[Amps] AMPS

[Ian White GM3SEK]

Sorry, my previous posting was an intermediate draft. What I meant to 
post was this version, so please ignore the previous one.

There is nothing technically different below, but I tried to make the 
explanations even clearer.



R L Measures wrote:
>
>> The manual includes an extensive, well written
>> discussion of tube protection, how it works and how to configure
>> it.  However, since it does not agree with your pseudo science you
>> don't acknowledge it - as usual.
>
>The Q1 is a TIP-147 and is rated at 5A.  Q1 appears to carry grid 
>current. Is this true?

(For reference, we're discussing the Triode Board user manual:
http://www.ifwtech.co.uk/g3sek/boards/triode/triode-manual.pdf )


Q1  is part of an 'electronic zener' circuit that provides cathode bias. 
It has the same function as a large fixed zener diode.  It is not part 
of the protection circuit, and it is NOT required to interrupt large 
grid currents.

In normal operation Q1 carries cathode current, and also normal levels 
of grid current. The current paths for normal anode and grid current are 
shown in Figure 1 on page 6.

As several people are pointing out, there are two completely different 
fault scenarios to be considered. A "soft" fault involves moderate 
levels of excess grid current - a few hundred milliamps. A "hard" fault 
involves possibly tens of amperes, coming from B+.

A "soft" fault involves up to a few hundred milliamps of grid current, 
and might typically  caused by incorrect drive/tuning/loading. Fuses 
cannot protect against such faults, because the current level is too 
low, but the tube will be damaged if the faults are allowed to persist.

The only way to handle these "soft" faults is by electronic protection, 
which will quickly take the amplifier off-line while also allowing a 
quick and easy recovery. The grid current trip is normally set at about 
150-200% of the maximum normal grid current, and operates in a few 
milliseconds so the grid is very well protected.

In this kind of fault, the bias device (transistor Q1 on my board, or a 
large fixed zener) only has to handle a few hundred milliamps of excess 
grid current for a few milliseconds,  which of course is no threat.


The scenario for "hard" faults is COMPLETELY DIFFERENT. The driving 
force for these faults is always B+, and the current takes a different 
path. This path is shown in Figure 2 (page 7) and it does not involve 
the cathode or Q1. A "hard" fault will involve both high grid current 
and high anode current, and the Triode Board senses both of these.

Once again, fast electronic protection takes the amplifier off-line and 
a relay breaks the mains supply to the HV transformer. While this is 
happening, the glitch resistor in the B+ line limits the maximum fault 
current.  Large diodes protect the meters and grid current sensing 
circuits, and steer the fault current safely back to B-minus. Such 
faults are normally completely recoverable with no component damage.

The effects of "hard" faults involving internal tube arcs are difficult 
to test for, because such arcs cannot be produced on demand. However, I 
have done extensive tests with external crowbar shorts caused by 
dropping a piece of copper tubing into the plate compartment. Typically 
the amplifier shuts down with a modes "tick" and doesn't even blow a 
fast mains fuse. This is the kind of protection I was aiming for.

I repeat: Q1 plays no role in this. It is simply a bias component.  It 
is not part of the protection circuit, and it is NOT required to 
interrupt large currents.

BUT

Everything I said above about current paths assumes that the grid stays 
at ground potential.  If the control grid is allowed to float, two very 
bad things happen (in any triode PA, and with anybody's control 
circuit):

1. The tube has no functioning control grid, so it acts like a diode 
with no current limiting capability (except the emission capability of 
the cathode).

2.  The current path switches to go directly from anode to cathode. In 
that case all the fault current goes through the bias device (zener or 
power transistor) and you'll probably lose it.

So don't ever aim to disconnect the grid deliberately in any fault 
situation.




In conclusion, this isn't a simple topic that can be debated by a 
barrage of one-liners. The subject does require some considered thought.

As this will be my last posting to AMPS for several days. I hope I've 
made my points clearly enough.


(Once again, apologies for having posted the previous version before it 
was ready.)


-- 
73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek