The AM Forum

I spent the better part of the day today trying to correct the intermittent arc-over problem in my remote antenna tuner at the base of the tower.  Suspect trash between the plates of one of the air variables.  But I ended up checking the calibration of the DC meters in both homebrew transmitters, checked two rf ammeters against each other, and attempted to get a close estimate of my power output.

I found the old Mirage wattmeter, typical of "ham radio quality" rf output meters, was reading about 33% high on the  high-power scale, compared to the thermocouple rf ammeter working into known accurately calibrated 50-ohm dummy load. Both rf ammeters read the same, so I have a little more confidence that they are accurate, although it is possible that both meters have the same %-age error.  The DC plate voltage meter and plate current meters are right on in the HF-300 rig, as measured with calibrated lab-type reference meters and cross checked with my Fluke DVM, but the plate voltmeter and plate milliammeter in the 8005 rig both read 3-6% low.

I determined that both homebrew rigs, running pushpull triode finals with plug-in coils and swinging link coupling, are  running only 50-60% carrier efficiency, as measured right at the output terminals of the transmitter.  The tube charts give efficiency ratings of about 70% for those tubes.  But then there is the feedline to the tower, and the remote antenna tuner at the base, plus the open wire line up the tower to the antenna.  I'm probably doing well to get 40% efficiency, comparing DC input to the final to the rf input to the radiating antenna.

The Converted Gates BC1-T measured at 70% efficiency, which appears about what it should be, but I haven't yet checked the calibration of the plate voltmeter or plate current meter since I acquired the transmitter, so that efficency reading is suspect.

I'd bet many of us have overly-optimistic estimates of our transmitter power amplifier efficiencies, and our carrier power output, and that's most  likely even more true in the slopbucket community.

Next, to figure out where the rf is going, when comparing the actual rf output to what the tube charts say it is supposed to be.  I had always thought the old swinging link tank circuits were highly efficient, compared to bandswitching tank circuits with all the unused coil turns in the vicinity of the active part of the tank coil.

How many class-E rigs have been measured to actually deliver 90% efficiency to the antenna?

Don,
Presently my class e 75 meter rig is at 88% and the 160 meter rig is at 90%.
Measure your efficiency at both ends of a band. Better efficiency at the low end means you have too much L in the tank. Better at the high end not enough.

I built a Class E RF deck (breadboarded and unmudulated) that put out about 700 watts at 93% efficiency, using 100V on the drains.  That was using #8 wire for the output transformer, a hi quality ceramic vacuum variable, and a tank coil made of 1/8 thick by 1/2" copper strapping.  The FETs were mounted on 1/4" copper plates.  Highly impractical but fun to build and play with to see how efficient I could get it.

My E-rig (QIX design)  runs at 89-90% ... and I have a very messy layout....

However,
Being a class-e amp has no effect on feed-line and antenna losses or efficiencies......


But, it wouldn't be too hard to deliver 90% at the tenna.... I wouldn't think....


Well laid out amplifier at 92% efficiency......
Low-loss feedline at a reasonably short length, and probably NO antenna tuner ......


Let's see....

An amplifier running at 500 watts input, and 460 watts output is 92% efficient....

So, to keep a 90% eff. at the antenna, we could loose no more than 10 watts from transmitter to antenna..

hmmmmmm...........

In this day and age of compact class E rigs, I'm waiting for the first extreme efficiency pervert to mount a rig on the top of a tower with the dipole legs coming off the rig's SO-239...

 ;)


T

All these so called efficiency numbers measured with a watt meter having a 5-10% error. Yea, right!

"In this day and age of compact class E rigs, I'm waiting for the first extreme efficiency pervert to mount a rig on the top of a tower with the dipole legs coming off the rig's SO-239..."

QRP nuts already do that....

Be a fun project ..... wish I had a tower! 

===============================

"All these so called efficiency numbers measured with a watt meter having a 5-10% error. Yea, right!"


Myself, I don't use a wattmeter....yes I have one, but it's really only good for watching needles bounce around.........
Haven't been able to find one worth a darn....(for a price I'm willing to pay...).

I just use the scope...............

