3CX800 amp(s)
Michael Owen W9IP/2
Well, it's a puzzler for me, anyway:
I want to build a couple of VHF power amplifiers using the
3CX800 triode. This is no problem, since there are several
designs published. But I would like to run them all from a common
HV power supply. Is it OK to just connect each amplifier's B+, B-,
and chassis grounds together? Actually, the real question is
what will happen to cathode-current metering? I gotta admit, triodes
in grounded-grid baffle me. Any help out there?
Along a similar line, can you think of any way in the world to
also run a grounded-cathode tetrode (ie 4CX250 or 4CX1000) from the
same HV supply as the triodes, at the same time? I don't think it's
possible, but I could be wrong. (where will the B- be, at chassis or
floating?)
E-mail direct if you don't want to QRM the no-coders.
Mike
Gary Coffman
>OK, all you vacuum-tube wizards. Here's a puzzler for you.
>Well, it's a puzzler for me, anyway:
>
>I want to build a couple of VHF power amplifiers using the
>3CX800 triode. This is no problem, since there are several
>designs published. But I would like to run them all from a common
>HV power supply. Is it OK to just connect each amplifier's B+, B-,
>and chassis grounds together? Actually, the real question is
>what will happen to cathode-current metering? I gotta admit, triodes
>in grounded-grid baffle me. Any help out there?
Well it's not clear if you want to run the amplifiers simultaneously
or not. If you are only going to run them one at a time you can
parallel the power supply connections, or better, you can switch the
supply from one to the other. If you plan to have more than one
amplifier active at the same time, then you must provide power supply
decoupling for each amplifier to prevent RF from one getting into the
other. This would be in the form of chokes and bypass caps mounted on
each amplifier chassis. To meter cathode current, you will need a
cathode resistor in each amplifier bridged with it's own metering
circuit rather than a single meter in the power supply across a common
cathode resistor.
>Along a similar line, can you think of any way in the world to
>also run a grounded-cathode tetrode (ie 4CX250 or 4CX1000) from the
>same HV supply as the triodes, at the same time? I don't think it's
>possible, but I could be wrong. (where will the B- be, at chassis or
>floating?)
Again you must use power supply decoupling to prevent RF feedback.
The B- will be commoned to chassis. For cathode metering of the
triode, you must supply a cathode resistor bridged by the metering
circuit. The two stages may be link coupled or a series resonant
input circuit may be used for the triode. You will need a screen
supply for the tetrode. You may derive this from HV with a suitable
regulator circuit or use a separate supply. The separate screen
supply has the advantage of allowing you to switch the amplifier
on and off with the screen supply for TR purposes. You will also need
a control grid bias supply. That should be separate.
A few safety notes. If your cathode resistor fails, the metering
circuits can become hot with HV. Mount them appropriately. Never
operate a tetrode with screen voltage and no plate voltage or
bye bye screen grid. Arrange your circuitry so that turning off
the HV also turns off the screen supply. There is still a danger to
the tetrode if you use a fuse in the HV. Consider a sense circuit
that interlocks HV with the screen supply. Your control grid and
screen supplies should be very stiff. There should be no change
in voltage with drive level.
Gary KE4ZV
Perry Scott
>3CX800 triode.
>and chassis grounds together? Actually, the real question is
>what will happen to cathode-current metering?
I've built the 2KV supply, with grid and screen supplies for a multiple
4CX250B HF amplifier. My strategy is to connect B-, chassis, and
electrical ground together. I'm using some sockets that allow me to
insert metering resistors between the chassis and cathode. Keep the
metering resistor and bypass cap leads short if you're doing VHF. I
drive the plate and screen transformer from the same point, so plate is
guaranteed whenever screen is available. (Thanks for the tip, Gary, I
think I'll just let the main fuse do the work, and not fuse B+)
>I gotta admit, triodes in grounded-grid baffle me.
