Driver Modulation

[K9ACT]

I have an on-air friend who is a crusader for modulating the driver along with the final and provides a long list of the virtues of doing this.

The one that got my attention was that it is a common cure for not quite 100% modulation peaks.

Apparently, it can be quite crude, the only requirement is that the drive goes up and down with the final modulation.

I am interested in other opinions on the subject along with suggestions on doing this to my 807 driver for my 811 and 8000 rigs.

Thanks,

js

[k4kyv]

Gates does it in the BC1-T and later tube-type transmitters, by using a tertiary winding on the modulation transformer.  It allows them to get away with skimpy rf drivers (a pair of 807's driving a pair of 833A's) and still maintain modulation linearity.

It has been argued that it is impossible to achieve perfect modulation linearity with triode tubes without simultaneously modulating the driver stage.

[Jim, W5JO]

Modulating the driver is done in some aircraft radios.  Bendix-King did it in the KX 173 series radios and I am sure other did as well.

[steve_qix]

I used a slightly modulated driver (about 40%) in my vacuum tube PWM rig I built back in '85.  This transmitter used a pair of 450tls modulated by a single 4PR1000 just for reference.

The 450TL requires a lot of drive, by the way, to handle modulation linearly up to 150% positive.  To avoid excessive grid heating, I modulated the driver (which was a 4-400 operating with no cooling, since it run at relatively low power).

I modulated the screen of the driver, in this case.

This was done in various broadcast transmitters, those that used triodes in the output.  I have not seen it done in any transmitter using a tetrode in the output.

Generally, the modulation is around 30%, from what I've seen.

Regards,

Steve

[ka3zlr]

The Schools of thought are Few OM, I would Start by stating First,That, "certain tubes make better transmitting Tubes by Convenience" and as well, "in keeping with ratings envelope for Class of service" a gud rule of thumb "is to underrate your tubes to 3/4 CCS RF Service", and they will live a happy life and your tube budget will thank you as well.

As to your circuitry Involved I quote the master:

More emphasis should be placed on the selection of proper rf circuitry to
handle the power you are trying to generate.  Good choice of the proper
neutralization, proper biasing, proper bypassing, proper plate/screen
voltages and currents are much more important than the particular type
of tube used.

A gud school of thought here OM.

Now as to your question on Audio, would you rather cheat yourself by a possibility of generating transients or build a better mouse trap..? it's that simple...

A properly designed, "Insert Here" (many circuits are accessible here and on Huzmans  Web Site for gud reference) Audio Stage(s) with Correct early stage Phasing and Feedback will go a much longer way in Feeding your Class of Service RF stage, Do Not Cheat Yourself here...is the proper answer...

Build Strong, Build Well, and Build it Right the first time, Time to do Some reading OM

73.

.

[K9ACT]

Build Strong, Build Well, and Build it Right the first time, Time to do Some reading OM

73.

.

I presume but it is not obvious, that you are addressing me, i.e. the original poster.

Now kindly point out what this has to do with the question I posed, viz., is driver modulation a good idea?

If you are addressing someone else, my humble apology.


js

[ka3zlr]

Hello,

There's no presumption, I just follow what the Masters state:

This is good Reading:

 k4kyv:
Q:  Can't the best quality audio of all be gotten from low level modulation and a linear amplifier?

A:    A linear amplifier has the same kind of distortion as a class-B modulator.

With tubes, the best quality audio can be had from low distortion plate modulators such as class-A series or Heising modulation, or pushpull plate modulators running class A or AB1.

Pulse-width series modulators produce perhaps the best audio.

I suspect the best quality of all comes from the new solid state class-E rigs.

According to the tube manuals, class-B audio service has inherent distortion levels on the order of 3-5%.  It can be reduced with negative feedback.  My Gates BC1-T manual claims less than 2% distortion at 100% modulation.

The signal driving a linear amplifier has its own distortion, since the original signal has to be produced somehow.  Control and screengrid modulation is inherently less linear than high level class-C plate modulation.  Pushpull transformer coupled class-A or class-A series modulators, with negative feedback, might be good candidates for the driver stage of a linear.  If the linear is run properly in class AB1, that would be near to the best possible audio out of a tube transmitter, even though the efficiency is not all that good.

Speaking of efficiency, an AM linear or grid modulated amplifier has close to the same overall efficiency as plate modulation, when calculated from the ratio of power drawn from the a.c. mains, to rf carrier output.  A linear amplifier running AM has exactly the same efficiency as when it runs SSB.  It's just that the duty cycle is different.


