[Elecraft] Interpretation of P3 display

[Guy Olinger K2AV]

Gu74B, aka 4CX800A, is a tetrode, a pair used in Alpha manual tune amps
since the 91B, replaced recently with 4CX1000A's in the Alpha 8410.  The
4CX1000A's have the same protection needs.  The same protection needs exist
for the 4CX150A's in my ancient Tokyo Hy-Power HL-1K, but it's not there
and if you spike 'em, you can blow 'em.

*** Spikes are bad with tetrodes. ***  Need a minor bit of theory to
explain.  Oversimplifying  just a little...

Some amps can tolerate spikes, but nearly all screen grid tubes commonly
used in ham amplifiers depend on running with VERY small control grid
current, single figure milliamps. The control grids for these tubes run
with negative bias that exceeds the normal positive going peak RF drive
signal value.  Therefore any level of signal whose positive peak exceeds
the negative grid bias will go into non-linearity and excessive current.
 Alpha's and other amps consider this a fault condition in and of itself,
REGARDLESS OF DURATION.  This fall-off-the-cliff point in drive voltage is
most UN-like the zero-bias triodes in grounded grid amps which always have
significant grid current under normal drive and loading, and get to
over-drive in a far more "gradual" way as you increase drive power.

SINCE most ham amp tetrode transmitting tubes are designed to work in that
no-grid-current zone (look up AB1 in the tube operation section of the ARRL
handbook), they do NOT have the huge clunky grid structure of a 3CX1200A7
or 3-1000Z.  Tetrodes have very fine control grid wires, JUST BECAUSE they
supposedly don't need to carry anything except very low current.  There are
some very useful engineering advantages to these fine wires which I won't
get into.

Amps with Gu74B's and like tubes in a grounded cathode configuration (as in
the 8410) will require a 50 ohm power absorbing circuit (aka swamping)  in
the input circuitry, in parallel with the control grids, so that the
transceiver does not see the very high natural resistance of the biased
control grids themselves.  Normal drive to full power output in these amps
does NOT require drive voltage that exceeds the grid voltage.  A pair of
4CX1000A with drive power right at the grid bias is hardly legal in this
country.  Pair of GU74B also illegal this way, though not so blantant.

If you look at the front panel of an 8410, you will note there is no way to
measure grid current.  Just a green light and a red light.  The green light
just barely flickers under correct operation.  Solid green means you are
pushing it a little, and there is a certain amount of capacitive current in
the grid.  Red means you are creeping up to the fault level and you need to
back off.  All of these conditions are tiny single digit milliampere
currents that can occur without going past the control grid bias level.

The full scale value for grid current on my HL-1K is THREE milliamperes,
where you are never supposed to be.  Little kicks on the meter below ONE
milliampere on the scale are what you want.

Now back to the spikes...

The spikes look just like fault conditions to the amp.  To a protection
circuit that depends on sensing what would be destructive continuous
overdrive in microseconds and then engaging in a few milliseconds, the
spike looks exactly the same as driving the grid with a continuous PEP the
same as the spike peak power.

Whether a single un-faulted spike could DESTROY the grid in it's time is a
kind of Russian roulette question.  If the spike was on the leading edge of
every single CW baud in a contest, you have your answer.  Kind of like if
someone in a bullet proof vest was shot over and over, how long would he
last?

Absolutely, when the spike-ee is a pair of now very expensive 4CX1000A or
4CX800A, you start fault detection proceedings the microsecond after the
input goes over the engineered threshold.  Meaning a spikey rig simply
could NOT be used with an Alpha 91B through 8410.  You could reduce power
output so the spike was the normal PEP for full power drive.  But on the
output, on average it would look like you were running 1/2 or 1/3 power.
 Zounds.  Who would put up with that?

Hams should be loudly, blatantly, zealously, continuously,
in-their-face-nasty, heavy-push-back, we-won't-put-up-with-your-crap
against rigs with spike problems.  Kind of like cars that lose wheels at
highway speeds.  Just simply not permitted, ever.  Do it and somebody goes
to jail for negligence.  And anyone making spikes by how they use their ALC
should be buried under a pile of rotten cabbage heads and tomatoes.
 Spiking things up should have roughly the same social value as repeatedly
loudly passing wind at Christmas dinner.

 Should I tell you now what I really think?  I guess I have an attitude
about that  :>)

73, Guy.

On Sat, Jun 9, 2012 at 3:41 PM, Roy Morris <w4wfb at carolina.rr.com> wrote:

> David, what are you saying?  Do the GU74bs dislike or do they tolerate the
> spike?  You left out an important word in your previous statement.  Thanks,
> Roy  W4WFB
>
>