NPC for AM Transformer radios is HERE... Tubed, Solid State, etc. ANY high level TX

[Toll_Free]

View attachment 3954

Here is a scope display of an AM carrier overmodulated by a continuous tone. If, as you claim, under these conditions the final collector voltage is reduced to zero, and the carrier is cutoff, then explain to me HOW THE UNIT CONTINUES TO TRANSMIT, AND HOW THAT TRANSMISSION CAN BE DETECTED BY A STANDARD AM RECEIVER IF NO CARRIER IS PRESENT. Common sense and logic tells me that the carrier is there.

The misunderstanding comes from the way a scope displays a modulated AM signal. There are three components to an AM signal: the carrier, the upper sideband, and the lower sideband. However, a scope cannot display these components seperately in a single trace. What the scope does is to ADD THE THREE SIGNALS TOGETHER AND DISPLAY THE VECTOR SUM. What happens is that as modultion % increases, the negative modulation peaks are subtracted making it APPEAR that the carrier is varying in amplitude. When modulation exceeds 100%, it APPEARS as if the carrier is cutoff, but this is not the case. If it was cutoff, transmission would cease.

If you look the conditions in the scope trace shown above on a spectrum analyzer, it will show that the carrier is present when overmodulated, and does not change in amplitude.

BTW, I posted the AMwindow link because I thought Steve did a good job of explaining how the circuit works, and that members might like to look at the additional info on the site.

- 399

Sorry, but I bet you don't have a spectrum analyzer. Because you would see, with a transformer or other high level modulated AM signal, the carrier DOES drop when overmodulated with a NON NPC or other type of assymetrical limiter in it. As I stated, I HAVE one, and have actually WATCHED it. Nor are you taking into account that NUMEROUS types of transmitters exist in both the CB / HAM and the commercial world that have controlled carrier, which is what BDW and Eitner mistakenly call NPC. For it to be TRUE npc, the cap would also need a resistor and a diode, so that it ONLY worked on the negative peaks....

See the little flat lines in between the "overmodulated" waveform? That's NO RF, COMPLETELY CUT OFF, ETC. 0 volts is 0 volts, my friend. PLEASE tell me HOW much current your little final transistor can sink to put power out at 0 collector volts (0 RF volts MEANS 0 collector volts. That little line in between the modulated peaks positive and negative is 0 volts). These whole laws of physics get in the way of Channel 19 speak a LOT, but if you do the math, or look at it GRAPHICALLY, then you would see how it really pans out, pun intended...

ALSO, you would see with the spec an that as SOON as the carrier level starts to waver, the AM bandwith (what is really called IMD, what a LOT of CBers confuse with "loudness") REALLY gets wide in high level systems. It's not AS bad with low level, as either the balanced modulator will phase invert the 0 carrier condition, or if you are modulating a driver stage typically you still pass the entire modulated signal through an IF filter of one kind of another... This is one reason why even when you overmodulate the HELL out of a 'ham rig', typically you'll only sound like shit on frequency and within the passband of the IF filter.... They run the tx 'backwards' through the receiver filters. This is another reason the ham radios sound 'pinched'.... This, and ALC issues on AM.

It's a hard subject to grasp, unless you actually HAVE the equipment there, or can take a look with the scope at the modulated DC being fed to the final and driver in the 29 style transmitters. Again, HOW can your transmitter be putting out ANY power when the modulated stage(s) see 0 volts?

(and to answer your question as to WHY you can still hear someone when this condition happens, the reason is this: The RF has AUDIO superimposed on it. 27 million times a second is a LOT more often than the typical 2400 times a second (2.4 kc, typical AM or SSB passband, or narrower in a good TX, and typical of the peakiness in the D104 or other power type mics) If you take a look at the RF, and not the AF, on the oscope, you'll also see this..... However, it takes a scope a LOT faster than 30 mhz to be able to actually WATCH what is happening to the carrier. A scope with at LEAST double the passband of the actual circuit being monitored is what most people say to use. BUT, if you DO have a scope fast enough, you'll see that you can actually COUNT the individual rf cycles, and see that if you vary the modulated tone, you can figure out the tone frequency by the amount of RF pulses that get "swallowed". Simple division. As long as you know the CARRIER frequency. You can also do this with multiplication, if you know the modulating frequency, but it isn't NEAR as accurate. In other words, if you adjust your scope to look at the actual CARRIER frequency, instead of the modulated signal, then you'll see individual pieces of the RF carrier 'disappear' on the scope trace.... And those pulses will CHANGE depending on the frequency of the modulated signal, as well as the carrier).

