252 UTC

Title: About Vpp, Vp, Vrms and P(R) [solved]
Post by: Zazzle on September 25, 2017, 2252 UTC
Heyho,

I need a slap onto the back of my head so the bits fall into place. Also, yeah, asking this is kinda embarrassing. I feel stupid. Google doesn't really help here. It all says that my instruments are wrong.

The Oszilloscope (actually five) says Vpp=30V. The Wattmeters says ~8W. I say 'eh?'

If I'm not stupid, then "calculating P on 50R from VPP" works like this:

Code: [Select]
Vpp / 2 = Vp
Vp / SQRT(2) = Vrms
Vrms˛/R = P

Which is:

Code: [Select]
30Vpp / 2 = 15Vp
15Vp / SRQT(2) =~ 10,6Vrms
10,6Vrms˛ = 112,3 / 50R = 2,24W

That's theory. But all five power meters say "approx 8W".
The circuit draws ~1A @ 12V. So 8W out divided by 12W in is n= 67%, which sounds resonable. Also, power loss (heat) feels normal. So it checks.

What is it that I'm missing here?

Some small voice in the back of my head says "theory only looks at the positive swing. But the negative also contains energy, which makes it twice" (double the voltage = 4 times the power. Which would check. I could also run this through Fourier, which also requires to ignore the negative swing but uses two times the positive swing).

Code: [Select]
30Vpp / SRQT(2) =~ 21,3Vrms
10,6Vrms˛ = 453,7 / 50R = 9W

So, gimme a slap please.

Edit: read values from an CB Radio with Po=4W. Powermeter says: ~4W. Oszilloscope says: Vpp = 20V. Checks with my theory. But I'm none wiser.

Greetings,
~Zazzle
Title: Re: About Vpp, Vp, Vrms and P(R)
Post by: redhat on September 26, 2017, 0232 UTC
The formula I always use is.... W=(0.707 ( Vpp / 2 )^2 / Zload

For example, I have a transmitter producing 636V peak to peak...

Divide by 2 = 318V

Times SQRT of 2 (0.707) ~ 224.8Vrms

Squared .... 50546.73

Finally, divided by your load impedance,

1,010.9W

This formula only works with relatively pure, steady state signals.  Those with lots of harmonic energy can screw up the measurements, as can a reactive load.  One thing that bites me sometimes is forgetting that my Bird termination/wattmeter combo begins to roll off at 30 MHz, so anything below that is uncharted territory.  Case and point, I was repairing a commercial broadcast transmitter with it on ~800 KHz and the bird said around 150W.    The forward power meter on the transmitter said 1,000W, which was confirmed by my fingers on the end of the load and also a scope making voltage measurements.

On a similar note, if your scope is set up wrong, or the probes have a problem, your accuracy will suffer.

+-RH
Title: Re: About Vpp, Vp, Vrms and P(R)
Post by: Josh on September 26, 2017, 1822 UTC
Yep, meters can show a lot of crazy readings. Use one designed for the frequency range of use, however just about any meter, assuming it has an swr function, can show swr nulls even if out of range. Then you have pep vs average mode that needs to match what kind of signal you wish to measure. Don't use a cb meter for lower hf work, even high quality v/uhf meters will not read accurately on hf. I used to use a bird for hf but now use a daiwa 801 as it's more versatile.
Title: Re: About Vpp, Vp, Vrms and P(R)
Post by: Zazzle on September 26, 2017, 2259 UTC
Heyho,

The formula I always use is.... W=(0.707 ( Vpp / 2 )^2 / Zload

That's actually the same way I use. Just a different way to phrase it (0.707 = 1/1,41).

Anyway, I found the issue. It was a bit nasty and tricky to get behind.

1) The probes, set to 1:1 mode,  introduce the full capacitance reactance of the probe lead to the signal. Which is, at 14MHz, 100-200R (depending on the C of the lead). This alone affects the signal.

2) The lead isn't terminated (10M at the Input of the Oscilloscope) so the wire starts to act as a Lecher line, which also affects the signal (transforms the impedance).

Switching the probe to 1:10 made a really biiiiiiig difference. But the reading was still off.

I ran tests with a 40MHz, a 50MHz a 250MHz and a 500MHz Oscilloscope. The bandwidth of all devices were sufficient.

Different probes lead to different readings. Starting from 30Mhz it got better. All below was... way off.

In the end I found a nice workaround that gives good readings:

I mount a 10W dummy load on a BNC T-Splitter which gets directly connected to the Oscilloscope input. A 2m RG174 BNC-wire connects the BNC T-splitter with the TX unit.

So I have the dummy as close as possible to the Oscilloscope input and no line in between that causes isues. Gives proper reading on all Oscilloscopes. 55.6Vss. That's what I wanted to read!

Good night!

~Zazzle


Title: Re: About Vpp, Vp, Vrms and P(R) [solved]
Post by: redhat on September 27, 2017, 0010 UTC
That's one of the reasons I always leave my probes in 10x.  Check your tilt correction on the probes as well and make sure that is set correctly, or you will get excessive error.

+-RH
Title: Re: About Vpp, Vp, Vrms and P(R) [solved]
Post by: Stretchyman on September 27, 2017, 1833 UTC
Sounds 'Complicated'

Simply multiply the peak to peak voltage by 0.35, square the result and divide by 50.

For measuring power with a scope, just use a BNC T piece and suitable adapters.

Power flows thru' the top of the T whilst the bottom leg has a probe stuffed in it.

Got a 100:1 probe cheap and can be used for powers for 0 to over a 1000W.

Simples...

Str.