It is always best to have the GG amplifiers tuned input circuit as close to
the cathode as possible.
There needs to be capacitance directly from the cathode to ground. A pi
network input is ideal and easiest to deal with. A parallel tuned input
circuit can also be used as the tune capacitor is directly from the cathode
to ground but it also needs to be very close to the cathode.
The rule of thumb as I recall, there needs to be about 13 pf per meter of
capacitance from cathode to ground. The 10 meter band would require about
130 pf of capacitance at the cathode side of a pi network.
Without this capacitance directly at the cathode the efficiency also suffers
as there is not a direct path to ground for the harmonic currents in the
cathode.
A tube type exciter with pi network output is better than nothing but is not
ideal as the output capacitor of the pi network is too far away from the
cathode of the driven amplifier.
Also using an external antenna tuner such as is found used with most
transistor type radios is usually a T type network. This doesn't provide the
required capacitance to ground for the cathode unless the amplifier has its
own built in pi network input matching system.
73
Gary K4FMX
> -----Original Message-----
> From: Amps [mailto:amps-bounces@contesting.com] On Behalf Of Manfred
> Mornhinweg
> Sent: Friday, July 25, 2014 11:54 AM
> To: amps@contesting.com
> Subject: Re: [Amps] SB-200 input.
>
> Hi all,
>
> adding to Carl's comment:
>
> > Tuned inputs provide the flywheel effect that makes the tubes easier
> to
> > drive and improves IMD 5-10dB.
>
> That's indeed to biggest reason to use a tuned input. The impedance of a
> class
> AB grounded grid tube's cathode varies dramatically throughout the RF
> cycle.
> During a good part of the cycle, the tube is cut off entirely, and the
> cathode
> impedance is essentially infinite. During the rest of the cycle, the
> tube
> conducts, and the cathode has an impedance that changes depending on the
> instantaneous current. The curves of triodes are pretty crooked, and
> this
> reflects on the cathode. Having a big flywheel there, meaning a
> relatively high
> Q network of some sort, helps a lot in keeping the waveform clean, and
> in
> providing a huge drive current during that part of the cycle when the
> cathode
> needs it, while accumulating energy from the driver not needed by the
> cathode
> when it's in cut-off.
>
> The point is that with any grounded grid amplifier, you need a
> reasonably high Q
> network between the driver's output tube or transistors, and the
> amplifer's
> cathode. If the driver radio has a tuned output network, such as the PI
> tank of
> a tube rig, that's really quite enough, and a WELL DESIGNED broadband
> drive
> matching circuit can work OK. Same thing if you have a solid state radio
> with an
> antenna tuner inserted between the radio and the amplifier. In that case
> the
> tuner provides the required Q and flywheel effect - or at least we can
> hope the
> Q is high enough! But if you have a bare bones solid state radio, whose
> only
> flywheel effect comes from a 5 pole lowpass filter having a Q of unity,
> you will
> see trouble with a broadband cathode driving circuit. In that case, a
> tuned
> drive circuit works wonders.
>
> We can also see it this way: When you load a radio with a correctly
> tuned
> antenna, or with a plain resistor (dummy load), it sees a constant load
> all
> through the RF cycle, and that load is ideally 50 ohm, making an SWR of
> 1:1. But
> if you connect it to the cathode of a grounded grid tube, it sees a
> terribly
> varying impedance, with the SWR being far above 1:1 most of the time,
> and at
> infinity for a part of the RF cycle. No amount of broadband matching can
> fix
> that! Any matching transformer can only change an impedance by a fixed
> ratio, it
> cannot match to a wildly varying impedance. But if you insert a tuned
> circuit,
> which acts as a flywheel, this tuned circuit evens out the varying
> impedance of
> the cathode, and presents the average impedance to the driving radio,
> smoothly
> and cleanly. This impedance may still be wrong, but at least it's
> constant
> throughout the RF cycle, and that allows matching it to a nice 1:1 SWR
> with
> either a broadband transformer or a resonant network. Usually, when you
> go to
> the trouble to install such a resonant circuit, it's very easy to tap it
> for 50
> ohm, or to elaborate it into a PI section having 50 ohm input impedance.
> So
> that's what's usually done, rather than combining a broadband
> transformer with a
> set of simple tuned circuits.
>
> Please note that there is yet another problem: The tuned cathode
> matching
> circuit can very well even out the cathode impedance over the RF cycle,
> but it
> cannot do anything when the average impedance over the RF cycle varies
> according
> to the amplitude of the signal! So, even with a tuned matching network,
> the
> input SWR can change according to the exact instantaneous drive level.
> That is,
> an amplifier might present a 1:1 SWR to the radio at 50 watts CW drive
> level,
> but at 100W or at 20W the SWR will be higher. During SSB transmission,
> the SWR
> will be varying all the time, according to the signal's envelope. Some
> radios
> cope better with this than others. If an amplifier needs far less drive
> power
> than the radio can deliver, it's a good idea to insert an attenuator, so
> that
> the radio runs at full power, and the attenuator dampens the SWR changes
> of the amp.
>
> But the best way to avoid all this trouble is to avoid grounded grid
> amplifiers!
> If you use a grid-driven amplifier in class AB1, the tubes almost don't
> load the
> driver at all. A dummy load inside the amplifier loads the driver
> smoothly and
> cleanly. No tuned drive circuits are required, no drive distortion
> happens, and
> many tubes can be driven without needing any impedance transformation,
> and just
> a low drive power at 50 ohm. The disadvantage, of course, is that the
> tubes have
> to be tetrodes, at least, with an additional screen power supply.
> Triodes aren't
> linear enough to use them in this way.
>
> When using MOSFETs rather than tubes, the situation in this regard is
> way better
> than with grounded grid tubes , but not as good as with grid-driven
> tubes in
> class AB1. By using proper circuit design, with enough negative feedback
> and
> resistive gate swamping, normally the performance is pretty good without
> needing
> any tuned input circuit. But in extreme cases, when milking a marginal
> MOSFET
> for all the gain it can provide, leaving no room for gate swamping nor
> negative
> feedback, a tuned input circuit can be required!
>
> Many of you of course knew most or all of this, which is all pretty old
> technology, but perhaps somebody learned something new to him. And I had
> fun
> writing it...
>
> Manfred
>
>
> ========================
> Visit my hobby homepage!
> http://ludens.cl
> ========================
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|