This is pretty complicated, and a LOT of people DON'T get it.... BUT, if you can find the equipment to actually test it out, you'll see what I'm saying is actually true. A 100 mhz digi hold scope would be adequate, but not cheap.... And spec-ans ain't cheap regardless. I'm lucky to have an IFR, which has a Spec An built into it, as well as just about EVERYTHING else needed to adjust and align a radio, transmit or receive.

--Toll_Free

 

[MrSuburban]

I have a question what is this mod actually supposed to accomplish to a 29?

Can we get a before and after audio sample from someone that has done the mod. All the NPC mods I have heard weren't very impressive not sure why this one would be much different.



"The Diode "Super-modulation" Modification

One popular change is the addition of a silicon rectifier diode, or a group of diodes in a fancy circuit, in the modulation transformer secondary. The thought or claim is the diode or diode network limits negative peaks and prevents splatter. It does this by preventing the anode from going below zero volts on negative modulation peaks. This is sideways-thinking for two major reasons:1.) Going to zero-carrier is actually not what causes splatter or excessive bandwidth. The slope of waveform abruptly changing, going in a new direction towards a straight line, causes the signal to get wide. It's really a "Fourier problem", where the rapid change in slope requires high-order harmonics to produce the waveform.

2.) A plate modulated tetrode tube, contrary to what we might assume from causal understanding of plate modulated stages, reaches zero carrier long before the modulated high voltage reaches zero volts. The diode limiter will limit too late, unless the modulated stage uses a hard class-C low-mu triode in the PA.

anode supply voltage johnson ranger plate modulated

This is the modulated high voltage supplied to a Johnson Ranger II, along with the RF envelope in the second trace.

Line A was set for zero anode voltage

Line B is approximately 40 volts positive for the modulated anode supply voltage (200 volts per division for this channel with zero voltage set at scope graticule A)

Point C shows the RF envelope cutting completely off (over-modulation) even through the modulated high voltage never reaches zero volts.

Peak anode voltage is around 920 volts, operating anode carrier average voltage is 510 volts. This includes modulator secondary voltage drop.

Notice the high voltage does not have to double on positive peaks, and does not go to zero on negative peaks even though the stage is slightly over-modulated. This is because the modulated stage is a tetrode, NOT a triode! The modulated stage has both screen and plate modulation applied, as all tetrodes require. Triodes would require the anode voltage going to zero, and the anode voltage doubling for 100% modulation. See amplitude modulation for an explanation of why this occurs.

The modulation in the Ranger is very linear and has very low distortion with any "negative feedback" or other tricks.




While amateur radio reference materials oftentimes tell us negative anode supply voltage causes splatter, that is a gross oversimplification. In an AM system, the device being modulated actually must behave as a linear response mixer. The audio in effect is the signal, the carrier is the local oscillator. It isn't the fact the PA receives negative voltage that causes splatter or over-modulation, it is the fact the "local oscillator" or mixer is shut off abruptly, with a rapid transition in waveform slope, as the PA reaches zero output. We can "go negative" as much as we like with anode voltage and the energy in splatter does not increase, UNLESS we change the angle of slope abruptly at the zero-crossing transition.

Just like with a CW envelope (CW is really 100% modulated AM), the rate of level change or slope angle at any point in the envelope sets the bandwidth. Moving the diode from inside the tube (the cathode-to-anode path is the diode) to the outside of the tube doesn't modify or correct the slope errors. It does not allow us to reach 100% negative peaks without generating splatter. The carrier still reaches zero-volts with an abrupt waveform slope transition, and splatter is just as bad, whether the diode is inside the PA tube or moved to the outside.