You and me both. After reading the ARRL Handbook's 4CX250B 2m amplifier
article, it appears the GG's biggest advantage is that it lowers stage
gain, making self-oscillation less likely. With careful design, you can
use common-cathode, and get outrageous power gains. I've taken up the
gauntlet and chosen the latter.
Perry Scott / KF0CA
Jim Grubs
> From: pe...@hpfcdc.HP.COM (Perry Scott)
> Date: 4 Jan 91 20:49:08 GMT
> Organization: HP Fort Collins, Co.
> Message-ID: <788...@hpfcdc.HP.COM>
> Newsgroups: rec.ham-radio
>
> >I want to build a couple of VHF power amplifiers using the
> >3CX800 triode.
> >Is it OK to just connect each amplifier's B+, B-,
> >and chassis grounds together? Actually, the real question is
> >what will happen to cathode-current metering?
>
> I've built the 2KV supply, with grid and screen supplies for a multiple
> 4CX250B HF amplifier. My strategy is to connect B-, chassis, and
> electrical ground together. I'm using some sockets that allow me to
> insert metering resistors between the chassis and cathode. Keep the
> metering resistor and bypass cap leads short if you're doing VHF. I
I read a terrific Eimac application pamphlet about 30 years ago describing an
amplifier design in which the chassis was at screen voltage potential and the
screen grids were "grounded" directly to the chassis at the socket pin,
eliminating (reducing) parasitic oscillation problems do to resonance in the
screen bypassing components.
Those Eimac application notes were a treasure house of great ideas in
amplifier design and mechanical construction techniques. I used to drool over
them by the hour. Of course, as an impoverished student I could never afford
to implement any of the ideas.
--
Jim Grubs - via the friendly folks at UUNET
UUCP: ...!uunet!w8grt!jim.grubs
INTERNET: jim....@w8grt.fidonet.org
Gary Coffman
>
>I read a terrific Eimac application pamphlet about 30 years ago describing an
>amplifier design in which the chassis was at screen voltage potential and the
>screen grids were "grounded" directly to the chassis at the socket pin,
>eliminating (reducing) parasitic oscillation problems do to resonance in the
>screen bypassing components.
The great advantage of this approach, from Eimac's point of view, is that
it sells a lot of tubes. Modern power tetrodes can't tolerate even momentary
excess screen current. Some method of near instantly opening the screen
circuit is essential to the survival of these tubes. Even overcurrent
relays are often too slow to save the tube. Also, if you lose plate voltage
for even an instant before dropping screen voltage while the tube is under
drive, bye bye tube. The best circuits I've seen for screen protection use
a programmable switching supply with very small capacitors that is turned
on and off by system controls including overcurrent sensing and power
sequencing. Even this is tricky because some power tetrodes normally run
with a slight negative screen grid current so your supply needs to be able
to serve as either a sink or source of current while maintaining a stiff
screen voltage.
The simplest circuit derives screen voltage from the plate supply via a high
value dropping resistor that doubles as a current limiter, with a shunt
regulator, either gas or solid state, that sets the screen voltage. This is
practically fool proof. If the regulator fails though, you'd better have
good plate overcurrent protection too. Of couse the plate circuit can stand
a lot more overloading before causing permanent damage to the tube so a
dropout relay or even a fast fuse will do nicely. You can TR the amplifier by
grounding the screen, a very good, fast method of TR. I've had a pair of
8890s running in this circuit for a couple of years with no problems.
Gary KE4ZV
Peter Beyer
>The great advantage of this approach, from Eimac's point of view, is that
>it sells a lot of tubes. Modern power tetrodes can't tolerate even momentary
>excess screen current. Some method of near instantly opening the screen
>circuit is essential to the survival of these tubes.
Don't agree with that.... First of all, grounding the Screen and putting the
Cathode on a negative voltage is just a way to eliminate the Screen bypass-
capacitor hence creating a stable environment. Second, never ever let the
screen or any elctrode in the tube get floating. It's a way of asking for
flashovers!
