The AM Forum

Simple question, how much audio power do you need to modulate a 250 watt AM transmitter?  Answer 125 watts. Yes, no.  I am finding several differing thoughts on that subject.  Some say, if you put 125 watts of audio into your calss C amplifier, you must also de-rate that amount by the efficiency of the class C amplifier, others have said other things.  I would like for you folks to discuss this.

Bob,

Here is my understanding of the topic. This applies to high level class C plate modulation only.

For 100% modulation with sine wave audio tone, modulator audio OUTPUT power must be 50% of class c rf INPUT power. For asymmetric voice audio, average output power will be somewhat less than 50% for 100% modulation. In either case, spare audio power capability is a good idea.

For example, if you build a transmitter with 500W input power, say 2000VDC @ 250mA, you would probably get about 350-380 watts of output power from the RF stage. To modulate 100% with sine wave tone, you would need 250 watts audio output. To modulate 100% with voice audio, you would probably need somewhere between 200 to 250 watts of average audio power. I would try to design a modulator capable of between 270-300 watts output to have some spare audio capability.

Regarding the modulation transformer, if you run the RF plate current through the secondary of the mod transformer, then you would need a 250W unit to fully modulate a 500W input RF section. If you use a mod reactor so that the RF plate current does not traverse the modulation transformer secondary, then you could probably get by with a slightly smaller unit.

In my recent build, I am running the 822 final at 2000VDC and 250mA and am getting 375 watts output. The mod transformer is a UTC S-22 rated at 250 watts. I am using a mod reactor and there is no DC current through the S-22 secondary. My modulators are 811A's at 1350VDC and are capable of about 320W audio output. So you can see that I have "overbuilt" my modulation system.

This is my crude way of looking at class C plate modulation, but it has worked for me. I'm sure the EE's on the board can give you a more technical explanation.

Ron

A 250 watt transmitter is around 335 watts input power, assuming average class C efficiency of 75%. So your 250 watt transmitter has a modulator that would at minimum put out 170 watts. However, you have to take transformer losses into consideration. The best advice I've heard is to just simply build a modulator that is capable of the same power output as the RF final. So if you have a 250 watt final, build a 250 watt modulator.

Just do it the way it was always done, use Input power for the Class C calculation and dont fuss over efficiency if its only 60% on 10M or less on 6M, and 75% on 75/160.

Build a modulator that can provide about a 125% minimum (this assumes no processing otherwise go for 150%) of the 50% audio Output needed since that is based upon tube manual audio output and not what comes out of the transformer. This overhead also allows for tube aging since hams are notorious for buying used tubes of unknown remaining life......as long as they are cheap or free ::) ;D

A pair of 811's is fine for a single 813 as an example and very little of their 340W theoretical capability will be wasted.

There is nothing that irks me more than listening to a strapping carrier and pipsqueak audio.

Carl

The plate dissipation rule of thumb...

Well, the books I've read all say that the power required to plate modulate a class C stage is 50% of its input power. The efficiency of the modulated stage is irrelevant and does not enter into the calculation once you're thinking input power. So the answer would be 50% of the input power of the modulated stage must be available at the output of the modulation transformer. However, some allowances should be made. The human voice is asymmetrical, not a sine wave, so add some power for that. A modulator running at its limit has more distortion than if it has some room to spare so add some more power for that. Something on the order of 100% would have a good safe design margin.

A common setup for AM broadcast transmitters is to use a pair of tubes modulated by a pair of the same tubes. So a comfortable rule of thumb would be to have as much plate dissipation in the modulator tubes as there is in the RF final tube(s). A classic AM ham setup (as Carl mentions) is an 813 modulated by a pair of 811As. 125 W Pd for an 813 and 65x2=130W Pd for the 811As. A good match. Anything beyond that is just for bragging rights ;)

Sure, less may work ok, but if you're going to build a rig - make it a good one!

-Don

This was my first reaction.  Pick the tubes.   Over the years I have often run a single RF final modulated by a pair of the same. This has produced excellent headroom and cleanliness.  I've run a single 813 by a pair -  a single 4X1 by a pair...   a single 833A by a pair... and many more.  All were able to do 150% positive easily and cleanly.  This is well suited for a 2:1 step down mod xfmr.

The the tetrode / triode equivalent might be a 4X1  X  3X1's.   Or a 4-400A X 3-500Z's.   Or as already suggested, an 813 X 811A's.  (Though, with air, an 813 is probably more of a 250 watt tube)


The other choice is a pair modulated by a pair. A 1:1 mod xfmr ratio is suited here.  I now like 2X2 the best. Over the last five years I have leaned towards running audio around 125% positive and have no need for bigger headroom.  But it MUST be clean.

After running THD tests on many AM rigs, I saw how quickly distortion gets bad when we get at or very close to modulator saturation. We want the modulators to loaf up to the peak power we require - or we must pay the price in the form of a wider signal.  It takes a good ear to hear the audio quality difference ON frequency - but splatter OFF freq can be heard by everyone.

Also, the old Tron rule -  the mod transformer needs 10 pounds per 100 watts of audio power. This is a  good idea, especially if extreme clean lows are desired.  The classic Gates, RCA-type  BC mod iron rated at 1KW, etc., can't be beat. (used with a Heising reactor)

T

It's the opposite that irritates me: those that insist on running a tiny carrier but big audio. Sure it may sound good on a sync detector, but not everyone has one. When done right, running less than 100% modulation (but not less than maybe 75%) can have its advantages. One thing I noticed is weaker modulation (less than 100%) is a little easier to copy in high static than heavy modulation (over 100%).

I have to agree. Most BA receiver detectors don't really like mod % over 100 (or 90). Sync or not doesn't matter. It's one thing to listen to 125% mod on a fine signal ground wave, and impressive. Beautiful. But here in TX, an awful lot of the 'east coast sound' stuff we hear is unintelligible mush, regardless of % modulation. More mod % is just louder mush.

In my humble opinion, the best (distortion free, loud, always intelligible ... and always intelligent!) ham AM signal on the air anywhere is K4KYV, period.  Whatever Don is  doing, he's doing it right.

Other than severe distortion, which is rare amongst the AMers, the transmitted EQ curve may be a big part of why it  may be hard to copy some.  Boosting too much low end or having a muddy middle because of no EQing at all can cause intelligibility problems.  We all have different voices and must find the best curve to give the desired result.

Running a high % of positive modulation will cause intelligibility problems when there is selective fading. But, yes, many older receivers can't handle over 100% positive from diode detector overload. Pity, cuz most male voices, when limited to -95% negative, naturally reach up to 125% positive or higher. I know mine does.  So if I limit my positive peaks to  + 100% positive, my negative peaks (pinched carrier) are modulated only about 70% or so.  Not an optimum signal to noise ratio.


There's not much that can be done on the receive end for poor transmit EQing, but by adding a simple SoftRock that has a sync detector, the higher modulation levels above 100% could be handled better.  It pays to have two AM receivers going in the shack for this purpose.

**  BTW, I noticed today for the first time, a spell checker is working as I type a reply. Fantastic!!!!  Tnx to Steve / HX or whoever got it working.  Lords knows I need it.  Now  a grammar / phrasing checker. Just imagine.  ;)

T

Wat spel cheker? I bet it's the spell checker built into your browser, because if it's on this site, it's obviously not working.


Anyway, one thing I have noticed that makes high positive modulation easier to copy on some receivers is to slow the AGC down if possible, or just use the manual RF gain.

It's in your browser. Grammar/phrasing checkers are generally in most word processors. If you have a long and wordy technical or otherwise dissertation, you can type it into a word processor, check it for grammar, punctuation, and phrasing (i.e. no dangling participles, etc.), and when satisfied, copy and paste it into a new post. Most likely, it can add an additional level of clarity to your post(s).

My current plan (subject to change based on input from you folks) was to run a pair of 811a's using a TRIAD 12AL 125 watt transformer, to modulate a single 4-125a running at 1400vdc @ 255 mA input.  357 watts in, maybe 250 out.  I can build the modulator power supply to do whatever, as I will also have about 60 HY of modulation reactor in the line.  Both designs are straight out of the 1956 (14th edition) west coast handbook. I have talked with several of you about this rig over the past years and I appreciate all of your input.

Sounds like a plan for piss weak audio; did you read and understand what I and others said?.
Half or 357 is 178.5 which is what you need for audio at 100% in a perfect world of theory. On a real earth at least 250W is needed and better yet at 300W for some reserve and not pushing the modulator hard.

Carl

Agreed, Carl.  In my original correspondence with Bob, I had recommended the following alternate choices: Thordarson T11M77, UTC CVM4 or its daddy the VM4 or the Stancor A-3898.  The problem is that old friend voltage across a given class C impedance load.  If we can assume approximate 125% positive peaks on modulation, we have to design in more than the classic ½ audio power for a given class C power. 

Interesting about HLR’s rule of thumb of 10 lbs per 100 watt af power rule.  Turns out that those modulation transformers I recommended weigh out about 20+ lbs.  Now – if one wishes to design the classic “yellowey” audio parameters of 300 to 3000 cycles than we can take out a little from the mod iron requirements but you still cannot fool the typical requirements of the male asymmetrical voice characteristics.  My positive peaks come out to about 150% positive peaks with 100% negative points.

