First-off., I know I'll be most likely be "stepping on some feet" here 
and to others, just "preaching to the choir".! But feel the need produce a somewhat mini-novel length of an "explanatory-rant" concerning a few "pet-peeve" subjects of mine, but will also WELCOME any additions, corrections, dis-likes, or likes.?
It's just that I've always enjoyed doing my best to take-the-time in explaining some basic (and some semi-complex) issues / principals surrounding just some of the MANY "urban-legends" floating-around for SO many years.!! Just ask ANY of my customers, of which I've been known (at times that is..) to "talk the ears off an elephant".!!
+++++++++++++++++
Pertaining to a "volted" final(s), disabled audio limiters (aka: "NPC" / negative-peak-compression / RF-envelope flat-lining, splatter mod) and Cobra-style "swing-kit" mods:
Well first-off., there's a GOOD reason as to why ALL manufactures (even since the early 1960s.!) of voltage-modulated, transistor-output, AM radios would apply the modulation voltage to BOTH the RF-driver AND the RF final., whether it be via the old-style dual-coil transformer (or autotransformer) architecture, or via the new-style, pass-transistor architecture.
Just check-out the oldest of the old, SAM's photofact books going back to the very first transistorized radios.!
Then, add to that., that EVERY volted-final radio that's ever come across my "surgery-table" also had its AMC circuitry (at best.!) trim-pot "cranked-up", or (at worst.!) had a part (or parts.!) "clipped" (cut) out of the AMC / audio-limiter circuitry. I once had an RCI-2950 that had been through 4 different shops before getting to me, and had found four (4) individual places where the AMC "chain" had a part removed or cut.!
I mean., it's like having a steel chain. If someone cuts-out just one individual "link" from it, it's now broken, and if someone (or in this case, each subsequent shop) cuts another "link" (in the chain) it's just as broken as before.! But yet, each "tech".?!, has their own favorite "monkey-see, monkey-do" component (from a CB-trick book or internet-page) that they remove or cut, in believing in (or "justifying" their service-fee for) their customer's "peak & tune" job
BTW, I just HATE that term., because it's like going to an auto-repair shop, and then having the auto-technician asking the customer: "Sir, would you like to have your vehicle's "front-end" to be "peaked & tuned".? -- Well, it's aptly called an "alignment" for a reason.!
And because of this volted-final modification (especially when combined with a typically "cranked" or "clipped" audio limiter) I've had a number of "blown-final" rigs come-across my bench. But luckily (in the case of the older BJT final rigs) I've still got quite a "stash" of genuine OEM (Mitsubishi) 2SC-1969s.!
BTW & FYI., just check the spec-sheets for the OEM Mitsubishi 1969 and the 2312, and although you'll find that both of them are rated at about the same ~18w output, the hFE forward-gain of the 1969 (depending on the suffix letter of X, A, B, C or D., but typically C) has at least a 30-40 point higher rating than a typical 2312, plus the OEM 1969 has that very nice internal reverse-protection diode, which the lower hFE gain 2312 doesn't.!
**Anyway., sorry ya'll for me continuing my mini-rant here., but after being the lead-tech for multiple Motorola company-owned, service-centers, which then lead to becoming a self-employed, sole-proprietorship business for 35+ years in servicing virtually only Motorola & GE "professional-grade" FM equipment & control systems, and then because of micro-sized, "surface-mount" components (and SUPER reliability.!) of such professional equipment, bench-servicing basically "dried-up", which left mostly just the "grunt-work" of doing "installs"
I've now (for 12+ years) become all TOO aware of the almost universal LACK of "electronics-101" knowledge AND the understanding to apply such concepts to actual, real-world "consumer-grade" electronics (radios, RF-amps, guitar / audio amps, inverters, etc) of the wanna-be / self-proclaimed, electronics-technicians "out-there"
But getting back to the 1st subject of this ranting-post (regarding "volting") and about the voltage-modulating of both the driver & final in a transistor transmitter, as opposed to modulating only the RF-final in a tube-type transmitter-chain.
A vacuum-tube is for the most part / cases, a voltage amplifier (measured as mutual-conductance / micomhos / Gm) versus a transistor, which is for the most part / cases is a current-amplifier (measured as beta / current-gain / hFE)
A tube-type, plate-voltage modulated, single RF final transmitter has (in most cases) a "clean-looking", symmetrical, AM-modulated, RF-envelope output. This is mostly because the "modulation-free", RF carrier that's "driving" the control-grid of that RF-final tube, contains virtually NO "power" (barely a few milliwatts.?) because of the extremely high input impedance / resistance of that vacuum "voltage-valve", and so as the plate B+ supply drops (during the negative-going halves, of the plate modulation voltage "swing") there will NOT be seen any "leak-though" (of power) from the non-modulated, RF-carrier / driver stage showing-up in the resulting RF-final output (at the negative-peaks / "mirror-image valleys") of the RF-envelope.
BUT., contrast that with the requirements to "drive" a typical BJT power-transistor, in which there NEEDS to be a relatively large amount of power (~0.5 - ~2.0w) to "drive" that RF-final transistor (depending on the make-up & power output rating of the RF-chain).
In such a typical transistor AM modulated transmitter-chain, if you would ONLY voltage-modulate the RF-final (as is done in tube-type, plate-modulation) meaning that the transistor driver stage would consist of a relatively large amount of un-modulated carrier power (of the driver) could then easily "pass-through" that final stage, even during the negative-going halves of the modulation voltage applied to the RF final's collector, which could / would result in a distortion of the negative-going peaks / valleys of the output RF envelope, by not allowing for the "mirror-image valleys" to drop to zero (which is what equals a 100% modulated "envelope")
Hence., virtually ALL manufacturer-designed voltage-modulated transistor, transmitter-chains have a 2-tiered chain of collector / drain modulation.