Grouding the Screen is used by Collins many years ago in their amplifiers
already. I'am running an EME amp using this method with no problems, even when
I deliberate let it flashover, the tube survives...
Peter Beyer - PA3AEF DEC Easynet: utrtsc::beyer
Digital Equipment bv UUCP: {decvax,ucbvax,allegra}!
Holland Internet: decwrl!utrtsc.dec.com!beyer
be...@utrtsc.dec.com
beyer%utrts...@decwrl.dec.com
Perry Scott
What I've built for the screen supply is a voltage regulator with a
current limiter. The current limiter puts a sense resistor in the
supply line, then steals base drive from the pass transistor when
the sense voltage forward biases a 2N2222A.
The basic circuit is described in the ARRL Handbook section on voltage
regulators. The only tricky part is finding a pass transistor that will
handle 400+ volts Vce.
Re: grounded screen circuit
EIMAC makes screen bypass rings (at least for the 4CX250B genre), which
look serious enough to eliminate all parasitics. The rings are clamped
directly to the chassis, and the capacitors connect to the screen ring
via short spring stock.
Perry / KF0CA
Gary Coffman
>In article <19...@ke4zv.UUCP>, ga...@ke4zv.UUCP (Gary Coffman) writes...
>
>>The great advantage of this approach, from Eimac's point of view, is that
>>it sells a lot of tubes. Modern power tetrodes can't tolerate even momentary
>>excess screen current. Some method of near instantly opening the screen
>>circuit is essential to the survival of these tubes.
>
>
>Don't agree with that.... First of all, grounding the Screen and putting the
>Cathode on a negative voltage is just a way to eliminate the Screen bypass-
>capacitor hence creating a stable environment. Second, never ever let the
>screen or any elctrode in the tube get floating. It's a way of asking for
>flashovers!
By running the cathode at a negative 1100 or so volts you've now got the
problem of adequate cathode bypassing. Since the cathode circuit may be
flowing an ampere or more of RF current this is a much more difficult task
than bypassing the screen grid with it's one or two milliamperes of RF
current. Also you now absolutely require a separate screen supply and you
still must limit screen current to a safe value under all conditions.
In addition you must now supply a control grid voltage about 50 volts
more negative than the cathode, yet another ground isolated supply, and now
your input circuit must be designed to block a much higher than normal DC
voltage requiring either link coupling or wide spaced variable capacitors or
both. Running the cathode so negative runs the risk of filament to
cathode arcing unless the filament supply is isolated from ground by
at least 2000 volt insulation. This seems a rather high price to pay to
avoid a screen bypass capacitor. Modern power tubes like the 8890 have
annular grid rings and it is easy to make the socket so that the grid
fingers are attached to a sheet forming one plate of the bypass with the
other plate being grounded and forming part of the output cavity wall. A thin
mica or mylar sheet forms the insulation between the plates.
I agree that letting the screen float is a sure plan for disaster. If your
screen protection circuit opens the path to your screen supply you must
supply a shunt path to ground. A high value resistor to ground directly
connected to the screen is a good idea when using a separate screen supply
in any case. A steady load like this on the screen supply will help it
maintain stiff regulation. A good value for this shunt is one that will
draw a static current ten times the maximum allowed for the screen grid.
>Grouding the Screen is used by Collins many years ago in their amplifiers
>already. I'am running an EME amp using this method with no problems, even when
>I deliberate let it flashover, the tube survives...
What power tetrode are you using that doesn't mind flashovers? Most modern
tubes are damaged by flashovers. Usually cathode material is sputtered
onto the envelope and grid structures causing eventual failure through
internal shorting. The problem is progressive, one flashover may not
kill the tube, but each additional flashover deposits more material on
the enevelope and grid structures eventually killing the tube through
a short.
Gary KE4ZV