Actually, since with the human voice, the average power is 7-8 dB lower than peak power (equivalent to around 30% modulation), the average efficiency of a SSB linear is similar to that of an AM linear because the efficiency of a linear is a function of the amplitude of the signal (0% at idling current, and a practical real-world maximum of about 67% at maximum peak output just below the point of saturation or flat-topping).  AM linears got their reputation as "low efficiency" on AM because the 100% duty cycle carrier runs about 30% efficiency to allow enough headroom for the positive peaks.  With an AM linear, you can see the glow on the plates DECREASE when you whistle into the mic to produce 100% tone modulation.  The DC input is the same regardless of modulation, but the rf output is higher with modulation, since sideband energy is now included.  That power has to come from somewhere, so the efficiency of the amplifier goes up to generate the sidebands.

A major advantage of plate modulation with amateur AM is the ease of tuning up and QSY'ing.  You simply dip the final and load to the desired carrier output, while maintaining enough grid drive to assure class-C service.  With low-level modulation (linear or grid modulated), the rf drive level and degree of antenna coupling are critical to the modulation linearity of the final.

K1JJ:
Good article for the archives, Don!


Don is a Master, anything that I've done I've followed his advice as Gospel and Never went wrong.

You need to decide what it is "you" want to do...there's no Holy Grail, it's always best to read, Help is in the Archives and decide for yourself.

[Steve - WB3HUZ]

The Masters state things on all sorts of subject. Quoting them makes no sense unless you quote them on Jack's question.

I am interested in other opinions on the subject along with suggestions on doing this to my 807 driver for my 811 and 8000 rigs.

.... as in modulating the Rf driver for a plate modulated Class C final. None of what you quoted answered Jack's question.

I think Don and Steve covered it pretty well.

[ka3zlr]

Ok,well He does that, then his final stage tuning becomes even more critical, to gain what..some theoretical percentage over a weak assembly to begin with, it's extra werk over the air is going to make little difference...if there's a stability problem or a weak link in the assembly fix the assembly.

Everything is right there in Don's article...I'm Not the Crusader here, I just believe in Building heavy and strong, Big iron Big Toobies Big Signal...Grrrrrrrrrrr..LOL...

[Steve - WB3HUZ]

Your assertion that modulating the RF driver stage is somehow poor or weak design is incorrect. In general, for triodes, it is the preferred method.

[k4kyv]

Your assertion that modulating the RF driver stage is somehow poor or weak design is incorrect. In general, for triodes, it is the preferred method.

It should improve modulation linearity of any triode class-C final.  Where the weak design comes in, is allowing good modulation linearity with marginal grid drive.  The pair of 807's is just barely capable of driving the 833A's, but the extra kick on positive modulation peaks gives the final the extra grid drive it needs under those conditions.

The weak design is not audio driver stage modulation, but the skimpy rf driver stage that is being modulated.

[WU2D]

The British designed SCR-522 aircraft transmitter (about 1939) which used an 832 tripler driver and an 832 Final modulated both stages so the idea is old and may come from light weight gear. I converted one of these and it had plenty of modulation. I think is used a pair of 6V6's and a tiny transformer for modulation.

Mike WU2D 

[ka3zlr]

On the receiving end of things a Good Clean Solid Signal does more for anyone than this Grail search, and that i will assert in this.

On the tall ships side of this on the air today what do i see on my Scopes Good Clean strong Signals and Attention to detail, and that Is something to Reach for.

and a cold 807 after I rototilled the garden...

[WBear2GCR]

Jack,

Try it, see if you like it or not.
See if you can make it sound worth a damn or not.
See what you make the feedback fix or not.

Go for it.

    to everyone else...   

                _-_-bear

PS. I've heard you on 75AM a few times here on the eastcoast... Q5 but not enough snot on this end to make the return trip!

[K9ACT]

Jack,

Try it, see if you like it or not.
See if you can make it sound worth a damn or not.
See what you make the feedback fix or not.

Go for it.

   

My first attempt was a miserable failure and I need some ideas.

I put the primary of an output transformer in series with the screen dropping resistor and connected the low end
across the output of the speech amp that drives the modulators.

I doubled the size of the screen resistor which cut the power about in half but still only got about 10% modulation of a sine wave on the 807. At this level, it does not seem to have any effect on the audio of the final.

My speech amp is a solid state module putting out about 20 watts.  This feeds a transformer through a large cap that drives the modulator grids which are 811's or 813's depending on the rig.

I put the new tranny across the same cap and assume I can rob enough audio to modulate the screen.

