Question abt Push-pull RF Amplifier plate tank circuits
Hipiket
I'm planning to build an RF amplifier using two tubes in push-pull and I
have a question about the tank circuit.
I've noticed that every design I have seen from the 1930's to the 60's
ARRL handbook has an RF choke at the center-tap of the tank coil, where
the B+ is applied to the circuit.
Why is this? When one tube is in conduction and that end of the plate
coil is approaching zero, the other end will be near two times B+, but the
center tap should always be at B+ but have no RF at that point.
Why can't I just use a feedthrough capacitor and a short lead to the coil
center-tap and dispense with the RFC?
Thanks for any ideas...de AC4AQ (Jeff).
W8JI Tom
In article <4v2o53$n...@newsbf02.news.aol.com>, hip...@aol.com (Hipiket)
writes:
>Why is this? When one tube is in conduction and that end of the plate
>coil is approaching zero, the other end will be near two times B+, but
the
>center tap should always be at B+ but have no RF at that point.
>Why can't I just use a feedthrough capacitor and a short lead to the coil
>center-tap and dispense with the RFC?
>Thanks for any ideas...de AC4AQ (Jeff).
Hi Jeff,
You can use a coil that is "RF grounded" at the center, but then you
should float the rotor of the tuning capacitor for RF.
You don't want two conflicting tuned circuits to be created, with each
half of the coil and each half of the air variable potentially tuned to
different frequencies by stray effects and component tolerances.
It is generally considered safer to float the coil center and RF ground
the rotor of the tuning cap, because the though is that gives better high
order harmonic suppression.
There is no reason in the world to use a 2.5 mH choke though, that value
is much too high and can lead to VLF parasitics. A few hundred
micro-henrys with a high value resistor across it would be a better choice
for the center of the coil at HF.
73 Tom
Clifford Buttschardt
Jeff and the group. This is a very old question and I guess only an oldie
can answer it. In any push-pull tank coil and split stator arrangement
there is two ways in which the ground can be accomplished. Either ground
the split stator rotor or RF ground the center of the coil BUT NOT BOTH!
One way to visualize this is as two separate single tuned circuits
magnetically coupled. Now you can see why both capacitor and coil should
not be grounded as they will be fighting each other for proper balance.
Therefore, if the center of the capacitor is grounded, an RF choke in
the center of the coil should not have a bypass. If the center of the
coil is bypassed to ground, then the capacitor need not be split stator at
all since there is no need for a capacitor ground! Of course, the tuning
capacitor often is split stator. Be careful here as one can get into
difficulty with high voltage on the tuning knob either permanently or when
the capacitor flashes over! Cliff Buttschardt K7RR ex W6HDO
On 16 Aug 1996, Hipiket wrote:
> I'm planning to build an RF amplifier using two tubes in push-pull and I
> have a question about the tank circuit.
> I've noticed that every design I have seen from the 1930's to the 60's
> ARRL handbook has an RF choke at the center-tap of the tank coil, where
> the B+ is applied to the circuit.
William E. Sabin
William E. Sabin wrote:
>
> Mark Mandelkern wrote:
> >
> > w8j...@aol.com (W8JI Tom) wrote:
> >
> > >You can use a coil that is "RF grounded" at the center, but then you
> > >should float the rotor of the tuning capacitor for RF.
> >
> > >the rotor of the tuning cap,
> >
> > my homebrew push-pull 4-400A, 50 mc. amp, neither is at r.f. ground.
> > No connection at the rotor. RFC, 6 uH, at the coil. Working fine for
> > 35 years! I don't know any theory, so except for the harmonic issue
> > Tom mentioned (no problem here), I don't understand why one point or
> > the other must be at r.f. ground. I hope Tom will explain it for us.
> >
>
> Let's see if I can guess. The RF choke has enough energy stored in its
> inductance that it acts as a constant-current source for the RF plate
> current pulses. The two tubes switch this current pulse back and forth,
> in the process energizing the flywheel tank. The DC supply re-supplies
> this energy to the choke. This works only for a balanced push-pull
> circuit. A single-ended circuit needs a ground reference point (a bypass
> capacitor).