What the authors and proponents of most "diode limiter" super-modulation modifications intend to do is produce a negative peak limiter. They think this allows the modulator to hit the PA stage hard with audio and produce high positive peaks without splatter, but they miss the most important points! Properly designed negative peak "hard-limiters" must always be followed by a suitable low-pass filter. The low-pass filter rounds off transitions, and causes the carrier to gently slope to zero on negative peaks. This gentle slope, caused by the low-pass filter, limits the maximum frequency of distortion products and the transmitter's bandwidth (although it does nothing for in-band distortion). If the low-pass audio filter is omitted, the diode or diodes do nothing to constrain bandwidth. It's a totally useless mod without a low-pass filter!

I actually spent considerable time with my spectrum analyzer (it makes direct measurements of adjacent channel power) and my Viking Valiant, and found no matter what I did with a diode or diodes in the modulation transformer secondary, there was absolutely no reduction in adjacent channel energy for a given level of audio. The diode, no matter how configured, did not make bandwidth better.

What did allow me to run asymmetrical peaks was a simple mod to the 6AL5 clipper in the Valiant. I disabled one section of the dual diode, leaving the clipping effective only on one audio polarity. Since this "hard-limiter" is followed by a good 3.5-4kHz low-pass filter, the transition is rounded and off-frequency energy caused by clipping is attenuated. I could limit negative peaks to 100% in my Valiant and have 120% positive peaks with very little increase in bandwidth, but I had to do the clipping before the low-pass filter."


http://www.w8ji.com/Johnson audio mods.htm

 

[Shockwave]

This is a good debate and valid points are being raised on both sides. Toll Free is correct that when you modulate an AM carrier while watching the spectrum analyzer, you can absolutely see the carrier go down as the energy increases in the side bands. If you zoom in with a narrow enough bandwidth you can even see the modulated waveform being imposed on the carrier.

When I read what Tom W8JI wrote, as usual I find myself in agreement with him. This man is a walking encyclopedia of RF knowledge. The negative peaks of an overmodulated waveform do shut the carrier off. Unfortunately keeping the carrier on at any level will not correct the harsh interference to adjacent frequencies.

As I wrote earlier it is the transition into or out of the clipped peak that causes the majority of interference. To reduce interference we must alter the slope of this transition to remove the harmonics the sharp transition creates. It's more then just holding the carrier on during the negative peak.

Simply keeping the carrier alive has no change on the sharp transition the modulated waveform creates. You're negative peak will still be cut off just as sharply as if there were no carrier. The carrier is also cut of on the negative peak at the frequency of the extreme modulation. This is at an audio rate not the RF carrier frequency.

The properly designed negative peak limiter must be followed by a filter that removes the clipped DC component from the modulated AF before it's mixed with RF in the final stage. Otherwise you have the same splatter even if you manage to keep the carrier alive during the negative peak.

 

[loosecannon]

thank you toll free for your explanations.
i always wondered why the dial a watt method never took off with the internet crowd.

seems to me that what this circuit will allow you to do in the context of a CB radio is to run the AMC just a bit "hot", like around 120% to 150%, and by doing that in combination with this circuit; a perceived loudness increase will be noticed on the receiving end without as much of the associated distortion that would be present had the circuit not been in place.

i know thats quite a sentence, LOL, but am i getting this right?

if i am, then we need to watch the direction this thread is taking.
this should not be turned into a debate on the merits of NPC and its affect on adjacent channel bleedover.

rather, since this was posted in the CB section, maybe we should direct the discussion more towards the issues that a typical CB'er cares about.

like, is my radio louder with this circuit in it than it was without it?

is my transmitted audio perceived as loud but not pinched, muffled, or otherwise noticeably distorted?

how about some before and after vids!!!
what, no one has a 29LTD lying around? LOL
LC

 

[longhaireddwb]

I'm going to give it a try. This is just to try out the loudness issue. A lot of drivers (most of them) just want a loud radio and dont care what it really sounds like. Even distortion isn't a problem with some of them as we all know and hear. I like a clean sounding radio myself. If this works and bumps the modulation up higher and keeps the audio cleaner than the cap and resistor mod than its worth trying.

I'd still like to see the before and after scope pics so please someone come up with this. There' so many 29s out there someone got to have another one laying around to test on.