Interesting dialogue

Al

The thang that really makes me take notice about modulator overhead is how much power it takes to modulate a carrier at various percentage levels.

Someone correct me if I'm wrong, but 100% modulation requires X4 peak power.   150% requires X6 power.   200% requires X8 peak power.   (corrected numbers)

So even if we are looking at a natural 125% positive peak modulation level, we need X5 peak power, no?  
So much for the X4 requirement.  

I think many of the older boat anchor rigs were designed with X4 in mind and show it as they flat top around 110%. Or they even start to show poor linearity when approaching 90% positive.   But when we build our own, we can do away with these limitations.

The problem sometimes is we are limited by our modulation iron. The tubes may handle it but we may crap out our mod iron in the process. The worst risk is voltage.   To produce more peak power often means running a higher voltage.  The mod iron may be able to take the power, but even using arc gaps, the risk of internal flashovers is real when pushed.

That old heavy-duty BC mod iron is "almost" worth its weight in gold ... ;)

T

Bob,  

Those 811A's can easily modulate a single 4-125 at 350 or so watts input well over 100%. At 1000VDC on the plates, they can produce about 250 watts of audio, and at 1500VDC, they can go up to 340 watts of audio. Your design has provided for plenty of spare audio capability.

The modulation transformer however, is a bit small for this design, but if you use a modulation reactor setup, you might be OK. I agree that it would be desirable to get a larger mod transformer if possible, or alternatively, run the final at a lower power input, say 275-300 watts input.

If you want to run 350 watts input, I have a Merit 175 watt mod transformer that might fit the bill nicely. Let me know if you'd like to have it and I'll send you the details.

Ron

If the bandwidth is rolled off below 300 hz the natural asymmetry disappears. I spelled all those words incorrectly.

Yes Carl, I understand, that is why I said: my current plan subject to change etc. and  I can build the modulator power supply to do whatever, I can get a bigger transformer and do what you and the other prescribe, or I will de rate the RF output of this transmitter, (all of which is 85% finished), and build another that is larger in all respects. And thanks for the offer Ron, I have sent you an email.

not sure if you are referring to the N times carrier = PEP (the PEP bs) but at 150% positive, your peak power out should be six times dead carrier power.

Thanks for the info, Rob.

So then it's X5 carrier power for 125% modulation -   X6 for 150%  - X8 for 200% and X16 for 300%.


T

300 percent is 10x I think.

Maybe we need a chart!  Haha

On modulation peaks, the electrical plate power input to the transmitter's rf output stage is (1+k)(1+k) x the electrical plate power input at carrier... where k is the peak modulation index (as a percentage)/100%.

For example, if the positive peak modulation index is 150%, then k= 150%/100% = 1.5, and (1+k)(1+k) = (2.5)(2.5) = 6.25

Therefore, if the peak positive modulation index is 150%, the peak electrical plate power input (on audio peaks) to the transmitter's output stage will be 6.25 x the electrical power input to the transmitter's output stage at carrier.

HOWEVER, there is a very non-intuitive and possibly confusing aspect to this.

While (for the particular example of 150% positive peak modulation index), the peak plate power input to the transmitter's output stage is 6.25 x the plate power input at carrier, the peak audio power that the modulator is providing is only: k/(1+k) of this peak power... i.e. 6.25 x 1.5/2.5 x the plate power input at carrier= 3.75 x the electrical plate power input at carrier.  

3.75 + 1 does not equal 6.25 (!). So where is the extra peak power coming from???

The extra peak power is coming from the stored energy in the capacitor at the output of the DC power supply that provides the unmodulated B+ to the RF output stage. [I.e., on 150% positive modulation peaks, the current being drawn from that capacitor is 2.5 x as large as the current being drawn from that capacitor at carrier... while the voltage across that capacitor is (essentially) the same as it is at carrier]

Therefore the DC supply for the unmodulated B+ is providing (from its stored energy) 2.5 x as much power on modulation peaks as it does at carrier.

3.75 + 2.5 = 6.25

As a second non-intuitive and possibly confusing aspect of this... the peak power being provided by the modulator's audio amplifier (i.e. not including power being drawn from the stored magnetic energy of the modulation transformer) is k/(1+k) x the total peak power being provided by the modulator (including power being drawn on positive modulation peaks from the stored magnetic energy in the modulation transformer).

Therefore, for the example we are considering (150% positive modulation peaks), the peak electrical power being supplied to the plate of the rf stage by the audio amplifier of the modulator is 3.75 x 1.5/2.5 x the electrical plate power being supplied to the rf stage at carrier = 2.25 x the electrical power being supplied to the plate of the rf stage at carrier.  

The total peak power budget is:

(k x k) x the electrical power being supplied to the plate of the rf stage at carrier from the audio amplifier

+

k x the electrical power being supplied to the plate of the rf stage at carrier from the magnetic energy stored in the modulation transformer

+

(1+k) x the electrical power being supplied to the plate of the rf stage at carrier from the energy stored in the capacitor of the unmodulated B+ supply

The bottom line is:

At a minimum, the modulator's audio amplifier must provide a peak audio power output of k x k x the electrical plate power input to the rf output stage at carrier.

Stu

The nice thing about using 811As is that you can swap to 572Bs in the future if you want more headroom. Leave some extra room for them. A pair of 572Bs will modulate your 4-125A right out of its socket. :o

-Don

Stu,

Thanks for the detail explanation on this subject of PEP.  I'll have read a few more times to understand everything.

Fred

Another view of modulation power is to use the AM carrier power and modulation index formula:

Pt = Pc X (1+m^2/2)

where Pt is the total output power consisting of carrier plus modulation power, Pc is carrier power, and m is the index of modulation.

Here is a table for determining the Total ouput power at percentage modulation and the minimum modulation power required:

For example 25% modulation is a modulation index of 0.25. m^2 = 0.0625, m^2 divided by 2 = 0.03125, 1 + 0.01325 = 1.03125, so total power output Pt for 25% modulation for a 1kW carrier is 1,031 Watts. Modulation power required would thus be 31 Watts.

Modulation index          multiplier coefficient (1+m^2/2)


0.25                           1.031

0.50                           1.125

0.90                           1.405

1.00                           1.500

1.10                           1.605

1.25                           1.780

1.5                             2.125


Taking for example the maximum power output for an AM broadcast station with a carrier power of 1kW and a modulation percentage of 125%;

Total Power Output at 125% Modulation is Pt = Pc X  1.780 = 1,780 Watts
Pm is Modulation Power is = Pt - Pc = 1,780 Watts - 1,000 Watts = 780 Watts.

Usual design considerations for this example would be to have an 800 Watt modulator circuit for a little headroom.

If one were to design a Class B P-P modulator circuit for a plate modulated 1kW Class C final, and assuming an efficiency of 70% for the 800 Watt P-P modulator,

the DC input power to the Class B P-P circuit would have to be 1,143 Watts.


Phil - AC0OB

To add to this, assume a Class C stage efficiency of 90% for the 1kW plate modulated final, the DC input power to this stage would have to be:

1,111 Watts.

A well designed (stiff) DC power supply for continuous 125% modulation would require a total DC input power of:

DCc + DCm = 1,111 Watts + 1,143 Watts = 2,254 Watts.

This does not include filament power or control power.

Pos.      Mult
Mod      Factor
100% = 4x - Normal
150% = 6.25x - Reasonably common male voice natural asymmetry reproduced through a high quality system.  A minimum design/implementation standard IMHO.
200% = 9x - Artifical enhancement required to actually modulate at this level (not recommended); generally implies waveform modification (AKA, clipping) for normal program material.  Good headroom.  My personal standard for AM transmitter design (lots of headroom).
250% = 12.25x - Eh?
300% = 16x - Are you sure you're not using carrier suppression of some sort??  RX better have a sync detector!

I'm certain for those running 150% etc. natural positive modulation that you're also reducing your carrier power so that you're not exceeding 1500w PEP output on AM when you're on the air. That would be 240 watts of carrier output.  ;D

Al VE3AJM

Well fortunately, AM signals are made up of 3 distinct components - carrier, upper sideband and lower sideband.  A vacuum tube 1kW power input transmitter, putting out - say - 750 watts of carrier, 100% modulated will show, on a spectrum analyzer, the carrier power of 750w, and each sideband will show a peak (not average) power of 375 watts.

Ok, now we have a field day operation with 3 transmitters, each operating at - say - 1500 watts PEP, on the same band (ok, it's an example) - when viewed on a spectrum analyzer will show the 3 distinct signals of 1500w PEP, and most rational people would say this is a legal setup.  Are they actually running 1500W PEP?  If, somehow, we combined similar signals into a single feedpoint, and amplified them with a single linear amplifier, and fed this to an antenna system so that all of the signals were present at the same time on a single feedline, would one say that we now have 4500 watts PEP?

If I have a 750 watt carrier transmitter, and a 750 watt PEP double-sideband transmitter phase locked to the carrier generator, and I generate the identical signal as in the first example, how would this be measured?  Is it 750 watts?  How does this differ from the field day operation?  If we suppose that the 3 1500W (PEP) field day transmitters operating at the same time are legal, then by the same measure, I'm only running 750 watts!!

Anyway, you can see what I'm getting at...and the slope is very slippery  ;D

However you want to measure your PEP power output, instantaneous or average peak power on AM, is up to you of course. For me, the more AM stations out there with good signals, the better.