>
> Bill W0IYH
I ain't at all sure this is right!!!!
Bill
WB3U
"William E. Sabin" <sab...@crpl.cedar-rapids.lib.ia.us> wrote:
>I ain't at all sure this is right!!!!
I think Tom was referring to the fact that the coil will not
be exactly symmetrical around its center. If both the center
of the coil and the rotor of the tuning cap are grounded, the
actual inductance seen in the tank will change as the tubes
switch back and forth. This will cause the resonant frequency
of the tank to be slightly different, depending on which tube
is conducting.
73,
Jack WB3U
Mark Mandelkern
w8j...@aol.com (W8JI Tom) wrote:
>You can use a coil that is "RF grounded" at the center, but then you
>should float the rotor of the tuning capacitor for RF.
>It is generally considered safer to float the coil center and RF ground
>the rotor of the tuning cap,
Tom seems to imply that one or the other should be at r.f. ground. In
my homebrew push-pull 4-400A, 50 mc. amp, neither is at r.f. ground.
No connection at the rotor. RFC, 6 uH, at the coil. Working fine for
35 years! I don't know any theory, so except for the harmonic issue
Tom mentioned (no problem here), I don't understand why one point or
the other must be at r.f. ground. I hope Tom will explain it for us.
73, Mark K5AM DM62
Mark Mandelkern
Las Cruces, New Mexico, USA
William E. Sabin
Mark Mandelkern wrote:
>
> w8j...@aol.com (W8JI Tom) wrote:
>
> >You can use a coil that is "RF grounded" at the center, but then you
> >should float the rotor of the tuning capacitor for RF.
>
> >It is generally considered safer to float the coil center and RF ground
> >the rotor of the tuning cap,
>
> Tom seems to imply that one or the other should be at r.f. ground. In
> my homebrew push-pull 4-400A, 50 mc. amp, neither is at r.f. ground.
> No connection at the rotor. RFC, 6 uH, at the coil. Working fine for
> 35 years! I don't know any theory, so except for the harmonic issue
> Tom mentioned (no problem here), I don't understand why one point or
> the other must be at r.f. ground. I hope Tom will explain it for us.
>
Ian White, G3SEK
Isn't it simpler than that? There are actually three ground points, not
two. The third one is the center-tap between the output capacitances of
the two tubes, and that one is always grounded.
In Mark's situation the tube capacitances are providing the one-and-only
center tap for the output circuit. That's just fine, so long as the
output capacitances are large enough to swamp the out-of-balance strays.
This is usually the case at VHF: the "Plumber's Special" and W1SL push-
pull amplifiers work on the same principle.
In some situations you can get away with grounding either the tuning
capacitor center-tap or the coil center-tap as well as the center-tap
built into the tubes, without producing large circulating chassis
currents between the grounding points; but the circuit is usually not
symmetrical enough to allow you to ground all three.
73 from Ian G3SEK Editor, 'The VHF/UHF DX Book'
'In Practice' columnist for RadCom (RSGB)
Professionally:
IFW Technical Services Clear technical English - world-wide.
William E. Sabin
Ian White, G3SEK wrote:
>
> Isn't it simpler than that? There are actually three ground points, not
> two. The third one is the center-tap between the output capacitances of
> the two tubes, and that one is always grounded.
>
> In Mark's situation the tube capacitances are providing the one-and-only
> center tap for the output circuit. That's just fine, so long as the
> output capacitances are large enough to swamp the out-of-balance strays.
> This is usually the case at VHF: the "Plumber's Special" and W1SL push-
> pull amplifiers work on the same principle.
>
That idea did occur to me also, but I am not sure that the tube plate to
ground C is large enough, compared to the other circuit capacitances, to
establish a credible ground return. The "cold" end of each tube's plate
load impedance should return to its cathode with a low impedance path.
I get the feeling that without a well established ground return, the
situation is rather messy and although it "works", it would not pass
muster in any engineering organization.
My idea of the constant current RF choke doesn't sound very good, on
second thought, and I withdraw it.