I don't even OWN a wattmeter.  I measure efficiency using a VERIFIED (at a standards lab) dummy load (in other words, I know the EXACT resistance and impedance of the load), and a similarly calibrated oscilloscope.  I measure the RF voltage AT THE DUMMY LOAD with the oscilloscope, and do mathematics to calculate the output power.  There is no other good way to do it because the complex impedance of an antenna (even if very well matched) will cause errors, and of course the accuracy of RF power meters is always in question.  The 'scope into a calibrated load is the only thing I'll really trust.

While we're on the subject, Don, your efficiency MAY be better than you think.  Have you calibrated your current and volt meters?  I have found some to be WAY off (10% off or more), which, when multiplied makes the error even greater.  I always verify (and adjust, if necessary) the calibration of all amp and volt meters using a verified instrument.  Something to check, anyway !!

Regards,

Steve

Thanks for proving my point Steve. How many other guys go to such lengths?

does the term RSS mean anything.
Sorry none of us own real calibrated instruments.
We all own a bunch of old crap that we hope is accurate.
A 10% scope , a 5% load, a 5% ohm meter is what about 12% error. 

Which is OK, Frank. After all, who gives a crap if my rig is 90 or 91 percent efficient. It's a matter of a few watts, which make zero difference in the received signal on the other end.

So admit it...you're saying my Midland SWRS meter is worthless, right?  

So why is it when I thump my mawl down and the mud ducks start quacking, it sez I'm 100% efficient? Caw Mawn.

Flav - Flavor of Love

yupper, BTW you were putting a nice signal into CT.  on 160 last night.

I just get up tight when I hear wild claims that can't be backed up.
 

Yeah?  Well why R U ducking us in the Super Bowl? Right now your final is 100% efficient and I'm black out.

Throw away the midland the mud ducks know all and if they are a quacking it means they are a lackin
against the qro.

You'll be sorry when FCC goes back to the old rules.

The world is ending tomorrow, so I'm not concerned.

Have you changed the QST under the birdcage yet? You could be mistaken about tomorrow. The world might not end for another few days.

It's already tomorrow in Zulu time. Maybe it's the next day then. No matter what, the ARRL is still evil.

Hiram spelled backwards is MARIH.

That's five letters. Hitler had six letters. That's pretty close if ya axe me!

It's scary.

T

And if you rearrange the letters in Santa, you get SATAN! 

God Bless Glenn.

Well, I don't know - probably not many... I got some standard resistances and calibrated meters from the old standards lab at WPI.  With these, I was able to verify [and adjust] the calibration of the measurement instruments I'm using.  I used to work quite a bit with a Professor Howe, who used to run the standards lab.  He showed me how to check and achieve calibration, and it is very interesting, actually.

But, the main reason I did all of this is because as we start approaching and reaching 90 + % efficiency, the errors become a larger and larger part of the equation as the numbers of get closer to 100% (0% loss - which can never happen, of course).

I'm doing experiments using different topologies, circuits, components, etc. and in order for these experiments to be meaningful, there must be a control which can be relied upon.

I suppose if one wanted to be REALLY accurate, you would use the SAME instrument (say, a 'scope) to measure everything - including voltage and current (using a small, precision resistor or a current probe), as well as the RF output into a known load.  I suppose doing things that way would yield the most accurate measure of efficiency, as any errors would appear on both sides of the equation. 
But, I'm not quite *that* anal :-)

Regards,

Steve

Below is a photo of a high efficiency ham shack -  no transmission line losses.

Geez, I hope I don't give Timmy any ideas.  If you feel the urge to step out and take a whiz, that first step is a doozy.

That ain't ladder line, it be 'lec-ric goin' up to the refrigerator and the maul.

Dude, you're killing my bragging rights!  Actually I measured my power with a scope also because my dummy load isn't exactly 50 ohms and it would change resistance as the oil started boiling.  But you're correct, my Fluke 77 has 1.5% accuracy for DC current measurements and the used Tek 2465 has at best 5% accuracy.

Actually I never had the SOB on the air so I could care less about received signal.  It was purely an exercise in futility  ;D

John,
 You can have the bragging rights of all time best efficiency because all I have is a BS-O-Meter referenced to another BS-O-Meter.

Hey Frank, Do you have the matching BS Step Attenuator?

Attenuators are the only accurate thing in my shack because I can bring them to work and check them against real test equipment with real cal stickers referenced to the NIST.
Even then I rarely trust my own dynamic range numbers unless I get the same number 3 times on different days.
As performance increases I trust measurments less.

