On Mon, 03 Oct 2005 16:11:05 GMT, "Asimov"
wrote:

My choice would be not to use anti-parallel diodes in a design but this
is what I found in the circuit so some engineer thought about it enough
to use them. I would have preferentially used a reverse biased diode
across the active device's junction because it is usually the reverse
polarity in a pulse that damages a junction.

Generally a good design and a robust device will be tougher than the
diodes frequenty used.

For example I've found JFETS will easily stand 40+volts on the gate
before failing, MOSFETS (dual gate 3n200 types and BFR998) 60-80V
were needed to kill them. Even a lowly 2n5176 bipolar will be grossly
overloaded in a wide band amp before possibility of destruction.

Remember .5V PP (I'll assume silicon diodes) is at or around -1 to
0DBm. Thats one milliwatt of RF. the real issue is with wide band
signals you can have multiple signals within the bandwidth that at
some point in the cycle are additive and as we've all said those
diodes will do bad things under those conditions.

In this question I was more interested on the effect of the protection
diodes' capacitance because as you know it is greatest at zero bias.
This wideband pre-amp's input stage uses feedback to define the input
impedance and the diodes' capacitance is at the node point, so I
wondered if there was little effect. As you know, in an inverter the
node point is ideally a voltage null.

The capacitance if of limited significance. Its' effects will be
tempered by topology and impedences at the point of use.
As pointed out, they must not see conduction or "bad things happen".

_IF_ you really must protect the input with diodes put a healthy
(6-12V) of reverse bias on them. That should adequately protect
the first device but still absorb transients.

For instrumentation applications (high impedence) they are a
commonplace thing as the source could be anything or known.

For RF use I tend to leave them out as most devices once in circuit
are robust as is. And usually uf there is an event big enough to fry
the device then protection diodes are just two more fried devices.

One exception: series diodes on the supply for reverse polarity
protection and Zeners(or transzorbs) on the supply for overvoltage
protection. Often destruction didn't come down the antenna but in
the back door via the power or other control lines.

Thank you both for the great debate!

As in most electronics there are subtle and gross effects. Being
aware of both are of importance. Hope it helps.


Allison
KB1GMX