Bill W0IYH
W8JI Tom
In article <4v7m2v$g...@news.lascruces.com>, k5...@lascruces.com (Mark
Mandelkern) writes:
Hi Mark,
>Tom seems to imply that one or the other should be at r.f. ground. In
>my homebrew push-pull 4-400A, 50 mc. amp, neither is at r.f. ground.
>No connection at the rotor. RFC, 6 uH, at the coil. Working fine for
>35 years! I don't know any theory, so except for the harmonic issue
>Tom mentioned (no problem here), I don't understand why one point or
>the other must be at r.f. ground. I hope Tom will explain it for us.
It needs to be at RF ground because one half of push-pull resistances
pulling on the tank are off during part of the cycle.
Look at a perfect case, drawn it out on paper.
If the tank truly floated at the center, and the center tap has infinite
RF impedance through an RF choke, the tank would have no reference for the
tube that was "on" fully since the entire rest of the tank would be
"floating". Even the other tube would be "off".
The amplifier would not function at all if the center tap truly floated
and the conduction angle of each tube was less than 180 degrees! We could
not make the circuit work at all on any frequency. The low frequency
analogy would be a push pull audio output output stage with the center tap
fed through a very high reactance audio choke.
The circuit must provide an "RF reference to ground" in the tank for the
circuit to work at all. For proper performance it should be the center.
The reference can be achieved "accidently" via always present circuit
strays (capacitance from the plate circuit of the opposing tube---or
tuning capacitor stray frame and rotor to chassis capacitance) will work,
but doing it purposefully is almost always better.
73 Tom
Ian White, G3SEK
William E. Sabin wrote:
>Ian White, G3SEK wrote:
>>
>
>> Isn't it simpler than that? There are actually three ground points, not
>> two. The third one is the center-tap between the output capacitances of
>> the two tubes, and that one is always grounded.
>>
>> In Mark's situation the tube capacitances are providing the one-and-only
>> center tap for the output circuit. That's just fine, so long as the
>> output capacitances are large enough to swamp the out-of-balance strays.
>> This is usually the case at VHF: the "Plumber's Special" and W1SL push-
>> pull amplifiers work on the same principle.
>>
>
>
>That idea did occur to me also, but I am not sure that the tube plate to
>ground C is large enough, compared to the other circuit capacitances, to
>establish a credible ground return.
The output capacitance is between the plate and elements in the tube
(screen grid or control grid) which need to be well-grounded, plus
direct plate-to-chassis capacitance, so there's no problem in the
'quality' of the ground.
Whether this center-tap is enough to balance the whole circuit will
depend on the degree of inherent (physical) balance, and on the size of
the stray external capacitances to ground compared with the tube's
output capacitance.
>
>I get the feeling that without a well established ground return, the
>situation is rather messy and although it "works", it would not pass
>muster in any engineering organization.
>
On the contrary, at VHF it *works* for push-pull quarter-wave lines -
and it usually stops working if you add extra grounding!
The problem is that an extra ground converts one parallel-tuned circuit
into two separate but closely coupled single-ended circuits. Unless
these two circuits are identical (which is the same as saying, unless
the whole thing is balanced) they will have separate resonances and
there will be significant common-mode currents flowing through the
chassis between the two different grounding points. The result is that
the circuit doesn't resonate correctly, and there is a large risk of
instability caused by coupling via the chassis currents.
It is very difficult to establish an exact electrical center-tap on a
coil, because the coil is inherently unsymmetrical, so the best advice
seems to be "let it float". If you want to force a center-tap in
addition to the one provided by the tubes, do it on the tuning capacitor
- but make sure that the capacitor itself is well-balanced. A butterfly
capacitor works best. Experience at VHF is that a more conventional
twin-gang capacitor (two sections in line) needs to be mounted on
insulating pillars, with a wide ground strap connected from the exact
center of the metal frame. The grounding point must be close to the two
tubes and on the centerline between them. The variable capacitor must
obviously have a low-capacitance insulated shaft coupler and preferably
a non-conducting shaft as well.
You may be able to ground the whole capacitor at lower frequencies.
But if you're wrong it could mean a major mechanical rebuild, so...
