2 wrote:
>
>2, 10-ohm, 10Wers in series would be enough for glitch protection,
In the Europe the "best buy" are the 14W Welwyn series, as shown in
http://www.farnell.com/datasheets/909.pdf
10-ohm resistors of this type are very comfortable at <1A intermittent
service - but they do run hot! Two or three in series make a nice glitch
resistor for amps of the 1-1.5kW class.
These resistors have a nominal voltage rating of 750V, but that's the
steady-state voltage limit for the higher-resistance values. In practice
the 10-ohm types will withstand their share of 3-3.5kV in a "crowbar"
glitch test, for as long as it takes to turn the power supply off.
> but a huskier, 3A diode would still be needed.
This diode only functions during a glitch - in other words, any severe
current surge (and let's not ask why :-) The maximum possible surge
current is limited by the glitch resistor, but it's still about 10x the
normal current rating of a 3A diode, so we're relying on the surge
current rating of the diode.
What makes me slightly uneasy is that the surge current rating comes
with a specified time limit. If the surge is larger and/or longer, the
diode will fail completely. And then we can only hope that diode will
fail short-circuit... which it usually will, but obviously this isn't a
property that the manufacturer guarantees.
The 6A diodes in the Triode Board may seem over-engineered, but 200+
users have invented some interesting practical tests! Two of the best
reported were a direct short from B+ to chassis, *before* the glitch
resistor; and a glitch in a YC156 amp where the resistor couldn't stand
the sudden 5kV and arced from end to end. In both cases it was like the
glitch resistor wasn't there at all, so the surge must have been
heading for 100A... but the 6A diodes came through OK, and so of course
did the meters.
--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek
|