Dick Green WC1M wrote:
> I'm a little confused about lightning suppressor theory.
You and a lot of other people...<grin>
It's called "overvoltage protection" in the literature.
"lightning suppression" is the thing where the medicine show guy in the
horse drawn wagon shoots magic dust into the air.
> Many of us have installed lightning suppressors from Polyphaser, ICE, K5FD,
> etc. to protect control lines running from towers to the shack. The
> Polyphaser units contain MOVs that turn on at 82V. ICE offers two MOV
> voltage levels, 50V and 20V. K5FD offers three voltage levels: 65V
> (standard), 35V and 17V. I have the Polyphaser and K5FD standard units in my
> suppressor network.
> So, if components in a control box are connected to these suppressors, the
> voltage on the pins can rise as high as 65V-82V with respect to ground
> before the suppressor starts to do its job. Some bipolar transistors can
> take that kind of voltage but many FETs and ICs can't. While the suppressors
> probably keep things from melting and vaporizing, isn't it probable that
> semiconductors attached to the line will be damaged?
Yes, but... There's also an energy aspect to the transient suppression.
Say you have a switch (like a sparkgap) that fires in nanoseconds,
and, after firing, has a very low forward voltage drop (that is, not an
MOV or back to back zener, which is a fixed clamp). The voltage might
rise high, but it's clamped so fast that the energy is low.
The other aspect of transient suppression is changing the power spectrum
of the impulse. For instance, turning a 1 microsecond rise transient
into a 100 millisecond rise transient, essentially takes a very high
fast pulse and squashes it out. (series mode transient suppressors do
this, and are the preferred thing for stuff like an incoming power line
filter.. doesn't necessarily work at RF though..)
Logic gate stuff that's designed to talk to the outside world (except
cheap stuff or homebrew) usually has something like a series resistor
followed by some clamps to the supply rails, so it can take a limited
energy hit of several hundred volts without trouble. (several kV in
100pF is a typical model.. down in the millijoules energy, though).
The simple "MOV across the pair" technique has some limitations.. for
instance, if you have multiple MOVs in the system, the lowest voltage
one clamps before the others. This is the problem with the plug strip
style surge suppressors.. they advertise low clamping voltage.. Great,
but if you have a whole house suppressor at a higher voltage, the plug
strip will protect the whole house suppressor.
> I ask because we took a massive lightning hit here about three weeks ago
> that damaged at least a dozen boxes in my shack. I've got an extensive
> ground system and suppressor network. The only way I can explain some of the
> damage is that the suppressors allowed the voltage to rise too high for the
> semiconductors connected directly to the control lines. Does this make
Yes. Particularly in ham equipment that originates as a homebrew
project, I've noticed that designers don't do much to provide decent
overvoltage protection. It's sort of like they get the breadboard
working on the bench with the bare microcontroller or logic circuits.
Then, they just package it up and start selling it. This is just fine,
but something to be aware of, and manage your expectations. Commercial
products intended to be sold in any sort of volume, or for high dollars,
tend to have a more rigorous design and prototype testing process where
they actually consider things like this (if you sell a million widgets,
you don't want to have to deal with 1,000 returns for static damage).
If it's line connected, there are usually things like hi-pot tests
required too in order to get that UL or other NRTL sticker which is
needed to sell devices in some areas (e.g. Los Angeles County).
This is why modern PCs are MUCH more tolerant of this kind of abuse than
they were, say, 15 years ago. Back in 1990, you could get away with
selling a laptop PC that died if you zapped it with a static after
walking across the floor. Not today.
TowerTalk mailing list