This is because, when the final AND driver stages are voltage-modulated together (in phase) the amount of un-modulated RF carrier power required (from the pre-driver) is SO small (~25 - ~100mw) that it should / will not affect the output RF-envelope wave-form "valleys" (near the 100% point) of the voltage modulated RF driver, much less the RF final.!
Sooo, if the RF-final is "volted" (tied to B+) and only the "driver" is modulated, whether it's just an AM-only "class-C" biased RF-chain, or especially in the case of a transmitter-chain with SSB capability ("class AB" biased) because in such RF-chains, there will be a pre-set bias-voltage applied to that RF-final and driver during transmit mode, which makes those transistors even more susceptible to "pass-through" of RF power from the preceding stage, and will result in some amount of that RF power from the modulated driver stage (~0.5 - ~2.0w) to leak-through (and be amplified by) the "volted" final.
Even when the modulated RF driver's collector-voltage is near the RF "cut-off" point (during the negative-going halves of the modulation voltage) some of that driver's RF-power will easily just "push-its-way-through" (and then be amplified by) the RF-final (which now has full B+ applied to it) and will show-up as some form of a distortion in the output of the resultant RF-envelope's, negative-peaks ("valleys")
Also, one must consider that in such a "volted-final" scenario, that each "power-transistor" (driver and final) are now "sitting" at a MUCH different part of their beta / hFE "output-curve", which is caused by the GREATLY differing collector voltages of driver & final during modulation.!
AND it will also do NOTHING to increase the transmitter's output (you're still limited to the spec-sheet, power-rating of that volted-final.!) except that now, it will display itself in the form of some (or a lot of.!) fake / splatter "air-watts" on your watt-meter (no matter HOW expensive someone's watt-meter may be.!)
***A good illustration / analogy that I believe explains this "fake-effect", would be like having a natural-gas meter outside of your house, that measures how many cubic-feet of natural-gas that has passed through it, to heat your house. It measures the total volume of gas (ANY gas or air.!) that passes through it, right.?! Well, what would happen if your gas company added other types of non-flammable (or less-flammable gases) to your supply line.? Simple.! Your gas-meter would now show you're using more cubic-feet (total volume) of gas, for the same amount of heat (or less) generated in your home.!
It's the same principal for RF harmonics & "splatter". They ARE in fact, actual watts, but are on other, much higher frequencies, which are then just simply added together in a / your wattmeter, but they're watts that are totally useless.! And given the fact that they are FAR away from 27Mhz (mostly 54Mhz & 81Mhz and even higher multiples.!) they also then just "add insult to injury", in the fact that all of those higher-frequency watts being produced, are just being "jammed right back down the throat" of your transmitter / amplifier as full-out reflections from your tuned-antenna giving you an SWR of 10+:1 on those other "out-of-the-ballpark" frequencies.!
Then another "add insult to injury" fact (that fools your watt-meter into reading those "air-watts") is that those higher frequency harmonics (2nd, 3rd, 4th, 5th, etc.) AND splatter, will "push" your watt-meter's needle (or digital read-out) MUCH further per-watt produced at those higher frequencies, on your meter, that's been designed / calibrated for ~30 Mhz.!
Meaning in general, that for every watt produced at some of those higher frequency harmonics & splatter can AND will "push" your 30 Mhz designed & calibrated meter (even a PEP modified Bird-43 meter with the proper "slug" inserted.!) will be displayed by as much as 3 or more watts, per 1 watt of that harmonic-splatter.!!
Due to "character-limit"., to be continued in just a moment...
Radioman56, your knowledge here impressed me and surpasses most that have tried volting a final. Believe it or not this entire thing about volting the final started as a 1986 high school electronics project when we were studying different modulators. Being a CBer, I was comparing this to what was being done inside the 148GTL and the 29LTD. I noticed the 148GTL had a larger final but was only applying modulated 6 volts VCC to the final collector and was essentially wasting the 5 watts of audio output from the TA7222AP, since it was only driving the darlinton pass transistor in AM.
The 29LDT put modulated 12 volts on a smaller final from that chip, through what first appears to be a modulation transformer. Turns out in TX, only one winding was being used as an AF choke to isolate it from the DC supply. All that was needed here was a Radio Shack 10 amp alternator whine choke. And, of course rewiring the mode switch so that the 8 volt bias line only turned on in SSB. Now the pass transistor ran cool only modulating the driver and the final 12 volt VCC was modulated through a coupling cap and choke, from the AF chip.
As I read your second part of the post, you mention very good phasing information when combining two modulated stages. The stock 148GTL takes audio from the TA7222AP and passes it through a coupling cap (90 degree phase shift), through two more transistors where the the phase remains the same as the input to them. The overall change is 90 degrees from the output of the AF chip to the finals. When modded, the final goes through one coupling cap and provides a 90 degree shift. To line the cap charge time up with the driver, the closest common value cap was a 470 uf. Messing this up still made the watt meter happy but produced noticeable distortion in the audio.
Eventually a local (T-Bird) purchased the modification, reverse engineered half of it and sent it into Secrete CB Magazine. He never did figure out you still needed to run the final in class C in order to properly modulate it at 12 volts on AM and every one I've seen done elsewhere left that part out too. Snake one wire into the stock harness and it gets easily missed. That mistake killed a lot of stock finals. If you took the time to bend the pins on an MRF477 to fit the pinout of the 2SC1969 and rematch the lower output impedance, this thing took off with the higher gain but killed internal, base power supplies. Today the mod has been reduced to nothing more than placing 12 volts on the final in most cases. That's just a bad idea all around.