I need some ideas on a simple way to modulate the 807 using this setup.  My understanding is that is does not have to be hi fi.

js

[Ralph W3GL]

No, it does not need to be hifi BUT it does need to be in phase...

[Opcom]

a resistive divider could be placed across the modulated plate supply and this tapped for the driver stage's plate supply. Alternately, the tap point could be capacitively coupled to a transformer winding, with another winding being in series with the plate supply to the driver stage. The capacitor would be large enough to minimize phase shifts. Just some ideas. It would not be necessary to 100% modulate the driver as others have already indicated.

[k4kyv]

The way RCA does it with one of their 5 kw transmitters is to use a tapped modulation transformer secondary, rather than a tertiary winding.  The transformer is used with a reactor in a conventional "modified Heising" arrangement, with the modulation reactor feeding DC to the RF final, and the audio coupling cap in series with the top (plate) end of the modulation transformer secondary.  The bottom end of the mod xfmr secondary goes to the positive lead of the rf driver supply instead of ground, with a large enough output capacitor on the driver power supply to thoroughly by-pass the audio to ground.  The tap on the secondary feeds plate voltage to the rf driver stage.  The driver DC passes through the modulation xfmr secondary, but the effect on performance of the transformer is inconsequential, since the driver modulation utilises a small fraction of the power output of the modulator.

This arrangement could be used with a multi-match modulation transformer, such as the VM-5 or CVM-5, by selecting one of the unused taps on the secondary side to feed the rf driver. It would be easy to calculate the percentage of modulation available with each tap, or this could be determined experimentally.

[K9ACT]

Some interesting results to report.

With one 811, the trap test shows perfect linearity and nothing improves with the driver modulation.  One can't make straight lines straighter, I guess.

With two 811's, there is a slight curvature to the lines and I can not get more than 85 ma with the 600 ohm grid leak.  Hard to believe that a 6% shortfall would matter much but that could be about how much the lines deviate from straighthood.

When I apply the driver modulation to the 807, there is a small but obvious straightening of the lines.

So, it looks like more screen modulation would do the trick.

I appreciate the insight on how this has been done but in my case, it has to work with two different rigs so I can't conveniently get into the high level area of the PA's.

I need to do it at the driver level so the switching can be done at the speech amp.

Again, I have watts to spare at the speech amp and just need a simple way to modulate the 807 grid with this.

js

js

[Bacon, WA3WDR]

The grid characteristics of a triode are strongly affected by its plate voltage.  A triode's grid operating characteristics change considerably over the range of plate voltages that appear during modulation, and this tends to cause some flattening of positive modulation peaks when plate modulation is applied.

At low plate voltages, the triode's grid tends to pull higher current at a given grid voltage.  At high plate voltages, not only does the grid draw less current at a given grid voltage, but the grid voltage needed for plate cutoff also becomes more negative.  Basically you need some grid drive at low plate voltages, and more drive and more bias on positive modulation.  Grid leak bias and a partially modulated drive signal are the answer.

A tetrode or pentode is a good choice for a driver, because they act like current sources because they have high plate resistance.  When the triode plate voltage is high (during positive modulation), its grid draws less current (it presents a higher impedance to the driver), and the tetrode or pentode driver provides more drive voltage because the triode's grid impedance increased.  This self-modulation of the driver stage tends to maintain the triode's grid current on positive peaks, which is a big step in the right direction - but more is usually needed, so some in-phase modulation should be applied to the driver stage as well.

Whether fixed bias is used or not, grid-leak grid bias is desirable in the triode stage, so that higher drive levels produce more negative grid bias on the triode stage.  Low values of grid leak bypass capacitance are desirable as well, so that the grid bias can closely track the modulating audio at higher audio frequencies.  It is possible to apply reverse-polarity modulating audio here directly, but I don't think anybody does that.

If modulation is not applied to the driver stage, then its screen voltage should be fixed or regulated, and not supplied through a dropping resistor.  A stiff screen supply will maximize the plate resistance of the tetrode or pentode.

[ka3zlr]

Hi Bacon,

 I have read those articles down through the years on the AM page as well, and agree with the practices, But, he's wanting to feed his driver from early stage, and have the capability of splitting between two transmitters...as I'm reading it or maybe i misunderstand, and my take is it will have to be series fed...or I'm reading something wrong here...I feel it would be better served taken care at the transmitter level, But he's wanting early stage...this outta be neat to watch what transpires.


Phase will be a big factor and feed control as well.....Controlled Carrier anyone  ..?

meaning No Disrespect here...Good Morning All, I haven't had my Coffee yet..