EMRFD Message Archive 103

Message Date From Subject
103 2006-08-23 15:01:32 Stan Zener protected final stage
I thought that if you had a high capacitive load for the final
transistor you did not need a zener. However, in the W7EL
transceiver written up in ARRL publications he has a 390 pf at the
collector to ground but still has the zener. I read someplace that it
worked due to the capacitance load from the zener (?). Since most
transistors switch 10s of times faster than a zener I would think the
output transistor would be blown out long before the zener conducted.

What is the best way to do away with the zener in a RF stage?

Should we go to transformer coupling?

Do we really need a zener or is this just a old wife's tale from the
early semiconductor days?
104 2006-08-23 17:00:08 Graham Haddock Re: Zener protected final stage
Stan:
I have also read that it is the parallel capacitance of the
zener that helps stabilize the output stage in certain designs.

As far as to whether it will conduct when the zener voltage is
exceeded, the answer to that is yes, and probably in time to limit
the peak voltages on the output transistor, even if it took
multiple RF cycles to go into full conduction. There is a
time dependency involved with the breakdown voltages when
you are dealing with sub-microsecond pulses.

As far as a design without one, yes, you can take it out, but
then you need to make sure that the PA you are building will
not exceed the breakdown of the transistor under some reasonable
range of VSWR, say 10:1 all around the circle on the Smith chart.
Most hams don't have access to the type of test equipment that it
takes to run this test.

One easy way, in the absence of the ability to do load-pull testing,
would be to not push the breakdown voltage of the PA device,
that is, make sure that the breakdown is at least three or four
times the supply voltage to the PA. That is overdesigning,
but it makes it very unlikely the PA device will see breakdown
voltages, even under a fault condition.

Very few commercial PA designs use Zener diodes bridging the
PA output, but they test the designs extensively under fault
or other mismatch conditions.

In my own experience, I have found that self oscillations
are an enemy of the the PA, I have had a 10 MHz PA break into
oscillations under no load, and blow out the Zener, but
leave the PA transistors intact, so they do seem to work.

105 2006-08-23 21:00:57 Bob Larkin Re: Zener protected final stage
Hello Stan,

While putting EMRFD together, I discussed this question of Zeners for
Class-C amps with Wes. In the process, I put together a test circuit and
found that the effect is real! The side bar on pages 2.34 and 2.35 tells
the story. In a bad case, seen in the 3rd 'scope picture, a 5VDC supply
results in 40 to 45 Volts peak at the collector of the little 2N3904. In
addition, note that:

1-The impedance seen by the transistor is the net of the amplifier matching
components and the external load. It can always be made inductive with some
load Z. Capacity in the circuit does not mean ensure a lack of big peaks.
Resistive loading will reduce the peaks.
2-If you can guarantee a proper resistive impedance, as seen by the
transistor, no Zener is needed.
3-If you use an adequate VSWR shutdown arrangement no Zener is needed. This
added complexity is probably more appropriate for a "big" amplifier than
for most QRP projects.

BTW, Im headed away for the weekend, so if I ignore any responses for a few
days, this is why!

73, Bob Larkin W7PUA

At 09:48 PM 8/23/2006 +0000, ak0b wrote:

>I thought that if you had a high capacitive load for the final
>transistor you did not need a zener. However, in the W7EL
>transceiver written up in ARRL publications he has a 390 pf at the
>collector to ground but still has the zener. I read someplace that it
>worked due to the capacitance load from the zener (?). Since most
>transistors switch 10s of times faster than a zener I would think the
>output transistor would be blown out long before the zener conducted.
>
>What is the best way to do away with the zener in a RF stage?
>
>Should we go to transformer coupling?
>
>Do we really need a zener or is this just a old wife's tale from the
>early semiconductor days?
107 2006-08-23 21:35:57 Wes Hayward Re: Zener protected final stage
Hi Stan, et al.

As Bob (W7PUA) mentioned, we did some measurements that we reported
in EMRFD on this Zener thing. Actually, Bob did the measurements,
for he had the faster oscilloscope. The Zener diodes really do
turn on and catch the high voltage spikes.

We don't see Zener diodes in our transceivers because, among other
things, they are usually push pull circuits. When one transistor
is turning off, the other is turning on, and this tends to restrict
the voltage spikes.

There is another thing happening in these circuits. The Zener
diode is usually embedded within a low pass filter. This filter
keeps harmonics from escaping to the outside world. It also
constrains the rise time. So we don't have nanosecond spikes of
energy. Rather, they are usually a few microseconds. And the
Zener diodes really will turn on this fast.

Roy, W7EL, reported in his classic article on his Optimized QRP Rig
in QST, August, 1980, that the usual Zener diodes we used on the PA
collector has quite a bit of capacitance. Hence, we need to account
for that when selecting capacitors for that part of the circuit.
But this is just one more parasitic (undesired) capacitance. It is
the Zener action that we are seeking, and not the capacitance.

73, Wes
w7zoi