Whether your FCC is in agreement with your method of measurement and conclusions you've written about here, may be an issue for them if there were any enforcements of the rules with regard to AM power. Slippery slope and one foot on a banana peel deal :)

Al VE3AJM

So, if Im using a CE-100V or any other rig that allows selection of DSB OR either sideband AM what is my PEP in both conditions assuming the carrier is the same at 25W? 100W on DSB and 50W with one SB......correct?

Since the receiver (most) only demodulates one SB it will sound the same if the listener has zerobeated the carrier and doesnt tune around.....correct again?

Now, if I reduce the carrier a bit but it remains constant, and I increase the mike gain, I can increase my modulation index to say 125% or whatever I want within reason and boost the perceived audio in the receiver as long as I dont exceed its detection threshold. Or I can boost the carrier and audio to 100W PEP and still stay in the rigs spec.    Am I still on the right track?

With slightly reduced carrier, the overall efficiency is better meaning less wasted heat and the slightly elevated single SB has less total power than both at 100% further reducing waste.  What is wrong with this scenario?

Since I dont have a Class AB or B modulator negative-positive cycle to worry about do I just have to clip the low level audio when exceeding 100%? The 100V has a clipper built in followed with a rather elaborate set of filters that limit the audio to 4KHz as well as substantially reducing higher distortion products.

Are there any other rigs with that flexibility? The 200V eliminated the switchable SB's on AM as part of their cost reduction. Other phasers from CE, Hallicrafters, Lakeshore, etc, can select a SB in SSB mode and then inject carrier which is a minor inconvenience while the 100V has a seperate carrier control for AM behind a front panel door.

With 20-30W of AM carrier available driving a big amp is no problem. It can really generate serious RF from an Alpha 77SX (a pair of 8877's) Ive been repairing and testing with the TS-950SD on AM  ;D

When driving a linear with only one SB on AM can the amps carrier output be increased a bit wiithout running into its distortion threshold? IOW since it will be driven at 50W PEP instead of 100W PEP with DSB can I diddle with carrier and audio controls to boost the PEP another 3dB?  I dont have the 100V ready to go yet so no scope or SA tests have been run.

Carl

Why would anyone want to run only one sideband and carrier?

It takes away a giant advantage on the air.

Also cant be demodulated by simple means without plenty of unwanted distortion.

See KWS-1.

Dont know as Ive not tried it. And Id of thought that a single SB with a 5-8kHz BW would be more desirable than DSB in the same space.

But some Googling answered some of my questions primarily in a 2006 post by our own K4KYV; thanks Don.

OTOH I wouda thunk that a simple plug in or wire in SS adapter for the 6H6/6AL5/1N34 would have been developed by now and corrected the high distortion.

Do all conventional AM detectors going back to the 20's suffer the same distortion with SSBAM? All I see mentioned is the half wave diode.

For weak AM that isnt FMing I find myself using the BFO or product detector anyway and leave it on when a strong station enters the conversation. Of course drift, FMing, and the inability of some to zerobeat doesnt help ::)

Carl

Steve,

I'm glad to see you have come  round to my way of thinking on this issue.  Remember a thread about a year ago, when you took exception to my posting that was essentially an identical argument, and you posted a couple of diagrams that analysed an AM signal both in the time domain and frequency domain? The time domain, as would be viewed using the envelope pattern of a wide-band oscilloscope, clearly demonstrated the vectorial sum of USB, LSB and carrier, which indeed indicated the 4X power at 100% modulation.  But the frequency domain, as displayed on a spectrum analyser, showed just as clearly the carrier with distinct upper and lower sidebands each peaking at an amplitude far below that of the carrier.

And who says all the transmitters even have to be on one band. During field day, each transmitter operating on the same band would most likely still be modulated by different audio signals. But what about W1AW, when they simulcast on SSB on several bands at the same time?  If you had an oscilloscope connected to a non-resonant pick-up, or a simple untuned field strength meter, wouldn't the indicated peak amplitude of the sum be far above that of any one of the transmitters?

The 4X argument assumes a carrier that actually varies up and down in step with the audio doing the modulation. This theory was disproved sometime in the 1920s. As we have long known, there is, instead, a steady unvarying carrier,  with two mirror-image sidebands, emitted independently above and below the carrier frequency.

Regarding the audio power required to plate modulate a carrier, with a sine wave tone the audio power is 1/2 the DC input to the final stage. I have seen estimates that even a very good modulation transformer is only about 90% efficient, so the modulator tubes need to deliver slightly more than 50% of the DC input.  With voice modulation, the PEAK audio power required to modulate 100% is 50% DC input, so the average power with voice modulation will be more on the order of 25%.  The exact figure depends on  the characteristics of the voice doing the modulation.  Because of the high peak to average characteristic of the typical  human voice, the modulator tubes may be rated at substantially less plate dissipation than what would be required for sine wave tone modulation, but the DC plate voltages and modulation transformer turns ratio must still be such that the modulator is capable of delivering the required power albeit over a short duty cycle. Just don't whistle into the mic or run a sine wave tone generator for more than a half second or so.

The audio power level increases as the square of the percentage of modulation. At 141% positive peak modulation, the required audio power is the same as the DC input. At 200%, the audio power is twice the DC input power. The 1950s era ARRL handbooks have a section on asymmetrical modulation, and show a drawing of what 300% modulation would look like on a scope. That level would require 9 times the power for 100% modulation, or 4.5 times the DC input power. I doubt if anyone's natural voice asymmetry reaches 300%; now we are talking about DSB reduced carrier, like what W3PHL used to run.

Don k4kyv

And such displays/descriptions have no conflicts with modulation theory.  They exist simultaneously when sampled by such respective instrumentation.

And Part 97 does not seem to proscribe time domain measurement method.

Maybe exalted carrier detection would work but those things were forgotten long ago.

Otherwise a simple detector isn't going to make reception much fun. Better off with that other phone mode.

Have there been any known or recorded cases of any enforcements of this by the FCC on an AM station. If so, do we know which method was used/applied by the FCC? They may just use a Bird peak reading wattmeter for all we know. That wouldn't work all that well if the station was operating a link coupled output transmitter and OWL all the way to the antenna feedpoint.

Al VE3AJM

Same number as DOC : )

 :D OK FIINE!

I do recall that an Industry Canada inspector shut down AAM back about 10 years ago. 5kw output got their attention.

Al VE3AJM

Of all of the points that have been made about PEP measurement.... I personally find this one the most interesting.

Steve posted:

"Ok, now we have a field day operation with 3 transmitters, each operating at - say - 1500 watts PEP, on the same band (ok, it's an example) - when viewed on a spectrum analyzer will show the 3 distinct signals of 1500w PEP, and most rational people would say this is a legal setup.  Are they actually running 1500W PEP?  If, somehow, we combined similar signals into a single feedpoint, and amplified them with a single linear amplifier, and fed this to an antenna system so that all of the signals were present at the same time on a single feedline, would one say that we now have 4500 watts PEP?"

In fact, if you measured the time-varying envelope power... which is the average, over 1 rf cycle, of the instantaneous power of the composite signal formed by the sum of the three signals that Steve described... then the peak value of the envelope power would be 9 x 1500 watts = 13500 watts. This result follows from the fact that since the 3 distinct signals are separated in frequency by only a very small percentage of the average of the three carrier frequencies... there will always be instants in time (that last many RF cycles) when the three signals add approximately in phase. During those instants in time when the three signals add approximately in phase, the combined RF voltage will be the sum of the three individual RF voltages (i.e. 3 x the individual RF voltages)... and the combined envelope power (whch is proportional to the square of the RF voltage) will be 9x the envelope power of one signal considered by itself.

However, there will also be instants in time (lasting many RF cycles) when (for example) two of the three signals are 180 degrees out of phase... and their RF voltages will cancel. At those instants of time, the envelope power of the composite signal will be equal to the peak envelope power of just one of the signals.

Therefore, at some instants of time, the envelope power of the combined signal would be 9x the envelope power of one signal. At other instants of time, the envelope power of the combined signal would be 1x the envelope power of one signal.

Averaged over a large enough number of RF cycles, the individual RF powers would add... but that is not the engineeriing definition of the envelope power of the composite signal.

This phenomenon is the same as what happens if two tuning forks at approximately the same frequency are vibrating simultanously in approximately the same location. You will hear, in each ear, the sum of the sound pressure waves from both tuning forks. You will also hear, in each ear, a "beat note" at a frequency equal to the difference of the two tuning fork frequencies. The beat note results from the pressure waves, falling on each of your ears, coming into and out of phase. The time to make one complete cycle of: in phase- out of phase- in phase is: 1/(the difference of the two frequencies).

Assuming the two individual acoustic signals are appproximately equal in amplitude... then at instants in time when the two audio signals are in phase, the associated instantaneous acoustic power falling on each of your ears is 4 x the instantaneous acoustic power of either of the two audio signals. When they are out of phase, the instantaneous acoustic power falling on each of your ears is approximately zero. The time averaged [i.e. averaged over an interval of time that is equal to 1/(the beat frequency)] acoustic power falling on each of your ears is the sum of the individual time averaged acoustic powers.  