Also note that everything that's been said about the plate circuit
applies even more to the grid. In addition to all the other problems,
unbalance in the grid tuned circuit will deliver unequal drive to the
two tubes. In push-pull grid circuits there are advantages in floating
both the variable capacitor and the inductor, and grounding the circuit
through a preset differential capacitor to balance out all the strays
(two separate preset trimmers can be used instead).
William E. Sabin
Ian White, G3SEK wrote:
>
>
> The output capacitance is between the plate and elements in the tube
> (screen grid or control grid) which need to be well-grounded, plus
> direct plate-to-chassis capacitance, so there's no problem in the
> 'quality' of the ground.
>
> Whether this center-tap is enough to balance the whole circuit will
> depend on the degree of inherent (physical) balance, and on the size of
> the stray external capacitances to ground compared with the tube's
> output capacitance.
>
If we have two tubes A and B in push-pull, then when tube A is
conducting, the plate to ground capacitance of tube B plus any ground
path inductance (assumed to be > 0) back to the cathode of tube A,
presents some *possible* problems with spurious resonances and stray
currents going where they are not wanted. I like your idea of the
butterfly cap with single-point ground. As an alternate, I would like to
see a fixed cap from each plate to a common ground point between the
tubes to augment the tube output capacitance. Also the RF choke in the
coil or quarter wave line B+ feed is good.
Bill W0IYH
W8JI Tom
In article <3217A8...@crpl.cedar-rapids.lib.ia.us>, "William E. Sabin"
<sab...@crpl.cedar-rapids.lib.ia.us> writes:
>
>Let's see if I can guess. The RF choke has enough energy stored in its
>inductance that it acts as a constant-current source for the RF plate
>current pulses. The two tubes switch this current pulse back and forth,
>in the process energizing the flywheel tank. The DC supply re-supplies
>this energy to the choke. This works only for a balanced push-pull
>circuit. A single-ended circuit needs a ground reference point (a bypass
>capacitor).
>
>Bill W0IYH
That could be true Bill.
If the choke were the sole component involved, the choke would store and
release a lot of energy to the tank. It would have to be a very good choke
or it quickly would become a very hot or perhaps an arcing choke.
It's hard to guess why a circuit that should not work very well actually
does. I can think of at least three things that help.... stray
capacitances from the center of the tank to ground, the choke being less
than infinite impedance, the tubes stray capacitance to ground (that's a
biggie).
The output capacitance of the tubes forms a center tapped capacitance in
parallel with the tank that is grounded at the center! If the tubes have
fairly low capacitive reactance compared to the tank capacitor values,
that's enough right there.
By the way, this is a second valid reason to float the coil and ground the
variable capacitor junction...but one that mainly applies at higher
frequencies. The original question refered to HF tanks.
At VHF, tube output capacitance forms a center tapped capacitance across
the tank, grounded at the center. If the builder uses a tank coil RF
grounded at the center at VHF, he is really asking for multiple resonance
problems.
73 Tom
Tom Bruhns
William E. Sabin (sab...@crpl.cedar-rapids.lib.ia.us) wrote:
..
: I get the feeling that without a well established ground return, the
: situation is rather messy and although it "works", it would not pass
: muster in any engineering organization.
: My idea of the constant current RF choke doesn't sound very good, on
: second thought, and I withdraw it.
For a practical use of a very similar circuit, have a look at what's used
these days to drive cold-cathode fluorescent lamps, such as back lighting
for LCD displays. Typically these are push-pull transformer coupled
arrangements with a choke in the Vcc lead to the push-pull stage. The
choke impedance is simply (more or less...) transformed by the transformer
to effectively limit the current in the lamp while allowing large voltage
to be developed to ignite the plasma in the lamp. A significant difference
is that the lamp drivers generally (from the circuits I've seen) do not
resonate the transformer as a tank. I believe the RF choke, as Bill
seems to suggest above, will limit the available power and significantly
raise the output impedance of the amplifier. Consider that if the tank is
shorted, if only one tube is "on" at a time and there is some time when
neither tube is on, you simply excite the choke at two times the amplifier
input frequency (plus harmonics).
