Valve Curve Understanding the Significance

oj Load and 0 ther Lines

Diagrams By "CATHODE RAY"

L AST month, in discussing cathode followers, I made

use of certain valve curve diagrams. It has occurred
ness men understand this, when the two quantities are
such things as time and commission on sales. But
to me that there may have been readers who quickly while we may all understand how it applies to voltage
shied off at that stage, or, seeing the diagrams in and current (for example, anode current and grid
advance, were non-starters. Others, though less voltage), what may not be quite so clear is how resis­
easily deterred, may through unfamiliarity have found tance, conductance and power can also be represented
them somewhat baffling, notwithstanding the clues I on the same diagram, or how several different voltages
scattered as freely as space permitted. in a circuit can be shown.
The first thing that has to be explained, perhaps, If one were to repeat Ohm's original experiment,
is why it is considered necessary to use up a lot of plotting the current passing through a piece of wire,
paper and drawing effort in this way instead of deal­ against the voltage between its ends, the resulting
ing with valve problems in a neat equation or two. graph would be the kind of thing shown as Fig. I-a
The reason is that valves do not behave in ways that straight line passing through the "origin" (0). (Of
can be represented accurately by neat equations. They course Ohm himself knew nothing about volts and
are not like resistors and capacitors and air-core amps, but we might as well make use of our modern
inductors. Oh, I know there is such a thing as an units.) The information conveyed by this line could
"equivalent generator" by which certain valve calcu­ be presented with much less effort as an equation:
lations can be reduced to simple algebra, but (a) that V = 31. Except for the number, the equation \yould
method takes account only of signal currents and be the same for different pieces of wire; a shorter
voltages, so is no use at all for finding the best working length of the same wire would give a smaller number
conditions, such as grid bias voltage, and (b) it doesn't than 3, and vice versa. If " V " is being used to denote
even deal with the signal part accurately, because it the potential difference in volts, and "I" the current
ignores the curvature or non-linearity of valves. In in amps, the number is the resistance in ohms. The
any case, certain types of mind are more brightly smaller the resistance, the steeper the line in the graph.
illuminated by a graphical diagram than by a row of If that fact is not obvious, try one or two different
equations. lines, and consider why the slope of the line is con­
In equations, quantities such as voltage and current nected with the resistance in this way. The reason, of
are represented by letters or numbers (depending on course, is that resistance in ohms can also be regarded
whether their values are being dealt with in general or as volts per amp. So the resistance represented by a
particular). On diagrams they are represented by dis­ line on a current/voltage graph is equal to the number
tances on the paper. I am assuming it is well known of volts it slopes along the voltage scale for each amp
how two such quantities are represented by distances up the current scale. In other words, resistance is the
respectively horizontal and vertical. Even tired busi- ratio of voltage to current, and on a graph the slope or
gradient of a line is the ratio of vertical movement to
I 2 horizontal movement or in this case current to voltage.
The easiest figures for finding the resistance in this

o
V'I example are 3 volts and I amp, but because the line is
Q.
:E straight-representing a linear resistance-the differ­
<
ences in volts and amps between any two points on the
line would do. If the resistance were not linear, the
slope of the line, and the resistance, would vary with
current (or voltage).
So not onlv the value of a resistance but also
whether or not it is linear, is clearly shown on a
-� -4 current/voltage graph.
And because conductance is the ratio of current to
voltage, it is shown too; the steeper the slope the
greater the conductance. The mutual conductance of
valves is, in fact, often called slope.

Representing Power

How about power? It is current multiplied by

-J' 2 voltage. Horizontal distance multiplied by vertical
distance gives the area enclosed by the vertical and
Fig. I. Graph of current against voltage for linear resist­
ance, represented by the diagonal line. The power used up horizontal lines at each end. For example, the power
in it when I amp is flowing is represented by the shaded area. released in our wire when I amp flows through it (Le.,
Negative currents and voltages are in the reverse direction 3 watts) is represented by the shaded area. With a
in the circuit. shorter piece of wire, only I volt might be needed to

