Id start with a known functional design and work from there. The theoretical
part is still black magic and I havent found any software except Microwave
Office to be of any use. And even Office can lead you a merry chase that
requires enough cut and try to make it almost not worth the effort. I no
longer have access to this very expensive program.
The only parameters I really depend upon is the published tube C and the
calculated plate load. Those are for the starting point only. The next most
important is the size of the box since it is part of the circuit and its
best to follow someones elses work if its of known quality.
I like the split line circuit for 144 and 222 as it lends itself to easier
modification to get it right and gets the HV off the output side of the
line.
Ive mentioned it here before, my 2M amp is a 3CX1000A7 (1500W Pd per Eimacs
revised specs) TV translator pull, a fairly scarce tube but in my case 25
years ago the price for tube and socket were far below a 8877 pull. It still
puts out the same power as when built.
And contrary to a lot of mythology a small vacuum variable works very well
at 2M. The end result is an easy 1500W out with around 60% efficiency and
absolutely no tuning drift or controls instability, even at full bore
maximum 1800W ( 3300V @ 900ma) for a full minute key down. The driver is a
RF Concepts 180W brick rebiased for SSB and driven for 100W out. The amp is
capable of quite a bit more but the pair of 12uF 4KV oil caps with good old
GE Pyranol keeping them happy is the limiting factor. The next tap on the 2A
CCS transformer results in 3900V...too close for comfort!
The 222 version is an 8877 using a handbook design that was tweaked for best
efficiency + IMD which is around 55% given the accuracy of the Bird.
The only test equipment I used besides the basics was an old Measurements
Model 59 GDO using a transverter/receiver as the frequency meter and the
3000W Bird/RCA load with a tapered input and useable to 1296 at reduced
power. I suppose an antenna would make a good load also using a common
birdie frequency. (-;
If using a different tube than published designs try and find one with the
same operating output impedance and then you have just the internal
characteristics and input to play with.
Final testing is done with a spectrum analyzer where its a huge help with
input and output tuning to get the best IMD. The nice thing about these
frequencies is to tune it once and leave it alone.
Carl
KM1H
.
----- Original Message -----
From: <TexasRF@aol.com>
To: <sub1@rogerhalstead.com>; <amps@contesting.com>
Sent: Friday, February 12, 2010 7:17 AM
Subject: Re: [Amps] Design of 2-meter strip lines or parallel lines
>
> Roger, I have built my share of VHF, UHF and Microwave amplifiers through
> the years. In the process, I have learned a lot about how they work and
> will
> share a few "secrets" with you.
>
> The book you are looking for does not exist. You are left with three
> options: 1>copy a known good design (such as the W6PO 8877 2m amp), 2>wing
> it on
> your own and be prepared for some cut and try (that is how the known good
> designs were developed) or 3> buy a new or used unit.
>
> You ask: why can't I just design the amplifier? The reason is that the
> internal inductive reactance of every tube model is different and get
> this:
> the inductive reactance changes with frequency. As the inductive reactance
> rises with frequency, the capacitive reactance between tube elements
> lowers
> with frequency. There is even a frequency that the two reactances are
> equal
> and look like a short circuit. Above that frequency the reactance between
> the tube elements turn inductive instead of capacitive. This is at the
> tube
> connection points, the capacitance is still there internally.
>
> The tube manufacturers for some reason have never characterized the
> internals of the tubes to allow us to accurately predict the inductance at
> various amateur frequencies. One of the reasons may be because the tube
> internal
> connections become part of the resonant circuits. If your plate resonator
> is
> a stripline the Zo of the line is very different compared to a coaxial or
> waveguide design.
>
> At HF, the variation of this reactance is not as severe and we can use
> the
> input and output capacitance and be close enough for matching network
> designs. Not so at VHF and UHF. Even a small 15 pF capacitor has the same
> problem. For example, one with .25" leads has enough series inductance to
> make
> the 15 pF work as if it were 1000 pF or so. With longer leads, it is no
> longer a capacitor for the frequency of interest but looks like an
> inductor.