 

[Toll_Free]

Although what W8JI writes is COMPLETELY correct, it has nothing to do with a typical collector modulated CB radio, since they DON'T have a screen grid.

You'd need a dual gate mosfet to approximate the same thing as a tet in the final. I only know of 3 CBers that would spend the money on a final big enough to be able to accomplish that, and none of them would do it knowing they couldn't hit 150+ positive peaks.

This circuit eliminates ONE type of problem encountered in the AM world, and that is baselining the modulator... It will PREVENT that, if properly implemented.

You CAN use it with NO limiter, but I wouldn't... I'd open a 29 wide open via the AMC pot, and then use this circuit ADJUSTED on the scope with the MICROPHONE that will be typically be used to eliminate the 0 carrier issue.

YES, their are companies that used this or similiar circuits commercially... They also introduced a 'splatter filter' after this to keep the problems the clipped AC mod waveform introduces out of the actual modulated signal: Problem is, when the splatter filter develops a problem (and they ALL do for one reason or another, it seems), they take the mod xformer out. With a typical CB, not a big issue.... When talking a 1-10 Kw mod xformer, it gets spendy..... Hence the reason this circuit became "the standard", and you just don't hit the mod circuit so hard.

If you REALLY want to make it LOUDER THAN HELL, you can use this circuit, then use something like Lou Franklins speech processor.... I modify them so they use a shottkey on the negative clipper side, and a standard rectifier diode to clip the positive side. This gives almost a 2:1 asymmetry to the modulating signal, and makes for a LOUD AS HELL xformer driven radio. This will allow you to run insane positive peaks, and the baseline preventer will keep the very few negative peaks that actually get through away from the final and driver... It ALSO allows you to NOT have to run a splatter filter on the output, since you only hit a half-cycle on the NPC circuit once in a great while, everything adjusted correctly

--Toll_Free

 

[Shockwave]

The point W8JI makes about the modulation levels in a tetrode are the only points that don't apply here. Because of the bias and screen modulation used on a tetrode it will reach RF cutoff long before the plate voltage hits zero. If this circuit worked, it could be adapted to something like a Tram D-201 by raising the keep alive voltage. Without reshaping the modulated waveform through a filter and removing the sharp transitions caused by clipping, you haven't removed the harmonic content responsible for the adjacent frequency interference.

It is also my understanding from theory and practice that you cannot effectively accomplish NPC in the speech amplifier stages unless they are all DC coupled to the modulator. Once the AF passes through a capacitively coupled audio stage, you lose the baseline reference point. Without DC ground as the reference point, the asymmetrical AF will begin to average itself out and appear symmetrical again after a DC blocking cap.

When we look at some recent home brew rigs that can use NPC like the class E AM rig kits on 75 meters, you will see that NPC is active in the last stages of the modulated DC feeding the finals and this is where the hash filter is placed to remove the harmonic content. All things considered, this 3 diode mod does not accomplish the main goal of NPC. NPC must remove the sharp harmonic filled transitions of the negative peak and assumes the modulator has the capacity to reproduce a larger positive peak without flat topping.

 

[Toll_Free]

Shockwave...

You can't even keep a signal faithful with DC coupling. Their is the 'repeaking' effect that happens in ANY amplifier stage, especially one that is NOT perfectly flat in freq response (and find one that is, we BOTH get rich!!)


The POINT is, you can GREATLY eliminate the lower AF waveform and keep the peaks on the positive side in greater amplitude. Yes, the amplifier stage(s) after processing WILL cause repeaking, but they are in NO way in even similar amplitude to an unprocessed signal.

If this was NOT the case, their would be NO cause for 'flipping phase' on the microphones, and seeing a subsequent increase or decrease in PEP Pout.... BUT, there IS. Nor would a clipper or compressor work. BUT THEY DO.

My point the ENTIRE time was this.... This circuit will eliminate the baselining common with overdriven modulator stages. It's NOT meant to be driven into saturation constantly, it's meant to be used on a PROPERLY functioning rig, and one that has been adjusted correctly. No, it ISN'T as easy as throwing a 10uF cap in series from the audio lead to the carrier control, nor does it effect the same thing... HOWEVER, if you think you need to drive every amplified stage to saturation 100 percent of the time in an AM modulator and RF stages, you've got issues beyond what ANY NPC type circuit will do.