As an aside, spectrum analyzers generally do not display peak envelope power... They display the average power in a narrow band of frequencies... averaged over a period of time that is approximately 1/(the width of that narrow band of frequencies). For example, if the resolution bandwidth of the spectrum analyzer is 100Hz... then what is displayed is the average power in that 100 Hz band of frequencies... averaged over a time interval of at least 10 milliseconds.

I don't know how the FCC would measure compliance... but my guess is that they would interpret the rules by measuring the RF envelope of the signal with an oscilloscope... observing the peak positive modulation index (k)... and defining the the peak envelope power as: the unmodulated carrier power x (1+k) x (1+k). This interpretation may not be "fair" or "reasonable" in some people's opinions... but there are lots of legally enforceable rules that are not viewed as fair or reasonable by various people who may be subject to those rules.

Stu

350% = 20x - They call me missa wurl wide audio drive!
400% = 25x - Creates tear in space/time and station gets crushed inside black hole. That's one way to clean the shack i guess.  ;D

It sure is great to have Stu in here!  Thank you!

Stu gave more support relative to the "AM Problem".  AM is, in fact, 3 distinct signals - which, just like the field day operation - if everything were combined into a single electrical path, would, under some circumstances (in phase condition) add up to more than is effectively there.

So, just like the field day operation - I could choose to operate 3 transmitters - a 1500 watt carrier output transmitter, and 2  750 watt PEP single sideband transmitters - one per sideband.  The sideband transmitters are phase locked to the carrier.  I run 3 coax runs up to my combiner which is located at the top of the tower.  The combiner output is connected to the antenna.  Peak reading wattmeters inserted into any of the individual coax lines show a maximum power at 1500 watts (the carrier transmitter).  The others show 750w PEP, and like the field day operation - all is perfectly legal.

Which points to the obvious conclusion that measuring the 3 distinct pieces of an AM signal, combined on a single electrical path (a coax cable or other output) is only valid when measuring average power.  A PEP measurement of the combined signals is not valid.

Steve

Minor comment (for clarification):

RF signals at different frequencies cannot be phase locked. They can, however, be forced to maintain a fixed frequency separation. If the frequency separation is fixed at B Hz, then they will move in and out of phase and back in phase again every (1/B) seconds.

When they are in phase, their sinousoidal voltages will add to produce (A+B) volts. When they are out of phase, their voltages will add to produce A-B volts... where A is the amplitude of one of the sine waves, and B is the amplitude of the other sine wave.

The composite envelope power (i.e. the instantaneous power averaged over 1 RF cycle) will peak at (A+B) x (A+B) / 2R... where R is the associated resistance between the two points across which the voltage is being measured. The minimum instantaneous envelope power will be (A-B) x (A-B) / 2R. The time averaged power in the composite signal will be [ A x A / 2R] + [ B x B / 2R]

Stu

Don

I very respectfully disagree with the following posted comments:

"The 4X argument assumes a carrier that actually varies up and down in step with the audio doing the modulation. This theory was disproved sometime in the 1920s. As we have long known, there is, instead, a steady unvarying carrier,  with two mirror-image sidebands, emitted independently above and below the carrier frequency.

Regarding the audio power required to plate modulate a carrier, with a sine wave tone the audio power is 1/2 the DC input to the final stage. I have seen estimates that even a very good modulation transformer is only about 90% efficient, so the modulator tubes need to deliver slightly more than 50% of the DC input.  With voice modulation, the PEAK audio power required to modulate 100% is 50% DC input, so the average power with voice modulation will be more on the order of 25%.  The exact figure depends on  the characteristics of the voice doing the modulation. "

The instantaneous rf power in an RF signal is proportional to the square of the amplitude of the signal [e.g. instantaneous power = v(t) x v(t) / R)... where v(t) is the instantaneous voltage, and R is the resistance between the two points across which the voltage is being measured].
The envelope power of an RF signal is defined (in engineering terms) as the average, over one RF cycle, of the instantaneous power in the RF signal, and is equal to 0.5 x the peak of the instantaneous power in that same rf cycle.  If the positive peak amplitude modulation index is k, then the peak amplitude of the RF signal is (1+k) x the amplitude of the RF signal at carrier (i.e. in the absence of modulatiion). The peak envelope power of the modulated signal is (1+k) x (1+k) x the envelope power at carrier. If k=1 (i.e. 100%), then the peak envelope power is 4 x the envelope power at carrier.

We apparently disagree here (that's what make horse races)... but I will reiterate my result from above. The peak audio output power (on positive voice peaks) that the modulator must deliver (not including losses in the modulation transformer, etc) to produce a peak positive modulation index of k (e.g. k=1.25 => 125% positive peaks) = k x k x the DC input power of the modulated rf stage.

Since voice signals reach their positive peaks only a small percentage of the time, the average audio power that the modulator will have to deliver to its load will be much less than half of the peak audio power it must deliver.
 
Of note, modern audio amplifiers are rated in terms of the peak audio output power they can deliver to a specified load... even though they cannot deliver anywhere near that much peak audio power if the audio waveform is a sine wave. [With a sine wave of that level, the amplifier would be be producing too much average power... and it would overheat/shut itself off/destroy itself]

Classical audio amplifiers were generally rated in terms of the average audio output power they could deliver to a load when the waveform was a sine wave at 1kHz. When the audio waveform is a sine wave, the peak audio power delivered to the load is 2x the average audio power.

In any event, when designing a modulator, it is (in my opinion) better to think in terms of the peak audio output power (on positive voice modulation peaks) that the modulator must deliver.

Stu

Yes, good clarification!  I should have said the sideband transmitters are derived from the carrier signal, but of course they are on their own frequencies, so cannot [by definition] be phase locked.

This is a good discussion, even if it is a deviation from the original thread  ;)

Excellent point. And this is where so many people go wrong in their frequency domain based "analysis" of PEP.

and to think all I was looking for is how big a mod transformer to use.  :)

This has turned out to be a very good discussion.

Stu,

Is that an error in the math,  (A+B)/R x (A+B)/R ??  I also see you corrected it later in the post.  Or maybe I'm wrong.

Fred

Grinning and thinking of how much fun it must have been for you to hear about that every morning even with the receive antenna disconnected.

Almost as much fun as I have with tourists at work AND at home in summer...

But they are phase coherent and that is all that matters when measuring the power of a signal. Otherwise, you could claim a harmonic is not part of the fundamental but a separate signal. Or that a SSB signal has zero power since none of the components are in phase. Try either of those with the FCC. Good luck.

Further, most (all?) wattmeters/power meters are phase insensitive. So if the FCC hooks one to the output of your TX, all arguments about phase are irrelevant.

Finally, let's look at Part 97. The definition of PEP is:

(6) PEP (peak envelope power). The
average power supplied to the antenna
transmission line by a transmitter during one RF cycle at the crest of the
modulation envelope taken under normal operating conditions

Note the section 'power supplied by a transmitter.' It does not say multiple transmitters at one station to separate antennas. It's one transmitter and one antenna.


phase coherence: The state in which two signals maintain a fixed phase relationship with each other or with a third signal that can serve as a reference for each.

Fred

Yes... I should correct/clarify (although this doesn't impact on the main results quoted regarding the required audio power to modulate a carrier):

The peak instantaneous power of a waveform that is the sum of two sine waves... whose frequency difference is a small fraction of either of their two frequencies... is: [(A+B) x (A+B)] /R.

The peak envelope power is (by definition): [(A+B) x (A+B)]/ 2R. I.e. PEP is the average power, averaged over 1 RF cycle of a sine wave whose peak instantaneous voltage (amplitude) is the same as the peak instantaneous voltage (peak amplitude) of the composite/modulated signal.

Note that the above is the defiinition that the FCC uses for peak envelope power: "Finally, let's look at Part 97. The definition of PEP is:

(6) PEP (peak envelope power). The
average power supplied to the antenna
transmission line by a transmitter during one RF cycle at the crest of the
modulation envelope taken under normal operating conditions"

The minimum instantaneous power is [(A-B) x (A-B)] /R.

The minimum envelope power is [(A-B) x (A-B)] / 2R

The time averaged power of the composite signal [the instantaneous power averaged over a time that is greater than 1/(the difference of the two frequencies)] is: [(A x A) / 2R] + [(B x B) / 2R].

Example: if a(t)= 200V sin (2 x pi x 3,800,000t) and b(t)=100V sin (2 x pi x 3,801,000t), and R = 50 ohms... then

The peak instantaneous power is 300V x 300V /50 ohms = 1800 watts
The peak envelope power is 300V x 300V / 100 ohms = 900 watts

The minimum instantaneous power is 100V x 100V / 50 ohms = 200 watts
The minimum envelope power is 100V x 100V / 100 ohms = 100 watts

The time averaged power (as would be measured with a non-peak-reading Bird wattmeter) is [200V x 200V / 100 ohms]  + [100V x 100V / 100 ohms] = 500 watts

Steve (K4HX)

Signals at different frequencies, by definition, cannot be phase coherent. For example, a sine wave at frequency 2f will incur 2 cycles (720 degrees) of phase change for every cycle (360 degrees) of phase change incurred by a signal at frequency f. By definition, phase coherent signals are (for example) phase modulated sinousoids at the same nominal frequency, whose phases a locked.  

When sine waves at different frequencies add.. their time averaged powers (i.e. when the composite signal's power is averaged over a time greater than 1/(smallest difference in the frequencies) add.