>
> At these frequencies, you do not get to choose loaded Q as one might wish
> to deal with various plate load impedance variations caused by changing
> the
> plate voltage or current. You have take what you get as the loaded Q can
> be
> 60 to 100 or more at 70cm. Even on 2m typical loaded Q can be 20 to 30.
> There is little you can do to lower these numbers other than using good
> common sense: minimize any shunt capacitance, use the highest quality
> resonator
> design (strip lines are not the best), Use the lowest plate load
> impedance
> possible with the tube you have (maximum plate current, less than maximum
> voltage). And finally, pay attention to rf current flow in the plate
> resonator. If a strip line is used, nearly all of the current flows on
> one side of
> the tube anode. There is that inductance mentioned above and every
> inductor has a given Qu and resulting loss resistance. If most of the
> current is
> flowing on 1/3 of the tube plate connection, there is 3X current compared
> to
> an optimum design. Losses being related to current squared, they are 9X-2
> or 7 times an optimum design. These losses are subtracted from your
> output
> power and can damage the tube or fingerstock. Also, in some tubes, such
> as
> GS23B, the material being heated by the losses will change dimensions
> enough that the resonant frequency drops. With the loaded Q being so
> high, it
> takes a few seconds for the power output to drop to half or less. Of
> coarse
> you can keep one hand on the plate tuning control and bring it back but
> who
> wants to do that? These problems go away with a circular resonator
> design.
>
> Did I mention that I am down on stripline designs?
>
> By now, you have found the ND2X website with all of the various Russian
> tube designs. There is a wealth of information there and most of it is
> good.
> A few of the published designs by YU1AW are from computer models and do
> not
> work in practice. Beware of any designs designated "lazy builder". I
> recently helped W9IIX with a GS35 222 MHz amplifier project started from
> the
> YU1AW info and it was a ball of flames at turn on.
>
> I will be happy to critique any design you might wish to try. In fact, I
> will offer unending moral support for your VHF or UHF amplifier projects.
>
> 73,
> Gerald K5GW
> eme capable on 6m, 2m, 135cm, 70cm, 23cm, 13cm, 9cm, 6cm and nearly ready
> on 3cm
>
>
>
>
> In a message dated 2/12/2010 3:36:42 A.M. Central Standard Time,
> sub1@rogerhalstead.com writes:
>
> I've brought this up before on the Amps reflector, so far I've not come
> up with any information.
> I'm looking for information (book or software) that doesn't cost a
> fortune, for designing a legal limit 2-meter amp with a fair amount of
> overhead so it can just loaf along using either one or two tubes in a
> strip line or parallel lines.
>
> I've spent hours searching, but most just lead me in a circle.
> I know absolutely zip about 1/2 wave and 1/4 wave strip lines and how to
> calculate their dimensions.
> The same is true of the parallel lines. How does the construction and
> dimensions change with power.
> How about a pair of GS-23B Parallel lines, or strip line, or a single
> GU-84B/4CX2500, or a pair of GU-74Bs/4CX800's. On HF or even six meters
> I can come up with a tank circuit, on two about the only thing I can do
> so far is to copy the dimensions of the strip line out of my Henry 2002A
> which uses a single 3CX-800. If I went to a pair of them, or the
> 4CX-800s how would that affect the dimensions of the strip line, or if I
> replaced the one or two 3CX-800's with a pair of GS-23Bs/4CX1600U (IOW
> it's a microwave version of the 4CX1600B which was so popular until it
> disappeared although I see Alpha has a few of them yet.
>
> At any rate, this is the kind of information I'm looking for. The
> problem is being able to find it, or construction articles with a bit of
> theory, books, or design softtware that doesn't cost a fortune.
>
> 73
>
> Roger (K8RI)
> _______________________________________________
> 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
|