--
Cheers,
Tom
to...@lsid.hp.com
W8JI Tom
In article <KF+qiDAd...@ifwtech.demon.co.uk>, "Ian White, G3SEK"
<G3...@ifwtech.demon.co.uk> writes:
>Also note that everything that's been said about the plate circuit
>applies even more to the grid. In addition to all the other problems,
>unbalance in the grid tuned circuit will deliver unequal drive to the
>two tubes.
Unequal load at the anodes causes a similar problem. One valve (I mean
tube) is undercoupled and one is overcoupled by tank unbalance.
>> In push-pull grid circuits there are advantages in floating
>both the variable capacitor and the inductor, and grounding the circuit
>through a preset differential capacitor to balance out all the strays
>(two separate preset trimmers can be used instead).
>
Not always, and certainly not unless the circuit's shunt C is a very low
reactance compared to the grid impedance. :-) It's a problem of
impedances, and general "do it like this's" can get us in deep trouble.
For an example, look at the grid circuit of a push pull class amplifier
that draws grid current. It has widely varying impedances at the grids
that gets pretty low, and very high capacitive reactance at each outer end
to keep things equal (unless the tube is *very* large). In class C (or
AB2) grid circuit, we want the coil center to be used as the drive
reference. Mutual coupling helps solve the problem, giving each grid a
stiff source.
The designer had best look at the actual application closely to decide
what method is best...especially at VHF. At a low enough frequency, we can
throw out some general rules and the system will work.
The original question was HF. In that general case with small tubes (like
20 kW on down), a center tap reference at the grid inductor and a center
tap reference at the grounded rotor on the output tank variable would
almost always be best.
73 Tom
Mark Mandelkern
I think it's clear that one should not r.f ground BOTH the
coil and the tuning condenser. But the question still is,
Why r.f. ground either?
My guess is that for older, smaller, lower-power,
lower-voltage amplifiers, it was very convenient to ground
the rotor of the condenser. So the RFC on the coil was
needed. But at the KW level it not so much of an extra chore
to mount the condenser on ceramic insulators and to use a
ceramic shaft insulator (and a grounding panel bearing for
safety).
One advantage of not r.f grounding the condenser is that
only the r.f. voltage, not d.c. plus r.f., will appear
across the plates. Also, there is less chance of a VHF
parasitic circuit set up via the condenser through ground.
The advantage of the RFC on the coil is that there is less
chance of any significant r.f. currents through the bypass
condenser. The simple feed-thru capacitor suggested earlier
would be questionable in regard to voltage rating (incl.
r.f. voltage) and r.f. current.
So it seems neither need be grounded, and all the mentioned
advantages obtained.
Repeating my earlier comment,
>In my homebrew push-pull 4-400A, 50 mc. amp, neither is at r.f. ground.
>No connection at the rotor. RFC, 6 uH, at the coil. Working fine for
>35 years!
73,
Mark, K5AM DM62
W8JI Tom
In article <4ve3cq$9...@news.lascruces.com>, k5...@lascruces.com (Mark
Mandelkern) writes:
>
>Also, there is less chance of a VHF
>parasitic circuit set up via the condenser through ground.
The VHF parasitic circuit is usually mostly internal to the tube in a well
designed and laid out PA. The tube's grid structure is the major player.
Tubes with long thin grid leads and large open grids (811A, 572B, 3-1000,
3CX1200A and D 7) all have low self-neutralizing frequencies and can be
prone to parasitics.
Tubes with small compact grids and short coaxial grid connections (8877,
3CX3000, 4CX1000, 4CX250) are relatively immune to parasitics when grids
are properly grounded.
So far as the anode circuit, what happens beyond the tank capacitor has
little effect on VHF stability. What we absolutely DO want is the
capacitor to be well grounded for VHF, and VERY close to the tube. If that
is done, only the connections between the capacitor and the tube remains
the major player in anode VHF resonance. Inserting a suppressor at that
point will allow us to control VHF path losses.
