Bill,
Agree with your numbers and conclusion. The magnitude of the effects will
depend on the swamping resistor. I calculate that at 30 MHz with a 50 ohm
swamp, there will be around a 2 degree phase modulation. As you say, so far the
assumption is that the tank is resonant i.e it looks like a resistance. Any
move away from that will change the input resistance, either higher or lower.
But the low Q means that it won't change much.
IF one were to worry about it, neutralisation would be added.
73
Peter G3RZP
========================================
Message Received: Aug 12 2012, 06:51 AM
From: "Bill Fuqua"
To: g8on@fsmail.net, "Larry Benko" , amps@contesting.com
Cc:
Subject: Re: [Amps] 4CX1500B project - swamped grid
Lets see, 50 volt swing on control grid produced 2000 volt swing on plate.
Gain during this conductive part of the cycle is 40 so the Miller
capacitance during linear part of the cycle is 4.1 pF.
The non-conductive part of the cycle it is 0.1pF. That is shunted by the
tube's input capacitance, 15.7pF. So the effective input capacitance ranges
from 15.7 pF to 19.8 pF.
Now that I have lead you down the garden path, it is all wrong unless
the plate load is a pure resistance.
The Miller capacitance comes from the fact that the actual current thru
the feedback capacitance is due to the voltage across it. In other words
the voltage across the .1pF capacitance is 41 times that
at the input of the tube. 2000V swing plus 50 volts swing, The add since
they are on opposite sides of the capacitor and out of phase. But here is
the catch.
The plate load is not purely resistive. It includes a resonante circuit
that continues the sinousoidal voltage swing even when the tube is cut off.
The Miller capacitance is always there and changing very little with
a tank circuit with a Q of 10 or more.
It is late and need sleep. Please pardon spelling or grammar errors.
73
Bill wa4lav
At 03:24 PM 8/11/2012 +0200, peter chadwick wrote:
>Larry,
>Thanks. According to Pappenfus, the change in input capacity in 4CX250B
>due to variation in electron density between grid and cathode is about
>0.1pF, which can cause problems with distortion ratios as low as -50dB -
>so is negligible in practice. But amplitude non-linearity leads, in the
>case of a incompletely neutralised stage, to the input impedance caused by
>Miller effect changing in capacity, and this leads to phase distortion. I
>would have thought that with reasonable neutralising, this effect would be
>negligible, but I suppose it does depend on how far down one is looking.
>It is suggested that swamping resistance is the answe - I asked because I
>didn't think tha poor neutralising would be accepted.
>In the case of the GG stage, running AB2, the whole thing is variable over
>the RF and modulation cycles of course.
>73
>Peter G3RZP
>========================================
>Message Received: Aug 11 2012, 02:02 PM
>From: "Larry Benko"
>To: amps@contesting.com
>Cc:
>Subject: Re: [Amps] 4CX1500B project - swamped grid
>
>Peter,
>
>I can't quantify the change which is probably small but a professor I
>had in a EE class was emphatic about this topic. At the time I probably
>didn't grasp it completely. There should be a continuum of change from
>best to worst as we go from heavy class A through class C. Perhaps
>someone else can quantify these numbers. If there really is NO change
>in input impedance through the RF cycle then there is no need for any
>matching circuit to have anything more than the minimum Q necessary to
>accomplish the math.
>
>73,
>Larry, W0QE
>On 8/11/2012 2:30 AM, peter chadwick wrote:
> >
> > Larry,
> >
> > Can you explain what causes a change in input impedance over the cycle
> > when operating in AB1 i.e no grid current? I can see Miller effect
> > causing some change as the amplitude varies, but not over the RF
> > cycle, and if neutralised, there won't even be that change. The change
> > in the amount od space charge and thus capacity should also be negligible.
> >
> > Peter G3RZP
> >
> >
> >
> >
> > ========================================
> > Message Received: Aug 11 2012, 04:55 AM
> > From: "Larry Benko"
> > To: amps@contesting.com
> > Cc:
> > Subject: Re: [Amps] 4CX1500B project - swamped grid
> >
> > Hank,
> >
> > I hope someone will correct me if I am wrong but I believe the
> > problem
> > is the changing input impedance the tube/FET etc. throughout the RF
> > cycle for all classes of operation except heavy class A. This dynamic
> > changing impedance causes the exciter to generate IMD to the
> > amplifier
> > which then gets amplified. This is why we pad the impedance and match
> > to it with circuits having a larger Q than would be necessary to just
> > do a match.