It's been commonly accepted by the SAME guys that brought the Class E stuff that this circuit DOES function as intended, but it isn't intended to be in conduction 100 percent of the time. Clip your neg peaks, check phase with an oscope, and use this circuit for the occasional peak that the audio circuits don't hit.

If you DON'T have audio processing, just leave the AMC intact.... BUT open it up. This circuit will produce a HELL of a lot less distortion (both even and odd order) than a typically AMC disabled radio.... Nothing will do that, not even Eitners NPC...

Run things wide open, and GIGO. Use limiters like this one to prevent the occasional peak, and the splatter we see on the airwaves would be greatly eliminated... Then again, you'd have to have 'techs' that could actually do the math associated

--Toll_Free

 

[unit248]

loosecannon asks two important questions:

"like, is my radio louder with this circuit in it than it was without it?"

I think the answer to this is: maybe.

An AM carrier that's modulated to beyond 100% is still "louder" than one that's modulated to no more than 100% whether you use NPC or not. The difference is that with NPC, your signal is cleaner, and your audio will not be as distorted.

Without NPC, your over-modulated negative peaks turn into square waves. If you look at the mathematical definition for a square wave, it's the sum of a bunch of other waveforms, which is a fancy way of saying that square waves are chock full of harmonics. Each of those harmonics is going to result in additional RF energy being generated above and below your fundamental frequency (i.e. splatter on adjacent channels).

Note that the worst case is to push the modulation up so high that you flatline not just the negative peaks but the positive peaks too. This turns your audio completely into a square wave (not just at the negative peaks) and results in the worst possible splatter and the nastiest sounding audio.

Of course, some people probably like that -- it's that whole: "when I key up the mic, I want them to hear me on all 40 channels" sort of mentality. And if that's what floats your boat, well fine. But it has drawbacks (not the least of which is that it makes you look like a bit of a jerk).

Consider the typical "keydown" situation. Station A and station B both have identical antennas, radios, mics and 1000 watt amplifiers. Station C is some distance away, with a receiver, on the same channel as stations A and B. Station A and station B both have their radios set up to produce, say, 150% modulation (with enough positive peak headroom to avoid saturating the finals), but station A uses a well designed NPC circuit (with filtering even) and station B does not. Now station A and station B both key up and shout their fool heads off while station C listens. Who will be heard?

Consider that station C's receiver is designed with an AM passband of, say, 5 to 6Khz. It has filters to block everything else out. The whole idea of course is that you only want to hear what's on the current channel, and reject everything adjacent to it. Also consider that a good amplifier will have output filtering to choke off spurious emissions. (11 meter amps probably don't, but that's a separate issue.)

Station B, without using NPC, is spewing 1000 watts worth of RF, but because of the flatlined negative peaks, a lot of that power is being wasted in the generation of distortion products. That is, instead of all the RF energy being focused onto the current channel (the fundamental frequency), only some of it is, while a large percentage is being emitted on other frequencies above and below the fundamental.

Meanwhile, station A, with NPC, is not exhibiting the same flatlining, which means his station is emitting fewer distortions products, so more of his 1000 watts is staying on the fundamental frequency.

Since station C's receiver is going to filter out the splatter on adjacent frequencies, station A will have an advantage, because more of his signal will fall inside station C's receive passband. Meanwhile, station B is scratching his head wondering why, with his watt meter showing 1000 watts of output, he didn't get heard.

But that brings us to:

"is my transmitted audio perceived as loud but not pinched, muffled, or otherwise noticeably distorted?"

I think the answer here is: it depends.

In terms of "loudness," yes it will sound louder without sounding pinched or muffled. The issue of distortion is a little more complicated though.

In an ideal situation, the NPC will result in negative peaks being compressed, but they will maintain the nice, smooth curve of a sine wave. While keeping the waveform smooth and curved avoids the RF signal distortion caused by clipping, you are still altering the audio waveform compared to how it looked when it entered the microphone. In the strictest sense, this is also a form of distortion. However, It's not anywhere near as bad as the flatlining distortion you would have without NPC, and I'm not sure how it will be perceived.