The resistor used in the suppressor MUST be large compared to the
impedance of the VHF path, and small compared to the reactance of the
suppression coil. Making the VHF path longer and less predictable by not
grounding the capacitor moves things in the wrong direction.
73 Tom
Hipiket
Wow! What a teriffic debate I sent careening down the mountain....
The plan I had was to use two RCA 833A's in a push-pull grounded grid
arrangement, with the objective of minimum distortion, rather than maximum
output. I intend to use a plug-in coil scheme, with doorknob "helper"
capacitors on wavelengths between 160 and 40 meters. Thus the split-stator
capacitor will not have to be of great size while providing a generous
breakover rating. Loading adjustment will be by means of a variable link.
160-15 meter coverage is planned (10 meters...if the band's open, you
don't need an amplifier and when it isn't an amp won't do much good
anyway. Really the same is true on 15 unless you're trying to out-holler
the other lids working a rare DX station.)
The real difference is that the power supply for this amp will provide
3250 VDC @ 850 MADC and only weighs 12 lbs (450 g)!!! It is silent,
isolated, regulated to 0.5%, ripple-free as can be. I conceived the idea
as a hybrid between a modern VCR power supply and that of the amp in the
so-called, "boom-car"...
Thanks for all of the valuable suggestions I've seen go by my CRT. Too bad
the Internet will make working DX stations seem quaint...
Hipiket
It was suggested to me that the RCA 833A was a poor tube choice for
operation above 10 MHz, due to internal lead length and "tube structure".
I guess that latter refers to interelectrode capacitance or possibly even
transit time. Since the length of the lead from K to the input transformer
will be on the order of 0.2% wave on 15 meters, I don't think that will be
much of a problem. The plate lead will be shorter still and the grid lead
shortest of all. Interelectrode capacitance is 12.3 pF for Cin and 8.5
pF Cout. These values are especially reasonable when you consider that the
push-pull input circuit will place the capacitance of the two tubes in
series while the parallel arrangement gives the sum of the tube Cinput. I
don't think I'm gonna worry about the transit time.
True, while I write this in hundreds of small towns high school football
games are being modulated and transmitted by 833A's in little Standard
Broadcast stations. But for operation at 21 MHz they should compare
favorably with the 4-400A or 3-500Z.
I thought distortion bore an inverse relationship to conduction angle and
this is effectively 360 degrees in an class AB push-pull amp whereas the
plate current is cut off during the negative peaks with a single-ended
amplifier. So I think that I will have a reduction of of odd harmonics,
but not as much as even order products.
Current sharing should be less of a problem...
But getting away from these Mason jars I intend to use and to the original
question. Does anyone think that there is an increase in efficiency in the
plate coil above 14 MHz with a plug-in coil design? I sense that there is
an advantage in not having the turns that are used for 160 to 40 doing who
knows what on the upper bands. Anyone have any thoughts on this?
I thank all for the knowledge shared here and urge you get in your shack
and build something. You can build better and cheaper than you can buy if
you just do a little research and a lot of scrounging and planning.
73 de AC4AQ
W8JI Tom
Hi Hipiket,
In article <508a7g$i...@newsbf02.news.aol.com>, hip...@aol.com (Hipiket)
writes:
>It was suggested to me that the RCA 833A was a poor tube choice for
>operation above 10 MHz, due to internal lead length and "tube structure".
>I guess that latter refers to interelectrode capacitance or possibly even
>transit time.
It's a problem of self-resonance in the internal structure. The grid
capacitance, along with inductance from the long grid structure and single
grid lead, lowers the self neutralizing frquency of the tube. This makes
the tube unstable at lower frequencies.
Examples of this can be found by looking at the parasitic suppression
methods required to stabilize various tubes. Tubes with long single grid
leads and large grids like the 811 or 572 require large suppressor
inductance, while tubes with compact grids (ie 8877) and coaxial grid
connections often don't require any suppression at all.
The 833 is the granddaddy of low self-neutralizing frequencies, so
stabilizing the tube without overdissipating the suppressor is a chore.
That isn't to say it can't be done (I've made sweep tubes work on 6 meters
and higher), just that you should be prepared for difficulties above 14
MHz.