The BS-O-Meter S meter is another good one. 

You just might be a redneck if...

http://amfone.net/Amforum/index.php?action=dlattach;topic=8910.0;id=1391;image

My dummy load consists of twelve 600-ohm glo-bar resistors in parallel.  Not sure of the wattage, but each one is about 18" long and 1" in diameter.  I acquired them new old stock from 1945, but the data sheet did not include power rating, but a similar size wirewourd resistor would be over 200 watts.  That would make the whole thing rated at 2400 watts, with each resistor mounted vertically to produce convection cooling through the hollow tubes.  I car run the BC1-T at full power for hours, and although it makes a good room heater, nothing appears to overheat.

I have noticed that the DC resistance, as measured with my Fluke DVM, changes when the resistors heat up, from 50.0 ohms to +/-  2 or 3 ohms (can't remember which).  Interestingly, the SWR reading stays exactly 1:1 as the resistors heat up.

My best estimate of power output is made using that dummy load, and a thermocouple rf ammeter.  I have several ammeters, so I use more than one, and go by the ones that give identical readings, on the assumption that they are not all off by exactly the same percentage.

My next step is to check the calibration of the kv meter and plate milliammeter in the Gates.  If the efficiency turns out to be close to the same as the homebrew rigs, that will tell me that my power readings are probably low.  If  it indeed gives a higher rf amps reading on the same frequency, into the same dummy load, at the same plate voltage/plate current as the other transmitters, that will tell me that the other rigs are running at lower efficiency.

I think the FCC calls this "indirect power measurement."

Speaking of BC 1Ts, in the years I was in broadcasting, I worked at a couple of stations that used that transmitter.  Both had a board mounted on the left side of the cabinet that had big Globar resistors mounted in that fashion.  The board also had a ceramic light socket that a light bulb of about 40 watts painted red attached. 

I forget how the power was switched from the antenna connection to the dummy load, but when you transmitted into it, the light would illuminate.  I thought that was standard in that transmitter.

It was missing from mine.  Didn't see any imprints from removed components, so I'm not sure if the dummy load was removed, or if the transmitter never came with one.  As I recall, those transmitters used a dummy load made of  specially wound wirewound power resistors.  They work ok on the BC band, but get flaky at higher frequencies.  My load uses Glo-bar composition resistors.

Wow, PROBLEM !

If the load is increasing in value, of course the current will drop and the error will be MULTIPLIED (I ** 2) * R  And of course, the resistance may be changing more than you might be able to measure because you can't measure it "in circuit", so to speak.

As Frank, myself and others have experienced, it is VERY hard to ACCURATELY measure efficiency.  I've got a calibrated, verified RF amp meter around here (2 of them, actually), but I only trust it at the current at which it was calibrated.  I've found, particularly with thermocouple RF amp meters, the accuracy is not linear across the range of the meter.  I've even found this to be true of standard milliamp meters.  So, I calibrate them at the current I'm going to be actually measuring - at least then, I can trust the reading at the set operating point.

I am partial to using a 'scope [assuming the calibration has been verified !!] for calculating efficiency because you can see exact, and extremely small changes in the output voltage (across a known load).  Of course, this assumes a pure sine wave........ because the RMS calculations will be inaccurate for other waveforms.

Yikes !

Regards,

Steve

When I discovered class e (some years after Nathan) I wouldn't say a word about efficiency for about 6 years because I didn't think it was true. 90% just sounded like I was cheating.

Bottom line if you are still blowing out parts you don't have it right.
Layout is very critical before you can make claims of high efficiency.
You can't cheat the laws of physics........capt. 

Just to confirm.... Am I using the right formula to find my output power?

Using a scope, and dummy load...
Measuring the voltage at the load....

I've been doing this calculation....

Vpk-pk  X  0.35355 =  Vrms

(Vrms X Vrms)  /  Rload   =  Powatts


Obviously dependent on whether the scope is calibrated......
But, that is the correct method, right?

That's one way to read it... or .7071 x the peak (not peak to peak), but either way, it's RMS assuming a true sine wave.

This is where the calibration becomes critical, becuase an inaccuracy in the measurement instrument is multiplied twice (squared !!).