Thus the tiime averaged power in a modulated signal is equal to the sum of the time averaged power in each of the components of the modulated signal (or, for random modulation, the time averaged power in the composite signal is the integral of the power spectral density over the band of frequencies occupied by the modulated signal).

Stu

Thanks for the clarification. You are correct. The point I was trying to make was that it is not proper to decompose a signal into some subset of spectral components and claim each of these is separate and distinct in terms of power and do not add to the total power.

I'm still a little confused with the definition though.

phase coherence: The state in which two signals maintain a fixed phase relationship with each other or with a third signal that can serve as a reference for each.

Wouldn't a signal at 2x f maintain a fixed phase relationship (a 2:1 relationship) to the fundamental?

Steve (K4HX)

I guess that one could create a different/extended definition of phase coherence, which would nevertheless be quite useful for some practical engineering purposes. For example, one could create an extended definition in which two sine waves are phase coherent if the phase change of one sine wave... in time T... is exactly N x the phase change of the other sine wave, in time T... where N is an integer.

However, to the best of my knowledge, the standard definition of phase coherence would not extend to the case you mentioned.

Stu

Stu,

Thanks for your reply.

I've studied your comments and see the difference in the math between the instantaneous power and the PEP.  I just noticed in that first equation you were dividing by R* (a number of posts back) which I think was just an oversight.

No wonder most hams, including myself, would have so much trouble trying to understand all this correctly.

Fred

The point I am making is that a power measurement that might be suitable for a relatively simple signal (such as ssb) is not suitable for a complex signal that is indeed made up of 3 distinct and frequency disparate components (signals) such as AM.  

The field day example works very well here.  I'm betting that if one were to "inspect" a field day operation, that only one component (the most powerful generated signal) would be considered when measuring the power of the operation, and the resultant combination of every signal simultaneously generated would in fact not be used.

I'd bet not. The power from each transmitter would be measured and nothing else. Read the section of Part 97. It's one transmitter and one antenna.


This doesn't square with Part 97. Once again, it's one transmitter and by extension the total power output from that transmitter. Parsing the signal is not legitimate. Further, SSB is made up of distinct frequency components. So is FSK, SSTV and even CW.

Steve,

If we're talking about staying below radiation limits at any given spot near many xmtrs, it is the sum of all signals hitting that spot.  How exactly all the signals would be summed up is another story.

Fred

Yes, if you are talking about RF exposure, it's a different story. I thought we were talking about measuring the power supplied to an antenna.

Ok, I bet so  ;)  But seriously,  In the case of the discussion about a field day operation, the term "component" that I used (not quoted in the above example, but it was there elsewhere in the text) means a signal as part of the field day operation, and not a sub-part of a signal.  I should have used a different word.  We are, in fact, in agreement about that.

Using PEP measurements of AM signals "constructs" a single signal out of otherwise discrete signal components, and the result is not a true measurement of the actual power.  This is not the case with CW, SSB, SSTV, and many others where PEP is, in fact, a valid measurement of the actual power.  I am not "deconstructing" a signal, I am in fact stating that I do not want to "construct" a signal that is not "really" there and then make measurements against that result.

If I arranged for a visit to your station by the FCC, you'd bet your license? It's easy to bet when there are no stakes. ;)



How so?

It is not constructing anything. It's the measurement of the total power output of the transmitter. And I don't see how it's any different for CW, SSB or any other mode. They all have "discrete signal components." The easiest example is 2-FSK. There are two easily discerned components. Should we throw away one of these when measuring the power? And what should we do with the discrete components of a CW signal?

I asked Riley Hollingsworth this some years ago at Dayton. The question was if I was in a contest or a DX pileup and controlling 2 transmitters on the same frequency into 2 amps and 2 antennas how much total power could I run?

His answer was 1500W.

Another FCC engineer might have a different answer since Ive not seen it posted in writing.

Carl

Hey, you try it first and let me know how you make out!  ;D

This is how I measure W1VTP for 1500 PEP operation.

Al

PS: Couple of typos corrected

Hey, can Steve really do that ???

Are you lonely Fred?    ;D

The point is that trying to come up with technically incorrect explanations for why one can run more power on AM than as prescribed by Part 97 doesn't stand up to reality (aka a visit from the FCC), let alone technical analysis. If those who think their method of measuing power is correct, let them run it by the FCC. Or let them petition not only the FCC but the ITU and the IEEE to have the power measurement system corrected (since they claim that the current system is wrong).

To claim that AM is some how different from other modes like SSB and CW when it comes to power measurement is incorrect. First, all these modes are AM and all will show an envelope when observed in the time domain. All will show multiple spectral components when observed in the frequency domain.

It's possible that the only thing that will matter is what the FCC employee sees on their meter.

Calibration ought to be simple for those without a peak meter - set the carrier to 375W, and set the scope to 2 divisions. Then when it hits 4 divisions, that is 1500W. This does not mean a 375W carrier is required for operation. It means only that twice the voltage, or 1500W, is a known position on the scope screen.

Unless the black helicopters are circling overhead, the scope ought to agree with the peak meter.

Reducing things to the lowest common denominator is a good way to get as the core. If a simple and transparent measurement shows operation is legal, then there ought never be a problem.

It's not a good bet to guess the FCC employee will give up and go away just because there is no 50-Ohm connector to be seen. There is probably a set of instructions for that situation. I'd like to see their measurement manual.

Should these measurements be made using the forward power sample of a good r-f directional coupler at the tx output connector, so as to ~eliminate the effect of load reflections on the measured values?

Or should the r-f sample source be a non-directional probe so as to include reflections in the measurements?  Of course, this would make the measurements more dependent on the load impedance and line length.

Al (W1VTP)

In my experience, peak reading wattmeters (including very expensive models, and models like the Bird that include a peak adapter board) have too slow a response speed to capture voice peaks... and therefore, are not accurate in measuring the PEP of voice modulated signals. It is typical for these meters to produce a reading that is only around 70%-80% of PEP when used with AM transmitters.

I once had a telephone discussion with a fellow who sells a popular power meter that uses sampling/digital technology. I asked him why he intentionally designed it to be too slow to capture actual voice peaks. He told me that he had to do that so that his meter wouldn't read a higher PEP than competing products (which don't capture voice peaks). If he made his meter fast enough to capture true voice peaks... his sales would drop, because the people who buy his meters want to run higher PEP, and they don't want a meter that discloses their true PEP. I once designed my own peak detecting adapter to work with an expensive digital wattmeter... and it produced a reading that matched the scope measurement. It wasn't easy to make the peak detector fast enough, using analog components, but I was able to do it. The key limitation is the skew rate of the op-amp that drives the integrating capacitor of the peak detector circuit. That op-amp has to deliver enough current to charge the capacitor to as close as possible to the actual voice peak voltage... while the RC time constant of the integrator must be long enough to allow the meter to respond to the voice peaks.

Steve (WA1QIX) might want to comment on the challenges he encountered in designiing the analog peak detectors in his first generation modulation monitor product.  

Using a digital oscilloscope, observing the RF sample... and triggered by the modulating audio... is the best way to measure actual PEP

R. Fry SWL

You raise a very good point. If the antenna is not resonant, I suspect that the FCC would define the power limit in terms of the forward power - the reflected power (i.e. the power that is actually leaving the antenna, neglecting any resistive losses in the antenna itself).

Stu

Stu brings up a good point about accurate peak reading monitors.
The Ham op is not impressed to see those little instantaneous 6 microsecond peaks, indicating that MAX PEP has been attained. They would like to see more of an "averaged" (peak) reading displaying the munkey swinging steady at the so called PEP.
Steve's (QIX) mod monitor showed me that I was way out in left field with my settings, for what I thought was a well adjusted AM TX.

This is slightly off-topic, but pertains to max peaks as seen on a broadcast FM monitor. (1977). College Radio 7KW ERP. And I understand that FM max modulation is bandwidth, but same consideration to FCC rules.
The Belar FM mod monitor, and that special design in the metering, would indicate 80% peaks, and the PEAK light (Real 100% FM modulation) would start flashing. The station engineer, at that time, would count how many flashes in a 5 minute period would occur.
He claimed that was the ABSOLUTE max the FCC would accept, if they came to the station for a visit. And you hope your mod monitor was recently calibrated. (Have the DJ play a song with little high-end audio).
The pre-emphasis for high freq audio would make the peak light flash more. That was the downfall of audio processing back in the 70's. The Audimax / Volumax were about the best in those dayz. The college bought an Orban processor and we could bring the modulation up to 95% looking at the meter, and the same count on the peak light.
Today FM broadcasters can take the entire modulation envelope right up to 100%.
I'm purposely not getting into any RDS or other uses of sub-carriers that would complicate adjusting the main stereo audio on the assigned freq.
Boring story over
Fred

Just to note that the total power output of an FM transmitter into a load Z having sufficient VSWR bandwidth is not a function of modulation, as it is in transmitters using some form of amplitude modulation.

Stu and others:

As far as I know this is the current accepted approach among amateur radio circles who care to measure peak envelope power (PEP).  A refinement might be to use Steve's mod monitor and backward calculate a correction factor or to use more expensive peak reading power meters that cost in the thousands (this expense will be unacceptable among most amateur radio stations and is, in my view, unnecessary).