>True, while I write this in hundreds of small towns high school football
>games are being modulated and transmitted by 833A's in little Standard
>Broadcast stations. But for operation at 21 MHz they should compare
>favorably with the 4-400A or 3-500Z.
The self neutralizing frequency of a 3-500Z is just under 100 MHz, with a
833A it's around 20 MHz. Imagine using a 3-500Z at 144 MHz. That's what
you're up against with a 833A on 28 MHz. You'll find all sorts of
regeneration because the anode to cathode feedthrough capacitance will
become very high above 7 MHz.
>I thought distortion bore an inverse relationship to conduction angle and
>this is effectively 360 degrees in an class AB push-pull amp whereas the
>plate current is cut off during the negative peaks with a single-ended
>amplifier.
Don't confuse distortion of the RF sine wave with IM distortion of a two
tone signal. Distortion of the RF sine wave is indeed less in a push pull
RF amplifier, and directly related to conduction angle. Reduced even order
harmonics of the driving frequency are the result of reduced RF sine wave
distortion.
On the other hand, IM products are directly related to gain linearity as
drive level is varied. Once a power grid tube is biased in conduction at
zero drive, very little change in IMD performance occurs with larger
conduction angles. Operating an RF amplifier in push-pull does not, as a
general rule, change IMD performance a great amount unless regeneration is
reduced.
>Current sharing should be less of a problem...
It will be just as much a problem. If one tube is operated in a non-linear
portion of its gain curve, it will contribute exactly the same distortion
in push pull as it would in parallel. Tube to tube balance will greatly
affect harmonic suppression. Remember IMD is caused by non-linearity as
the driving signal is varied in amplitude, NOT by non-symetrical
reproduction of the RF sine wave. It is possible to have VERY good IMD
performance in a single ended PA, in excess of -40dB, even though the
conduction angle is just barely over 180 degrees.
If conducton angle was reduced to 120 degrees *per tube* at full drive,
and two tubes were used in push pull, IMD would be horrible. It would be
little or no better than a single ended amp operated at 120 degree
conduction angle. That's because at a certain level of drive, the tube
simply quits amplifying. It's not only flat-topping that causes IMD, it's
low level clipping. Solid state exciters are good at this, they often have
POORER IMD performance at low power levels than at high power levels. If
low level non-linearity is a problem, reducing drive levels won't reduce
splatter...it will actually make it WORSE as a percentage of peak power.
That's one of the reasons asking someone with a NASTY amplifier to "turn
it down" sometimes does no good at all.
This is one of the reasons authors should be called to task for suggesting
people built tetrode amplifiers that can be driven with ten or twenty
watts. The idea of a ten watt drive PA being used with a 100 watt rig is
***stupid*** (in the most polite term).
>But getting away from these Mason jars I intend to use and to the
original
>question. Does anyone think that there is an increase in efficiency in
the
>plate coil above 14 MHz with a plug-in coil design? I sense that there is
>an advantage in not having the turns that are used for 160 to 40 doing
who
>knows what on the upper bands. Anyone have any thoughts on this?
A typical amateur grade multi-band tank circuit only consumes a few
percent of the total RF power. By far the largest concentration of loss is
in the PA tube. If all tank loss was eliminated, there would be only a few
percent improvement. The tube type, and even waveforms in the cathode
circuit, affect efficiency much more than eliminating a few shorted turns
in the tank.
73 Tom
W8JI Tom%1:2619/211.9
To: W8JI Tom%1:2619/211.9%12:320/100.666
From: w8j...@aol.com (W8JI Tom)
Subject: Re: Question abt Push-pull RF Amplifier plate tank circuits
Organization: America Online, Inc. (1-800-827-6364)
Hi Hipiket,
73 Tom
# Origin: Usenet:America Online, Inc. (1-800-827-6364) (1:2619/211.9)
# Origin: Gateway ARNet <-> HamNet by HB9EBW (12:320/100.666)
--
|Fidonet: W8JI Tom%1:2619/211.9 2:301/249.666
|Internet: W8JI..Tom%1:2619/211.9
|
| Standard disclaimer: The views of this user are strictly his own.