If you listen to a 1000 hz sine wave tone, and a 1000 hz square wave tone, they will definitely sound different. But here, you should still have a sine wave, just with asymmetrical amplitude. Whether or not you'll be able to notice the asymmetrical amplitude depends on the quality of the receiver on the other end and the ear of the listener. But given that broadcasters have resorted to this trick in the past, and that they would be more motivated than anyone to maintain good fidelity, I would think that if you did it correctly, it wouldn't be very obvious to the listener.

I think a lot of this depends on your goals and how much effort you're prepared to put into it. Remember how I asked about maintaining linearity? If it was me, I would do the following:

- Determine the linear operating range of my radio (i.e. how much power can I transmit without pushing the drivers/finals into saturation)

- Implement the NPC mod, with filtering if needed, so that the negative peaks not only don't flatline, but also retain as much of their sinusoidal properties as possible

- Adjust the audio input so that positive peaks achieve greater than 100% modulation _without_ exceeding the upper bound of the RF stage's linear operating range

(Note that if an amplifier were involved, there would be additional complications since I would want to avoid driving the amp into saturation too.)

I would do it that way, because I'm more interested in fidelity than I am in being super loud or cranking out the most watts. If I have to choose between 125% modulation with perfect linearity vs 150% modulation with some saturation, I'd choose the former. Of course, someone who's more interested in just "keying the maul down" would choose differently.

Unfortunately, my radio (Kenwood TS-850) uses a balanced modulator for AM, so none of this does me any good.

-Bill

 

[cbradiomagazine.com]

I love this video - which pretty much shows that while we know the cap/resistor mod in the 29 lowers deadkey and gives a little swing it can get ugly quicky. And even worse when you max the AMC and add a power mike.

Modulation Video by Jonesing - Myspace Video

When people do the cap/resistor mod and max the AMC they need to realize they can run a stock mic and the dynamike at 12 o'clock and get pretty good sound of out it. Still won't look super clean on the scope but most people aren't worried about that.

But from listening in DX land I'm pretty sure everyone turns the Dynamike to max.

 

[Shockwave]

If the negative peak is already being clipped off flat by RF cutoff and you are able to keep some level of carrier present, all you did was move the flat portion of the waveform farther into the sharper slope. The 3 diodes do not in any way I can see accomplish "negative peak compression". I'm also having a difficult time understanding how it can reduce adjacent channel interference when it does nothing to eliminate the flat DC component.

The clipper or compressor does not lose effectiveness when passing through capacitively coupled stages because it effects both the positive peak and negative peaks equally. It does not require that you maintain a stronger positive peak through the coupling caps. I agree that you can see differences in some transmitters from flipping the phase. It is also true that some of the asymmetrical properties are lost through the coupling caps.

As I understand it, this interference is not caused simply because the RF carrier is pinched off. It's the slope of the turn off and on points that causes that harmonic interference within the audio. Until you filter out those 90 degree corners on the modulated waveform and make them rounded, you still have the splash. Simply turning the carrier on and off at audio rates does not by itself mean it's increased bandwidth.

Consider the typical CW transmitter and the common problem of key click interference. The slope of the turn on and turn off points are rounded on a good CW transmitter to avoid the same type of interference. Having sharp corners in your modulated waveform creates the same problem. To reduce interference, we must round those corners off with a filter after the clipper.

 

[unit248]

If the negative peak is already being clipped off flat by RF cutoff and you are able to keep some level of carrier present, all you did was move the flat portion of the waveform farther into the sharper slope. The 3 diodes do not in any way I can see accomplish "negative peak compression". I'm also having a difficult time understanding how it can reduce adjacent channel interference when it does nothing to eliminate the flat DC component.

The clipper or compressor does not lose effectiveness when passing through capacitively coupled stages because it effects both the positive peak and negative peaks equally. It does not require that you maintain a stronger positive peak through the coupling caps. I agree that you can see differences in some transmitters from flipping the phase. It is also true that some of the asymmetrical properties are lost through the coupling caps.

As I understand it, this interference is not caused simply because the RF carrier is pinched off. It's the slope of the turn off and on points that causes that harmonic interference within the audio. Until you filter out those 90 degree corners on the modulated waveform and make them rounded, you still have the splash. Simply turning the carrier on and off at audio rates does not by itself mean it's increased bandwidth.