I have had a lot of contact in my professional life with Agilent, Weinschel and others and this approach is, in my view preferred, to no effort to monitor one's PEP.  After performing the calibration procedure as outlined in my document, I have seen reasonable correlation between the scope and the 43 in the peak reading.  I understand the principle; it's a matter of accepted practice by the FCC in the amateur community.  I prefer my technique over no monitoring of PEP.  My scope shows no discernable flat-topping and reasonable correlation between the scope which I run concurrently with the 43. The degree of accuracy is another topic of discussion and probably does not fit in amateur radio circles.   My next step would be to do some calculation based on the mod monitor that Steve made that I am using or employ yet another crosscheck as outlined below regarding adjusting the duty cycle of a CW keyed transmission.

The point is to do SOMETHING if one is running enough power that might produce more PEP than is allowed in part 97.  To do nothing based on some argument of peak reading equipment inadequacy without checking the degree of that inadequacy is not, in my view, good practice.  It is not preferred to at least to have some idea what is happening with one's AM PEP.

One thing I might do is change the duty cycle of the keyed CW emission to see how the peak reading 43 performs as the duty cycle is decreased - all the while monitoring the scope to see that at least some of the RF output touches the pre-calibrated 1500 PEP point.  I think I will see the same result on the 43 as I am now seeing with a 50% duty cycle of the keyed output.  My concern is that many AMers will hide behind the argument that required equipment is too expensive or would hide behind some undefined inadequacies of currently accepted peak reading modifications of amateur grade power meters.  I prefer making the measurement based on a properly calibrated scope and if I can include a peak reading Bird all the more convenient.  But do SOMETHING!

Al

If so, the only accurate means of doing that is to measure the power delivered to, and accepted by the feedpoint of the radiator.

AM broadcast stations measure power using an r.m.s. r-f ammeter in series with the lead from the matching network to/at the base of the tower.  The unmodulated r-f current into the impedance measured at the feedpoint is set so that it results in the power authorized for that station by the FCC.

When that transmitter is modulated +/-100% by a sine wave, the current reading on the r-f ammeter increases by a factor of SQRT(1.5), i.e., 1.225.

Your post is petty close to correct for AM B'cast. The FCC allows B'cast 125% pos peaks MAX. They are not concerned about PEP ratings for b'cast.
It's us poor AM Ham ops who got screwed into the 1500 W PEP crap. The SSB folks felt we had an "unfair" advantage over them and the 1500 W PEP regulation.
Canada is even more confusing in their rules lately.

And R. Fry, the RF power output of an FM transmitter is constant no matter if it's 10% modulated or 100% modulated. It's interesting to see the modulation of a B'cast FM transmitter on a real spectrum analyzer. The louder the audio, the fatter the bandwidth. And there are times the carrier actually disappears.
link to Bessel Null before ya'll think I'm blabbering:
http://www.fmsystems-inc.com/manuals/BESSELart.pdf
Fred

That is the point I was making, Fred, prompted by the statements in this thread about modulation peaks on a 7 kW FM transmitter.


There is no "carrier" in FM as in AM.  The energy at the assigned FM center frequency can go to zero for some modulation conditions, and that energy is totally distributed in Bessel sidebands on each side of the center frequency.

This can be useful in calibrating an FM modulation monitor by selecting a center frequency null produced from FM deviation by a known modulating frequency.

That doesn't match up with Part 97, which says:

(6) PEP (peak envelope power). The
average power supplied to the antenna
transmission line by a transmitter during one RF cycle at the crest of the
modulation envelope taken under normal operating conditions.


It's power supplied to the antenna transmission line, not the antenna feedpoint. It would seem the FCC is not concerned with the feedline other than what is going into it.

Actually, I did the math and if one measures the carrier accurately and uses Steve, QIX's mod monitor, one can derive the PEP quite nicely.

I have done an Excel spread sheet on it and it works slick. One needs only to enter the carrier power and the positive % modulation. Gotta bounce it off Steve QIX some time.  So, while my previous doc is correct, given Stu's correct analysis of the inadequate response of the peak reading Bird 43, I think I will use Steve's excellent peak reading response characteristics of his mod monitor as a preferred method of measuring PEP.

Al

Based on Steve's (K4HX) latest comment:

"That doesn't match up with Part 97, which says:

(6) PEP (peak envelope power). The
average power supplied to the antenna
transmission line by a transmitter during one RF cycle at the crest of the
modulation envelope taken under normal operating conditions."

I've revised my opinion of how the FCC would interpret their rule in the context of reflected power coming back toward the transmitter from the antenna system. I think they would say something like the following:

"We are going to measure the power being supplied to the transmission line at the point where it connects to the output port of your transmitter. We are going to make this measurment using a directional coupler... and all we care about is the forward power traveling in the direction from the transmitter toward the antenna system. We recognize that there might be reflected power traveling back from the antenna system toward the transmitter. We recognize that there might be resistive losses in the antenna system. We recognize that there might be ground losses associated with the antenna system. We don't care about those effects. All we care about is that the peak envelope power supplied to the transmission line by your transmitter's output port, traveling in the direction from the transmitter toward the antenna, is less than 1500 watts."

In respsonse to that, I would argue that if I use an antenna tuner... consisting entirely of passive components... it should be considered as part of my antenna systems (along with the transmission line between the tuner and the antenna). Therefore, if I can adjust my antenna tuner to ensure essentially no reflected power flowing in the section of transmission line between the output port of my transmitter and the tuner... I can supply the full 1500 watts PEP into that transmission line ... even though there will be points along the other section of transmission line between the output side of the tuner and the antenna in which the forward power exceeds 1500 watts PEP.

I think the FCC would agree that the tuner (consisting of passive components) is just a part of my antenna system... and all they care about is the PEP supplied by the output port of my transmitter, into the transmission line connected to that port, traveling in the direction from that output port toward the antenna system.

Stu

Maybe that is all the FCC cares about for amateur radio service, but my guess is that they are (or should be) more concerned about radiated power.

After all, it is radiated power that produces interference to other facilities.

Why penalize those with high gain antenna systems?

If the FCC is using a Bird 43 they already know it is only accurate to +/- 5% of the full scale reading.

How accurate is the LP-100 peak reading meter? They also claim a true peak hold capability.
http://www.telepostinc.com/lp100.html

Why not just use a dummy load (50ohm) and whatever meter the FCC thinks indicates the full PEP.  This way no need to consider reflected power or any other component of the antenna system.

Fred

OH my!  Let's NOT go down that road.

 :o

And Fred.  Yup

Because a dummy load doesn't radiate as much EM energy as most antenna systems, and may not have the same input Z as a given antenna system.

Probably the FCC would have no concern with anyone operating a ham-band transmitter of any power level into a non-radiating dummy load.

But substituting an antenna system for that dummy load may change the power delivered to the tx output connector when it is connected to the antenna system vs. the dummy load.

What facilities? Perhaps all facilities??? Seems like your using a very broad brush here!

R. Fry,

As you probably know, when the load matches the internal impedance of the generator you get maximum power out to the load.  If I'm using a xmtr that the specs say has a 50ohm output impedance, I would think that a load with a matching impedance is pulling full power from the xmtr.

Now having said this, you're right, maybe my set wasn't made exactly to the claimed spec, and some other load resistance will pull more power than the standard (or commonly used) 50ohm load.

I guess at that point, let the FCC do their testing with an adjustable load and figure out at what load my set delivers the most power out.

So, where does that brings us, back to your statement,  do the measuring with my antenna as the load.  I agree with your point.

If my xmtr has an internal impedance of 50 ohms and I connect a 75 ohm antenna system to it, I'm not going to be able to pull the full capable power from the xmtr.  And before everyone starts to jump up and down, if my antenna system is 75 ohm (cable, antenna, SWR meter) I will have no reflected power (1:1 SWR.resonant). I just won't be getting the maximum power from the xmtr.

In contrast to all of the above,  most of the xmtrs that us AMers are using have Pi network tank circuits which can be maximized for a limited range of loads.

Fred

What do the folks at the FCC's Office of Technology Policy really care about?

I know from first hand experience... having served (several years ago) as a member of the FCC's Technological Advisory Council (If you serve on this Advisory Council, you are not an FCC employee, you don't get paid, and you don't get your travel and living expenses reimbursed) that what the folks in the FCC's Office of Technology Policy really care about is: the inefficient use of many portions of the frequency spectrum. Spectrum is very valuable... and the FCC is promoting (among other things) the increased use of spread spectrum techniques and things like "cognitive networking" to increase the number of users per unit of bandwidth and per unit volume of space.

http://www.fcc.gov/encyclopedia/technological-advisory-council
 

Fortunately, most of our HF bands are of no great interest to other prospective users or the FCC... but we need to "keep our heads down".

If the FCC were to revisit HF amateur radio (which I don't think they are likely to do), they would likely focus on how "empty" the bands are most of the time... and on how high power users with large transmitted bandwidths "unnecessarily"  interfere with adjacent users when the bands are not empty. Most AMers would stand out as sore thumbs with respect to "efficient" utilization of their allocated spectrum.

With respect to converting concerns into rules...

The FCC knows that rules have to be structured in terms of things that the FCC and radio operators can measure... even though what can be measured may not be 100% correlated with what the FCC is concerned about.

You can measure the peak envelope power traveling from the output port of a transmitter into the connected transmission line, toward the antenna system. At least until the present time, it has been very difficult/impossible to come up with a practical/enforceable way to measure of how much interference a particular transmitter is causing.

