>
>Paul Christensen wrote:
>>By comparison, my PA70-V uses the same darlington EBS circuit to the
>>cathode, but Q2 does not exist as this was a feature
>>"enhancement" in the PA-77 which later replaced the PA70 series.
>>Instead, a 1/4-amp grid fuse is placed in series with the 8.2 V
>>zener diode to the tube's cathode.
>>
>>Now, to Rich's point (I believe)...the fuse may in fact blow faster in
>>this arrangement than in the case of the PA-77 where a
>>sampled switching transistor operates a relay to force a standby
>>condition. The relay uses no special acceleration circuit.
>
>>So, it
>>would be an interesting test to measure the "break" time of the fuse
>
>The break time of a fuse depends crucially on the size of the overload.
>Remember that a fuse operates in a very simple way, by heating up a
>piece of very thin wire until it melts. For the fastest break times, it
>requires a huge overload that flashes the wire to vapor almost
>instantaneously.
>
>That's why fuses are great for protecting DMMs against the common
>mistake of trying to measure voltage when the meter is still set to
>measure current - the fuse blows very quickly because the overload is so
>big. That's what fuses are very good at.
>
>But an overload of only 50% can take a v e r y long time to melt the
>wire - even an "ultra-fast" fuse can take several seconds. As a reality
>check, I just tested a 250mA F-rated fuse (a spare for my DMM, but you
>might use that same fuse for grid protection) at 50% overload. Well, the
>fuse got warm, but after 60 seconds it still hadn't blown!
>
>If a fuse is only protecting something like a line cord, a 50% overload
>for several seconds doesn't mean much - but it means a helluva lot to
>the grid of a tube! That's why most amp designers prefer an electronic
>trip circuit with a sharp threshold.
>
>>versus a transistor/relay combination.
>
>Just tested that too. The circuit is a bipolar transistor sampling grid
>current, into a CMOS trigger that drives the cathode bias relay. There
>is a 1 millisecond delay on the input to the transistor, to prevent
>nuisance triggering, and the data sheet for the relay (nothing special,
>a regular Potter & Brumfield / Schrack RTE240xx 8 amp power relay) says
>the break time is 2 ms.
>
>I hit the tube with barely enough drive power to take it above the trip
>threshold. Sure enough, about 3 ms later the tube was safely into
>cutoff.
>
>The important thing is that the delay of an electronic trip can be only
>a few milliseconds for *all* levels of overload. This is what an
>electronic trip can do, but a fuse can not!
>
// Autopsies inddicate that some electronic trip circuits are not be
able to prevent gold-sputtering in an 8874, 3cx800A7 or 8877.
Circumstantial evidence is that a 0.25A fast fuse makes an 8877 hard to
bruise.
The 8877 is rated to carry c. 440mADC grid current, assuming 80Vrms
cathode drive. The 8877's grid is hardly delicate. It iweighs c. 2oz
and is made from 108 segments of molybdneum bar stock. At its max-raing
of 250MHz and 4000v, in AØ mode, the grid carries >42Arms of RF
circulating current from the tank.
- R. L. Measures, 805.386.3734,AG6K, www.vcnet.com/measures.
end
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|