To expand on last night's comments from Tom W8JI :
This kind of "glitch" is a big surge of current, out of the power supply
and through the rest of the amplifier.
For this discussion we don't need to ask why a current surge might
happen. Even in the most bitter disputes, there has been an underlying
agreement that such things *can* happen, and that many amplifiers *will*
experience a big current surge at some time.
But even if it does happen, it doesn't have to wreck your amplifier.
Most glitches should be completely recoverable, without even taking the
covers off.
The aim of the "glitch" resistor is to limit the peak current out of the
power supply, and thus to limit the possible damage. This buys enough
time for something else (a breaker or a fuse) to disconnect the power
supply from the mains.
Eimac Application Bulletin 17 is very clear about the need for a glitch
resistor:
http://www.cpii.com/eimac/ab17.htm
However, many older amplifiers don't have any glitch protection, so
there is a lot you can do to make things better.
>
>If you want a glitch resistor, use a high voltage pulse rated resistor
>of some reasonable value. 10-20 ohms would be the minimum useful value
Why "high voltage pulse rated"? Because for a few milliseconds, anything
up to the full B+ voltage can appear across that resistor. It must not
arc from end to end.
The glitch resistor is ABSOLUTELY NOT intended to be any kind of fuse.
Exactly the opposite - it has to HANG IN THERE and do its job of
limiting the immediate surge of current, for a few milliseconds until
something else has had time to disconnect the power.
There always has to be a genuine fuse or circuit breaker to disconnect
the source of power... but the glitch resistor gives it time to do that.
I would favour disconnecting the mains supply to the transformer
primary. There are many kinds of fast fuses and breakers that can do
that. (Another tip: if you also install a step-start, it allows you to
reduce the rating on the primary breaker/fuse, which protects the whole
amplifier much better.)
Personally I don't like high-voltage fuses. They certainly don't take
the place of a current limiting resistor - remember, a fuse has NO
current limiting capability for all the time it takes for the wire to
melt, pull apart making an arc, and for the arc to die away. Simple HV
fuses are very prone to prolonged arcing, so they actually aren't very
fast. By contrast, the current limiting that you get from a resistor is
instantaneous - and that is where the main protection comes from.
>The resistor belongs in the HV supply line just after the filter caps.
That's right. When using a remote HV supply, always put the glitch
resistor in the power supply - never in the RF deck (think what happens
when there's a short in the connecting cable).
Tom suggests 10-20 ohms as a minimum, and allowing for the internal
resistance of the power supply, this is consistent with Eimac's
recommendation of 20-50 ohms.
A good glitch resistor needs to have a continuous power rating of about
40-50W to handle the normal tube operating current reliably, and it
needs to be physically long to give it a high pulse voltage rating
without arcing over from end to end. In other words, DON'T use resistors
with both leads coming out the same end, and DON'T use the metal-bodied
resistors (they have quite poor breakdown ratings to the metal body,
which is 'almost' connecting both ends together).
Tom has frequently recommended the 'globar' tubular silicon carbide
resistors (as also used for RF dummy loads). Another type of resistor
that is very useful for upgrading older amplifiers is the long-bodied
glass-coated wirewound kind, such as:
http://www.farnell.com/datasheets/909.pdf
Although even the largest 14W size doesn't have high enough power or
voltage ratings, these resistors are very convenient for connecting in
series. Three 10-ohm or 15-ohm 14W resistors in series will make a very
nice glitch resistor for amplifiers up to the 1500W class.
In smaller amps where there is very little space to break into the
existing HV wiring, you can generally manage to squeeze in two of these
resistors connected in a /\ layout. Don't worry too much about the 750V
maximum end-to-end rating; in practice it is much more.
OK, handing it back to the afternoon shift...
--
73 from Ian GM3SEK
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|