Consider the typical CW transmitter and the common problem of key click interference. The slope of the turn on and turn off points are rounded on a good CW transmitter to avoid the same type of interference. Having sharp corners in your modulated waveform creates the same problem. To reduce interference, we must round those corners off with a filter after the clipper.

A couple things:

The flat line where the negative peaks bottom out during over modulation is not really a "DC component." That line is at 0 volts. There's no energy there at all. (If you actually end up with DC voltage getting to your antenna you have a bigger problem than over-modulation.)

Your understanding that the real problem is the square shape of the clipped negative peaks is correct. However, I think you're under the impression that the NPC circuit still results in the negative peaks turning flat -- they just turn flat before dropping all the way to 0 volts.

That's not what happens. The intent is to squash the negative peaks so that their amplitude is less than the positive peaks, while still allowing them to retain a sinusoidal shape (which avoids the splatter problem). I think the idea is that the diodes conduct some energy through the resistor only during the negative peaks, which allows the resistor to become part of the modulation load. In other words, this _dampens_ the negative modulation peaks: it does not _clamp_ them.

In the process, some audio energy is wasted in heating up the resistor, which is why Toll_Free said it should be rated to handle "1/20th of the total audio power fed to the final."

Imagine a sine wave of +/- 10 volts. Now imagine a second sine wave of +/- 5 volts. Also imagine that both are oscillating at the same frequency. Now cut the +/- 10 volt sine wave in half at the 0 volt line and discard the negative voltage peaks. In their place, paste in the negative peaks from the +/- 5 volt sine wave, lined up so that the crossover points meet. What you have now is an asymmetrical sine wave, where the positive peaks have a higher amplitude than the negative peaks. It's not immediately obvious, but that's what the NPC circuit should do to the modulation waveform.

-Bill

 

[Shockwave]

A couple things:

The flat line where the negative peaks bottom out during over modulation is not really a "DC component." That line is at 0 volts. There's no energy there at all. (If you actually end up with DC voltage getting to your antenna you have a bigger problem than over-modulation.)

Your understanding that the real problem is the square shape of the clipped negative peaks is correct. However, I think you're under the impression that the NPC circuit still results in the negative peaks turning flat -- they just turn flat before dropping all the way to 0 volts.

That's not what happens. The intent is to squash the negative peaks so that their amplitude is less than the positive peaks, while still allowing them to retain a sinusoidal shape (which avoids the splatter problem). I think the idea is that the diodes conduct some energy through the resistor only during the negative peaks, which allows the resistor to become part of the modulation load. In other words, this _dampens_ the negative modulation peaks: it does not _clamp_ them.

In the process, some audio energy is wasted in heating up the resistor, which is why Toll_Free said it should be rated to handle "1/20th of the total audio power fed to the final."

Imagine a sine wave of +/- 10 volts. Now imagine a second sine wave of +/- 5 volts. Also imagine that both are oscillating at the same frequency. Now cut the +/- 10 volt sine wave in half at the 0 volt line and discard the negative voltage peaks. In their place, paste in the negative peaks from the +/- 5 volt sine wave, lined up so that the crossover points meet. What you have now is an asymmetrical sine wave, where the positive peaks have a higher amplitude than the negative peaks. It's not immediately obvious, but that's what the NPC circuit should do to the modulation waveform.

-Bill

It's not that the antenna would have to cope with any DC component from hard clipping. It's things like the primary of the modulation transformer that would have to cope with that problem. When the AF modulation is clipped on either peak, it's placing that DC component on this winding which has an impedance closer to a short at DC.

At least we agree the interference problem is created by the shape of the clipped waveform. In particular, the fast rise and fall time of the clipped waveform and it's sharp corners at the clipped peaks. The intent with NPC is to squash or compress the negative peak but I still fail to see how any resistive loading is going reshape the clipped waveform?

It seems the creator of this circuit misinterpreted this as meaning just keep the carrier alive and paid no attention to waveform. I hope I'm wrong here and this simple 3 diode circuit is the easy way to NPC. All it will take to convince me is a short video. You know what we would like to see.