338 WIRELESS WORLD, JULY 1955

pass 1 amp"and the corresponding area would be one­ The point p, where the two Iines cron, is the only one
third the size, rep re senting 1 x 1 1 watt. Equal
= common to both, and indicates that the current
powers in different resistances are represented by equal through both must be 0.615A, the voltage across the
areas of different shapes. 30 must be 1.85, and across the 10il, 6.15. Having
Incidentally, if the voltage in Fig. 1 were doubled, checked this by calculation, we can have some faith
from 3 to 6, the area would obviously be four times as in the graphical method and go on to apply it to
big. The diagram helps the weaker brethren to situations where calculation fails.
visualize the fact that (with a linear resistance) the But before we do that, let us see how Fig. 2 can,be
power dissipated is proportional to the square of the used to answer different kinds of questions. If we
voltage (or current). knew the value of the current but not R2, it could tell
Our Fig. 1 line represents a certain resistance or con­ us what R2 would have to be. Try it for Rl = 3 and
ductance, but does not by itself reveal the actual 1=0.5. In this case the point on the Rl line is fixed
current flowing in it. That depends on the voltage, by the fact that 1=0.5, so what we have to do is lower
which we do not know. It might be anything. What the slope of the R2 line to make it pass through that
the line does show is that if 3 volts were applied the point and then find what resistance it represents.
current would be 1 amp. Suppose we don't know the Or suppose we are told to find the value of R2 that
voltage applied to this 3-ohm resistance, but we do results in 2 watts being dissipated in RI" That means
know the total voltage applied to it and another known
resistance in series. With linear resistances it is a
simple exercise in Ohm's law to calculate the voltage
across each resistance and the current through both.
With non-linear resistances, to which Ohm's law
doesn't apply, we would probably be stuck-if we
didn't have the graphical method to fall back on. But
o's
before taking a non-linear example, let us first try a
linear one, which we can check by calculation.

Two Resistances

Suppose 8V is applied to our 30 in series with 100.

We know that the resulting state of affairs must be ----
represented by a point somewhere on the resistance
line in Fig. l. It must also simultaneously be on a
line representing the 100. If we were to draw a 100
+ 8
line through 0, that would be, the only point common
VOLTS
to both lines, and of course it would not represent
the situation at all. The clue is the fact that the volt­ Fig. 3. The top right-hand corners of all the rectangles
age applied to the 30 is 8V minus whatever is dropped representing a given power (2 watts in this case) trace
in the 100. The voltage dropped in the 100 is, then, out a curve like this.
from the point of view of the 30, a negative one,
beginning at 8V. So we draw the 100 line as shown
dotted in Fig. 2. To emphasize that there is nothing 2·0
wrong about putting the zero-current point at 8V, I
have added a second voltage scale to apply to this
1'8
resistance. The dotted line shows on this scale the
voltage to be deducted from 8V to give the voltage
across the rest of the circuit, whatever the current. 1"6

1·4

1·2
R,

II)
II)
0.. 1"0
0..
� ..
�

..
0'8 \
\-1'
\..
0'6 \
\
\
0·4- \
\
\
\
l 4- 5 8 0·2 \
\
VOLTS ACROSS R, \
, 6 5 4- l 2 o 0
8 7 4- 0
VOLTS ACROSS R�
VOLTS
Fig. 2. A circuit with two resistances in series can be investi­
gated by adding a second resistance line, sloping from the Fig. 4. Power curves can be used to find the value or R2
point representing the total voltage. receiving maximum power in the series circuit.

WIRELESS WORLD, JULY 1955 339

 �o

[§
40
OVER HEATING AREA

,-.. 30
«
1=
-... .;
H
.. 20

2
10

v... VOLTS

Fig. 6. Typical set of anode-current/anode-voltage curves

-2 -I
for a small power triode. showing the areas that for various
VOLTS reasons are out of bounds.

Fig. 5. Graph of a non-linear resistance-that of a diode

infinity, but at higher voltages than shown here it is
valve.
practically linear and therefore constant.
This is where the new boy may get confused. The
drawing a constant-power line. A power of 2W can ordinary "d.c." way of reckoning resistance is the
be made up of 2V, lA, or 4V, 0.5A, or 5V, O.4A, or ratio of applied voltage to current flowing. At point
8V, 0.25A, and any number of such combinations. a, the voltage is 2.8 and the current 3mA, so the resist­
The 2-W line can be obtained by plotting a few of ance is 2.8/0.003=930D. This resistance is equal to
them and drawing the smoothest curve through the that represented by a straight line joining a to O. It
points, as in Fig. 3. This fixes a point on R" through is not equal to the resistance represented by the slope
which the R2 line can be drawn to the applied voltage of the valve curve at a. This slope resistance is some­
mark on the voltage scale as before, and the value of times called the a.c. resistance, being the resistance to
R2 follows. Alternatively, if R2 is known, a line of small alternating currents superimposed on the steady
the corresponding slope is drawn through the Rl-P 3mA at a. The reason they are supposed to be small
intersectIon, and where it crosses the 1=0 axis it is that the bit of curve involved by them should then
indicates the total voltage that has to be used. be as near straight as makes no matter. Both these
A rather more difficult problem would be: Given kinds of resistance are significant; the d.c. kind when
Rl and the total voltage, find the value of R2 in which considering the "working point" of a valve (anode
maximum power is developed. One way of doing this voltage, bias, and so forth), and the a.c. kind when
is to draw several different power curves for R2. This considering signals being handled by it. At a there
means that they have to be drawn with reference to is not a great deal of difference between them, but at
the"volts across R2" scale, as in Fig. 4. The point b the d.c. resistance is zero, whereas the a.c. resistance
on the R, line corresponding to the highest power is is far greater than at a.
q, somewhere between 5 and 6 watts (actually 5t), A diode is normally used as a rectifier, and rectifiers
and if the diagram has been drawn well enough it will are always more difficult than you think, so despite the
tell us that R? for this condition is 3D. As we prob­ apparent simplicity of the diode I am going to hurry
ably knew alf the time, it would invariably be equal past it to the triode. The anode current in a triode
to R whatever that was, because a well-known and depends simultaneously on two voltages-anode and
impd:tant circuit theorem says so (the Maximum grid-so really needs a three-dimensional diagram,
Power or Load Matching theorem). for the making of which one would have to employ a
sculptor, and the Editor would object to the expense.
Diode Characteristic So, although a triode's current/voltage characteristic
is really a 3D su'rface, for economy and convenience
I should think that's about enough for linear resist­ it is usual to make do with a series of cross-sections
ances, for all the problems so far (except possibly the of this surface in two dimensions. Which two depends
last) can be solved more easily and neatly without on what one wants to show most clearly. Sometimes
graphs. A diode valve is a simple example of non­ they are anode current (Ia) and grid voltage (Vg), at
linear resistance. As Fig. 5 shows, regarded as a a number of evenly-spaced values of anode voltage
resistor it has several features not according to Ohm. (Va); and sometimes Ia and Va at values of Vs. The
First, a negative voltage does not cause a negative latter (Fig. 6) are the more generally useful.
current; i.e., one in the opposite direction to that
which flows with a positive voltage. (This is not strictly Forbidden Areas
true, but one has to have a very super-sensitive micro­
ammeter to discover it.) On the contrary, the current The shape of the la/Va curves is very like the diode
when the negative voltage is small is positive. Next, one. The effect of making the grid negative is, roughly,
the slope of the line (which is visually, as well as to push the curve bodily along to the right. What the
mathematically, a curve) increases as the voltage in­ effect of making the grid positive is, one does not
creases positively, which means that the resistance usually bother to find out for ordinary receiving valves,
decreases. Near zero it decreases very rapidly from because grid current flows and greatly complicates the

