Re: Question about grid behavior with triodes

Hello RAT friends,

Hello Wessel

I've looked in the books I have but there is not really a
useable
explanation over the practical aspects over grid behavior,
particularly with the onset of positive grid current. A
few questions
for the knowledgable.

1 Is there a formula over what the load impedance is of a
triode
control grid is "normally" and when driven positive?
2 Does grid current only occur when the input signal peak
voltage
reaches bias voltage or are there other scenario's?


The factors governing grid current are the same as for other
electrodes. Check out "Child-Langmuir". The formulae are not
often used in design because they are too complicated.
That's why characteristic curves, rather than formulae, are
published in datasheets. If a valve is designed to tolerate
grid current, you might expect a decent data*** to include
some pertinent graphs, at least sufficient so you know how
not to exceed the limit.

As for "other scenarios", datasheets will normally give a
value for the maximum allowable grid leak resistor. This
prevents the usual small "negative" grid current from
significantly depressing the voltage of the grid. Grid leak
values range from around a Meg for little valves, to around
100k or less for common power valves, so you can figure that
grid current must be on a scale of microamps. If you keep
your grid leaks at or below the maximum, you should be able
to forget about grid current as long as you have a touch of
negative grid bias.

If you really need formulae, then it may be worth searching
for spice model documentation. The worst thing about Duncan
Munro being AWOL is that he has removed the excellent
documents explaining his generic triode model from his web
site "until further notice", since years ago.

However, if you like algebra, you may be able to glean the
gist from the model itself. Download the PSpice model from
Duncan's site here:

http://www.duncanamps.com/spicevalvesgt.html

1. Note that you can *ignore* the word "limit". (a
programming fudge to suppress negative values)
2. Ignore the second and third columns. (they refer to nodes
in a circuit comprising the valve and several voltage and
current sources)
3. The lines of spice code can then be viewed as a list of
parametric equations, with each variable parameter on the
left, and its formula on the right.
4. Currents and voltages in the formulae are labelled with I
and V, respectively. In the parameters, they are labelled G
and E. Sigh...
4. You need to follow back the parameters until you get a
big formula in terms of currents and voltages at the valve's
electrodes.
5. PWR(a, b) means "a to the power of b".

For example, looking at the section of the model devoted to
grid current:

Egf GF 0 VALUE {PWR(LIMIT(V(G,K)-{GCO},0,1E6),1.5)*{GCF}}
Gg G K VALUE {(V(GF)+V(LD))}

You can see that grid current Gg depends on V(gf) and V(ld).
To find values for these parameters, you must look for them
elsewhere in the list; V(GF) is given by the formula for
Egf, on the previous line. V(LD) is given by Eld, somewhere
else in the model. Etc. If you expand the equations to
eliminate unknowns, you end up with a long
Child-Langmuir-type equation with a few bodges added.

Constants in the equation reflect the essential
characteristics of the particular valve, such as its
perveance and transconductance.

Some constants, such as GCO and GCF are essentially fudge
factors to force the Child-Langmuir equations to fit the
wiggles of real valves.

Ian




.

Quantcast