> >
> > If you build an SWR circuit in your simulation software such as the
> > "tandem match" circuit which consists of 2 coupled transformers
> > and no
> > capacitors you can see the change in SWR throughout the RF cycle.
> > Also
> > since this is a simulation a resistive SWR circuit will work fine to
> > show the changing SWR. A truly constant 2:1 SWR does not generate IMD
> > but a varying 1.2 to 1.5:1 SWR will generate some IMD.
> >
> > This is how I understand the problem.
> >
> > 73,
> > Larry, W0QE
> >
> > On 8/9/2012 11:12 PM, pfizenmayer wrote:
> > > I looked at 50 ohms shunted by 80 pfd -- 2 - 30 mhz - --- series
> > L of 271
> > > nhy and shunt C of 71.5 pfd at input gives better than 23 db
> > return loss
> > > from 2 to 30 mhz.
> > >
> > > I was wondering why Alpha did not look at that for the 8410 but
> > the 4CX1000
> > > is 77 to 90 pfd input so a pair is something like 2X the 4CX1500
> > or about
> > > 160 pfd.
> > >
> > > If I add another LC section to the 4CX1500 I get virtually no
> > improvement -
> > > but if I add some series L to the 50 ohm shunting resistor it gets
> > > incredibly good , better than 30 db return loss 2 to 30 mhz.
> > >
> > > The network becomes the 80 pfd tube , 50 ohms with 53.7 nhy in
> > series with
> > > it to ground and then the L net of 245 nhy series L , 59 pfd to
> > ground at
> > > input .
> > >
> > > One would want to be a bit careful about lifting the grid off
> > the 50 ohms
> > > too much but 53 nhy may be Ok . Actually I suspect you might
> > want to use a
> > > pair of 100 ohm 50 watters .
> > >
> > >
> > > 73 Hank K7HP
> > > ----- Original Message -----
> > > From: "Bill, W6WRT"
> > > To:
> > > Sent: Wednesday, August 08, 2012 11:15 PM
> > > Subject: Re: [Amps] 4CX1500B project
> > >
> > >
> > > ORIGINAL MESSAGE:
> > > On Wed, 08 Aug 2012 21:50:42 -0700, you wrote:
> > >
> > >> Because as I understand it, the tubes are essentially the same
> > except for a
> > >> slightly
> > >> larger anode cooler which accounts for the additional
> > dissipation. The
> > >> parameters for
> > >> typical operation on the 4CX1500B data sheet are based on a
> > load line
> > >> closer to class A
> > >> than those on the 1000 sheet -- they call for lower bias and
> > screen voltage
> > >> and more
> > >> idling current. The intention is to produce much less IMD.
> > Since I'm a CW
> > >> operator, I
> > >> would rather have full legal output than reduce IMD.
> > > REPLY:
> > > Ok, but "typical operation" is just a starting point. Don't feel
> > bound by
> > > that.
> > >
> > > If you want max output and max gain, put the screen voltage
> > close to max
> > > (400
> > > VDC) and set the bias for a small amount of idle current, say
> > 100 mA or
> > > less.
> > >
> > > The higher the screen voltage, the higher the gain. The lower
> > the idle
> > > current,
> > > the higher the efficiency but the higher the IMD too. For CW, as
> > you imply,
> > > IMD
> > > doesn't matter.
> > >
> > > One thing to be aware of when using swamped grid design: This
> > tube has a
> > > fairly
> > > high input capacitance in grounded cathode mode, aprox 80 pF.
> > You will no
> > > doubt
> > > have an SWR problem on the higher bands so you may have to
> > switch in a
> > > parallel
> > > resonant circuit to compensate. Or if your transceiver has a
> > good antenna
> > > tuner,
> > > or if you have an external one, that may take care of it. Just
> > something to
> > > keep
> > > in mind when you're doing the layout. Leave room for the
> > parallel resonant
> > > circuit and switch if you need it. Better yet, breadboard it
> > first and check
> > > the
> > > SWR with an MFJ SWR analyzer or the equivalent.
> > >
> > > 73, Bill W6WRT
> > >
> > >
> > > _______________________________________________
> > > 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
> >
>
>_______________________________________________
>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
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|