The scope showing a clear indication of carrier level going directly into something like 150% positive peaks and 90% negative peaks with no flat lines anywhere in the waveform. I don't think this negative peak clipper can accomplish this goal without being followed by at least a inductor and capacitor (LC) filter network.

 

[jammerdave]

switched power

In the diagram, the 50 ohm resistor goes to "switched" power. Is that switched on keyup, or just switch with radio power? Thanks!

df

OK, here's the 'final' circuit. This is for the 'standard' model, no variable to mess with.

The unlabeled diode is the existing 'anti spike' diode all radios have coming from the modulation transformer. You wrap the two other diodes around it.

This is the way the circuit works.

The diodes and resistors form a divider that allows the D2 to ONLY conduct during NEGATIVE mod peaks.... The HARDER the negative mod peak, the HARDER the diode conducts. During this conduction, it allows B+ (tube speak... Same as "the hot lead" in 12 volt stuff) from the resistor and capacitor to conduct through D2. By varying the resistance of the approximately 50 ohm resistor, you can vary the amount of negative peak limiting that the circuit provides.

I did NOT invent this... This is an adaptation of different peak limiters.

This circuit does NOT work 'just like npc' on the 148 or other class A series limiter radios.... IE, it DOES work at an audio rate, but it will NOT prevent negative 'peaks'... It prevents the carrier from being COMPLETELY pinched off. THIS is what causes the MAJORITY of splatter and garbage. This circuit will NOT clean up the waveform... I can add a splatter filter (to clean up the pos and neg peaks), but those add distortion of their own.

Also, the approximate 50 ohm resistor. It will run COLD if you don't 'bang on the limiter', so to speak. The MORE negative peaks you add (ie, the HOTTER you run the audio gain / power mic / etc), the HARDER this resistor has to conduct..... With a stock mic, a pair of 1/2 watt resistors in parallel (100 ohms) can be held in the hand. With the 'typical' levels a power mic produces, they will get WARM....






FOR THOSE THAT WISH TO MODIFY THE CIRCUIT!!!

If you want variable negative peak limiting, then take the "approximate" 50 ohm resistor and replace it with a simple LM317 or other 'variable' regulator. Copy any typical schematic for the voltage regulator.....

The 5.6 ohm resistor I made with a pair of 10 ohm half watts. The total package needs to be able to pass 1/20th of the total AUDIO power fed to the final... In a typical CB, this means it needs to pass ABOUT a half watt, average. The diodes need to be able to stand 4 times the unmodulated input voltage (13.8 on a transistor rig).

IF YOU WANT TO DO THIS TO A DIFFERENT 'STYLE' CHASSIS, here is the math:

The diodes are 1/2 the final plate impedance.... IE, (Voltage x voltage) / (2 X Power out) and take that answer and divide it by two. For a 50 watt MOSFET, it would work like this:

13X13 = 169. 169 / 100 = 1.69 ohms. Half that. .8 ohms. This is the MATHEMATICAL way to find the resistor value. once it is found, you can place a RHEOSTAT or pot there to vary it to get the level you want. Be careful, in tube stuff, full B+ can be across this resistor, depending on mod index. You can use the same LM317 to feed the circuit..... Tubed or transistor, you only need a few volts to ensure the carrier is NEVER pinched off.

Want to put this in a series modulated radio? Like the 148 / etc? Do the standard NPC volt the final mod, and the wire that goes to the driver, you cut. Insert a "anti spike diode" in series, and otherwise wire up the circuit above. Is it easier than a cap? No. BUT, if you have a buttload of these made up, it makes it easier, and if you put a variable power supply on it, you have a LOT better method of keepalive than NPC...

Incidentally, the power supply that feeds the keepalive circuit MUST be low impedance. Therefore, even though most other schematics that are similiar to this show it with the cap being unnecessary, I keep mine in circuit ALL the time.... Low ESR cap means you get a low R power supply.

Enjoy. Any questions, please feel free to ask. Simplest circuit I've come up with , and this will do all that is needed.

Merry Christmas.

--Toll_Free

 

[Shockwave]

Just switched on 12 volts from the radio. Not keyed only on TX.