However, I know first hand, that the FCC and the industry are exploring concepts like "local radiation temperature" as a means of coordinating the use of the spectrum among "cognitive" spread spectrum devices... which will adjust their respective output powers and radiation patterns, in a coordinated way, to minimize interference among shared users of a specified band of frequencies in a specified volume of space.

Compared to those initiatives... our concerns about 1500 PEP limits for relatively short distance AM communication on HF bands would not likely fall on sympathetic ears at the FCC. If anything, the SSB folks would receive a more sympathetic hearing regarding the efficient use of the HF amateur bands... but, again, I doubt if the FCC would spend any time revisiting the use of the existing HF amateur bands.

Stu

 

Most transmitters are rated for the power they can safely deliver to a specified load Z, such as 50 ohms.  But that doesn't mean that the source impedance of the output stage of the transmitter is 50 ohms.  If it was, then the r-f power dissipation within the tx would be the same as it delivered to the load, and the tx never could have a PA circuit efficiency exceeding 50%.

The source Z of most transmitters usually is much lower than the load impedance they are specified to drive.  This permits Class C amplifiers (for example) to deliver 75% or more of their d.c. input power as r-f power to the load.

As SWL points out:

The concept of a "matched load" does not apply (as just one example of when it does not apply) to a complex, non-linear situation like this where:

a. The source is essentially a current source producing a periodic sequence of pulses of (plate) current at some repetition frequency... i.e. a tetrode or a pentode with an RF bypass from screen to ground. Note that a tetrode or a pentode acts like a high impedance source because it delivers essentially a fixed plate current, independent of the plate voltage, provided the plate voltage is positive. In the special case of class A operation, the periodic sequence of pulses of plate current takes the form of an average current plus a sinusoidal current. In the special case of class C operation, the periodic sequence of pulses of plate current consists of pulses whose duration is less than half a cycle.

b. The load (through which the current flows) is required to be a tuned circuit that has a much higher impedance at the fundamental frequency of the current source than its impedance at DC or at any of the harmonic frequencies. I.e. we are trying to maximize the power delivered to the load at the fundamental frequency

c. Across this current source, there is a fixed voltage source in series with the load

d. There is a switch in the circuit that turns off the current source if the voltage across the current source goes below zero volts


I.e. in the case of an RF output stage, the fact that the plate current goes to zero (turns off) if the plate voltage is not positive leads to the inapplicability of the matched load concept. If, hypothetically, the plate current were independent of the plate voltage... whether the plate voltage was positive or negative... then the matched load concept would apply... and the value of that matched load would be very high (because the output impedance of a tetrode or a pentode is very high). Of course, the tube doesn't work that way... it "turns off" (i.e. no plate current) when the plate voltage is not positive.

However, it is true that there is a value of load impedance that causes the voltage on the plate of the RF output tube to drop to zero at one instant during each RF cycle... and that is the value of load resistance that is close to the optimal value for maximum RF power output.

Note that if the tank circuit is tuned to resonance, the time at which the plate voltage drops to its minimum value in each RF cycle coincides to the time when the plate current (i.e. each of the pulses of current that make up the plate current) is at its maximum value (provided the plate voltage is greater than zero volts). Therefore, if the plate voltage drops to a value below zero volts for too long in each cycle, the current source (that represents the tube's plate current) will be able to deliver less and less energy to the tank circuit during each cycle. That is why the value of load impedance (at the fundamental frequency) that causes the plate voltage to drop to zero for one instant in each cycle is approximately the optimal value.

The current passing through the RF load (i.e. into the input side of the tank circuit) is not a sine wave... but it contains a component at the fundamental frequency. The impedance of RF the load (at resonance) is much higher at the fundamental frequency than the impedance of the RF load (at resonance) is at multiples of the fundamental frequency. Therefore the voltage across the RF load is approximately a sine wave at the fundamental frequency. The voltage on the plate of the RF output tube is approximately the sum of a DC (the B+) and a sine wave at the fundamental frequency. The optimal value of the load impedance (at the fundamental frequency) is approximately equal to the DC plate voltage divided by the amplitude of the plate current component at the fundamental frequency.

In class C operation, the plate current is a periodic sequence of narrow pulses. It contains an average value, a component at the fundamental frequency, and components at harmonics of the fundamental frequency. From Fourier analysis, the amplitude of component of the plate current that is at the fundamental frequency is approximately twice the average plate current. Therefore, the optimal load impedance (looking into the tank circuit) in class C operation is approximately: the DC plate voltage / [2 x the average plate current]

Stu  

So how did the topic of "Modulation question" morph into FCC enforcement of power rules, and whether Hams have the right to keep our HF spectrum?  :D

So since the topic is all over the place, what about transmitters like the big Wilcox (96D?) that has OWL output at 600 ohms? Go a step further, and use resonant feeders.

So as hams, are we in any way obligated to have on our transmitters a 50 ohm output port that makes it convenient for an FCC enforcement engineer to connect their 50 ohm peak reading wattmeter?

Jim
WD5JKO

Given that the FCC has never put any limitation on antenna gain in an amateur station, it's clear that they don't care about radiated power. And there are good reasons for this.

Being concerned with interference and thus radiated power makes sense in broadcasting since there are strict limits on cochannel and adjacent channel interference. The public is being impacted by the interference.

In the amateur world, there were never any interference limits within the amateur bands other than things like "good amateur practice." This is because the public and other radio services were not being impacted by any amateur radio interference.

Coverage area is also a concern in broadcasting, again requiring the measurement of radiated power. Coverage is not a concern in amateur radio (at least not in the regulations) other than to the individual stations desires (local, DX, long path, etc) and is often dictated by propagation more than radiated power.

On the other hand, the FCC has historically been concerned with out of band operation, harmonic and spurious radiation from amateur stations because these could impact other radio services and the public.

What does Part 97 say?

Maybe you didn't mean it this way, but you seem to make a contradiction. If the bands are so empty, how could they ever be crowded and how could interference be a problem? Seems there would be plenty of room to spread out and avoid interference.

Interference has always been a part of amateur radio. Part 97 says nothing that would indicate that one should expect interference free operation at any time (bands empty or crowded). Riley Hollingsworth stated something similar in the past. So, I'm not sure the FCC cares about interference, unless it is malicious or due to an improperly functioning transmitter.

I don't think all the HF bands are empty or under used (recognizing that these are fairly nonspecific and ambiguous terms). Over the past 10 years I've noticed a decrease in the density of stations on 75 meters. I think there is less activity, but there is also more room to spread out with the increased phone allocation. In that same period, I've seen an increase in activity on 160 meters. The portion between 7.1 and 7.2 MHz most certainly has more activity than it did before 2009 due the most of the broadcasters departing.

If protecting amateur spectrum is of concern, we should worry about our allocations in the VHF, UHF and SHF ranges. Those are the ones useful to the modern wireless/broadband services that are currently pushing the demand for more spectrum.



It seems to me that comparing single channel, fixed frequency, simplex radio systems using low efficiency modulation schemes (all amateur comms to a greater or lesser extent) to frequency agile, cognitive radio systems using high efficiency modulation schemes will always result in the former looking poor when using spectrum utilization as a metric. In other words, all amateur comms are in the same boat in this regard, AM or otherwise.

That said, given the nature and purpose of the amateur service, spectrum efficiency shouldn't be a metric or most certainly not the only one. See Part 97.1. You could argue that the purpose in paragraph (a) could require improving spectrum efficiency and to some extent paragraph (b) too (although there are many ways to advance the state of the art in radio outside of spectrum efficiency.) I don't think the rest argue explicitly for spectrum efficiency, but I suppose one could always find a way to make these fit if they were really pushing spectrum efficiency.


§ 97.1 Basis and purpose.

The rules and regulations in this part are designed to provide an amateur radio service having a fundamental purpose as expressed in the following principles:

(a) Recognition and enhancement of the value of the amateur service to the public as a voluntary noncommercial communication service, particularly with respect to providing emergency communications.

(b) Continuation and extension of the amateur's proven ability to contribute to the advancement of the radio art.

(c) Encouragement and improvement of the amateur service through rules which provide for advancing skills in both the communication and technical phases of the art.

(d) Expansion of the existing reservoir within the amateur radio service of trained operators, technicians, and electronics experts.

(e) Continuation and extension of the amateur's unique ability to enhance international goodwill.

"(e) Continuation and extension of the amateur's unique ability to enhance international goodwill."


Time to do our part again, Steve - hold court on 75M and 40M into Eu.    Tell the Pascals of the whirl what a good job they're doing...  ;)

T

Yes. Very inefficient use of the spectrum but very fun!

R. Fry,

Thanks for your comments.

I did give some thought that the usual stated 50 ohm output impedance may not be the source impedance.  You correctly and simply point out the usual greater than 50% efficiency of class C finals.  I guess I failed to connect the dots on that one.

Thanks again

Stu,

Thanks for your comments.  As usual, one finds that things are quite as simple as one thinks.

I guess that's why some folks are students and some are professors.

Your input here on the forum is a great benefit to all us students.

I'll re-read and study your comments further later today.

I'm sure, given ample time, I'll be able to come up with more brilliant concepts.

Fred

Stu,

I read through your detailed answer to my earlier statement.  Had to read it a few times.  I can see the relationship of the plate current, voltage and impedance is complex.  I was aware that there are only short duration pulses into the tank circuit.  One thing I finally got an answer to is why we have to divide the plate voltage by twice the plate current to find the plate load.  Something I've done many times calculating Pi Network circuits but never knew exactly why.

Thank you for all the time you give to accurately answer the questions we know-it-alls come up with.

The FCC makes it simple for broadcast folk and their rated power. 

AM's are licensed for their carrier power, with no mention of anything else. 

FM's are rated at their radiated (ERP) power. MUCH more simple.

Well Bob,  as you can see, the threat you started is now on page 4 and growing.  I hope by this time we have answered your question.

BTW Bob, what was the question ???  I seem to have forgot.

Fred

True, as far as it goes. The BC stations must also make measurements on their radiated power or coverage patterns. These are far more involved and difficult than the simple power measurement in the amateur radio realm. Hooking a PEP reading wattmeter to the output of most amateur transmitters, transceivers or amplifiers is about as simple as it gets. I can read my power with nothing more than a glance at the meter. Performing a proof of an AM directional station requires a few orders of magnitude more time and effort.

WAY back when I was in broadcasting as chief engineer, we measured AM station power using an RF current meter at the common point.  The common point was either the input to the phasing equipment (in the case of a multi-tower array), or the antenna matching network.  The meter was often directly at the base of the tower for a single tower system.

For directional stations, we also had to measure the base current of EACH tower in the array.  In the case of a 4 tower directional array, that was a lot of measuring, as it had to be done in not-so-infrequent intervals during the broadcast day (which, in some cases was a 24/7 operation).  This was called the "direct method".  

If your common point RF current meter failed, or if the common point impedance was off for some reason (there are reasons), you were allowed to measure your power using the "indirect method".  This used the power input to the transmitter RF final amplifier(s).

Actual field strength measurements of critical points in the radiation pattern were also required on (as I am trying to recall from so far back) a weekly basis.  These measurement points were specified in the station license.

All measurements had to be logged, and the FCC would inspect these logs, and the entire operation, from time to time (I had a number of FCC inspections during my career in broadcasting).

We also had to perform a "skeleton proof" of directional arrays YEARLY, where you would take field strength measurements at  various compass points (I forget how many - it has been many decades)  and also what is called a "partial proof" at less frequent intervals (years), which required checking the field strength of all of the compass points.  A full proof would involve taking measurements at every compass point, and going out many miles along each compass radial.  That was quite a process and could take a week or more.  This was usually only done if the station was looking to increase power or change the radiation pattern.  It was also required to measure the field strength of the co-channel, 1st adjacent and sometimes the 2nd adjacent stations as well if a change in power or radiation pattern was to be requested.

It was also required that stations measure their common point impedance, and also sweep the common point and note the impedance change at various frequencies near to the station's frequency of license.  This was to verify that the common point impedance was reasonably stable over the audio range.  Many directional arrays are quite sharp, and the common point impedance can vary significantly over the audio range if steps are not taken to compensate.

From that I understand, many of the logging and measurement requirements have been significantly relaxed over the decades.

At the time I was in radio, there had to be a First Class FCC licensed engineer on duty at all times if the station was a directional station, or ran over a certain amount of power (a number which I do not remember).  

Because of this requirement, there were many "6 week wonders" employed at various medium-market directional stations - these were typically DJs, who took a course and studied like crazy to pass the First Class Radiotelephone Operator's exam, and promptly forgot everything once the license arrived.  But, this allowed the station to meet the requirement of a First Class licensed "engineer" on duty at all times.  Of course, if ANYTHING at all went wrong, these guys were absolutely clueless, which is when I would get the 2 AM call that a big storm had put the station off the air!

Oh well, a little history.  We have it easy on the Amateur side as compared to broadcast, although it was easier (and certainly less expensive) when Amateur power was measured as power INPUT to the final RF amplifier - and you used your plate voltage and current meters - rather than PEP output, which requires more sophisticated and more expensive equipment.

To clarify,  non-directional AM stations are not required by the FCC to routinely measure their radiated fields.  Such measurements were/are required only for directional AM stations -- where the critical points typically are in a pattern minimum on a bearing showing that the station limits their radiation in that direction so as to protect the service area of co-channel and 1st-adjacent stations.

The attached clip is an example of this for several stations on 710 kHz (red = full time, green = daytime, purple = critical hours, black = nighttime).  These patterns show the shapes of the radiation launched in the horizontal plane only, and do not show the effects of ground conductivity or the relative fields of nighttime skywaves.

(Continuing with serious thread drift, I know, but some may be interested in what broadcasters need to do vs. ham operators.)

On 60M, the ERP is a factor.   If you have antenna gain you must lower the power so it is 100W PEP ERP.   Since there is no AM allowed, we here don't have to consider it a ham band.

Good point!  No regular field strength measurements for non-directional.  Over here, mostly worked at medium market directional stations (when AM).  It just became part of the weekly landscape  ;)  During that time, we also built a directional FM operation in Worcester, MA.  That was very interesting!!

If you get a chance, check out the radiation pattern for WRKO (680).  50kW out of Boston.  When they change to their night pattern, their signal literally and completely disappears in this direction.  Turns out, a deep null is "pointing" right at my location.  The array is very sharp, and sometimes it is possible to hear the high frequencies of the modulation (when they're on night pattern), but no carrier.

Sure - WRKO patterns attached (click on thumbnail for a bigger image).

The green pattern is daytime and black is night, correct??

The pattern is altered by shifting the phase to the towers (2 or more).

Is the phase altered by changing feed line length in some way or is some other method or equipment use to do this?

Fred

Correct.

Typically, directional AM broadcast stations change patterns by using an alternate configuration of the power dividing and phasing networks at the common point of the array.   These networks usually are installed in a cabinet assembly inside the transmitter building.

The same transmission lines are used from those networks to the matching networks at the base of each tower regardless of the day/night directional pattern that array produces.

Here are the values for the WRKO array:
  
           Twr 1 (Ref)      Twr 2           Twr 3
    DAY
Field         1                   0.9               0.5
Phase       0                   40               104
  NIGHT      
Field         1                 1.904               1
Phase       0                  55.8             104

Also some directional AMs use higher transmitter powers during the day than at night, and/or are non-directional day, and directional night.

An exception in this category is KFMB in San Diego, which is 5 kW non-directional day and 50 kW directional at night.

I can see WCAP 980 towers from above the tree line about 8 miles away and when they switch to night power/pattern they are in the noise unless there is snow on the ground.

WCAP is directional with different patterns day and night, but they use 5 kW transmitter power in both cases.

Their nighttime pattern has greatly reduced fields toward the west compared to their daytime pattern.

Look up any station here.

http://www.radio-locator.com

Steve I feel your plain on doing directionals. I had an 8 tower 1% array in Atlanta (680) that was fairly stable. (It had to be, we protected WMPS in Memphis and WPTF in Raleigh.)

Seems to be quite a few 6-8 tower arrays on that frequency. WCBM in Baltimore runs a six tower array and has a very sharp pattern at night. I've driven by their tower site in years past when I lived in that area. I wonder who the lucky guy is that has to keep that thing in spec?

Imagine the pain of a 12-tower array...

My friend Dave Hultsman had a 12 tower for the old KLIF in Dallas. And I think there is one in Florida that has twelve.

Im to the right of Nashua on the pattern map and right on the red line (just about a perfect due North from the towers) but when they switch the pattern it is in the jumble of noise. On a good S meter it looks like more than 40dB down. With snow on the ground I can hear them fine but weaker.

Considering that going from 15W to 1500W is only 20dB it puts things into perspective.

When driving on I-495 when WRKO switches they are unreadable to the West at I-290 and not useable again until almost at Lowell. Daytime they are gone shortly after I get on I-84 off the Pike. Its a huge difference from the coverage of WBZ and a single tower in a salt water marsh.

I seem to recall that, at least at one time, WBZ had a Westerly cardioid pattern - "beaming" over the city, and the remainder of the US (the pattern looked like something that could easily be created by a 2 tower array).  I haven't looked at this in many years, and they may have switched to a single stick.

That site and some eyeballs when driving by say one tower. They sure put a good signal down the Med when I was fighting for the USA ::)

Yeah, they've probably changed the antenna setup since the map I saw of their pattern was made, and that map was not new!!

According to current FCC data for WBZ, they are using two towers, DA-1.

Tower 2 is located 90^o east of tower 1, and driven with field = 1 and phase = 86.4^o.

Their pattern, and a Google view of their site are attached.

My bad, I was thinking of another station down on the South Shore.

That's the pattern I saw on the map more than 40 years ago.  I guess it's still the same!!  They really have a strapping signal most of New England.

And then there is this three tower dog leg array about a 1.5 miles from my house:


http://www.radio-locator.com/cgi-bin/pat?call=WMT&service=AM&status=L&hours=N

They have some interesting phasing and current partitioning to each tower for their nighttime pattern.

Phil - AC0OB

At an azimuth bearing of zero degrees and 1 km from the center of the array, FCC data for WCAP shows a field of 805 mV/m day and 48.8 mV/m night.

This is a field reduction of 24.35 dB, which means that WCAP radiates only 0.0037X as much power on that bearing at night as they do during the day.