340 WIRELESS WORLD, JULY 1955

 situation, as well as spoiling the valve for most of its us say) is 250V. And if the load were to be a resist­
uses. So the whole of the area to the left of the ance, fed from the maximum supply voltage (400) it
"Vg=O" curve is reckoned as out of bounds. In fact, would be represented by the sloping line through 0
as Fig. 5 shows (for the grid and cathode of a triode and 400V OmA, as in Fig. 7. From its slope we find
together equal a diode) the forbidden area may have it is 6,2500. We note that the working point is on
to extend to Vg= - I V, or even a little farther, to the"Vg= -15 " curve, so that is the grid bias. And
make sure that no appreciable grid current flows. if we allow the signal input to swing the grid right
Next, again assuming that distortionless amplifica­ up to 0 and down to -30, the load line shows that the
tion is wanted, it is advisable to fence off the sharply corresponding V. swing is between 140 and 350 (=
curved part at the foot of the diagram, marked 210 peak-to-peak) and I. is 41.6 and 8 ( = 33.6 peak-to­
"Bottom Bend Area." The remaining parts of the peak). The voltage amplification is therefore 210/30
curves are not dead straight, but are tolerably so, and =7. The power output (into the resistance) is equal
can be made much straighter by negative feedback, to the r.m.s. signal voltage multiplied by the r.m.s.
as we saw last month. signal current, and since an r. m.s. value is 1/';2 times
Th(.! ceiling is imposed by the valve makers'"maxi­ a peak value, which in turn is half the peak-to-peak
mum anode dissipation "-the maximum power" value, this power is equal to peak-to-peak Va X la,
Va X la, that it is safe to inflict on the anode. Suppose divided twice by 2'; 2, that is to say by 8. So it is
in this case it is 6 watts. Then we draw a 6W curve (210 X 0.0336)/8=0.88W.
on the diagram as shown, to rule off what can be called
the Overheating Area.
Lastly, the valve maker usually specifies a maximum Voltage Amplification Line
anode supply voltage (V'(b) max). This must not be
From a practical point of view all this is rather
confused with the maximum anode working voltage
absurd. Is it voltage amplification or power output
(V.,m,,) which is the voltage between, anode and
we are trying to get? We have adopted a usual method
cathode when no signal is coming through, or the
for voltage amplification-a resistance coupling-but
average when it is. When there is a resistance coup­
the valve is clearly unsuitable for this and is intended
ling, this anode voltage is less than the supply voltage
for power amplification. However, what we are really
-by the amount dropped in the resistance. But it is
out for just now is a quick understanding of graphical
a voltage that is liable to get at the anode occasionally,
technique for valves, and I hope I haven't confused
at signal peaks or while the cathode is heating up. A
you by explaining two things at once. The procedure
vertical line should be drawn at this voltage (say 400
just described, if applied to a suitable high-mu valve,
for example) to close up the remaining gap in the
is correct for voltage amplification. One would not
boundary.
actually bother about a maximum power curve, how­
ever; the aim would be to slope the line as little as
Power into the Load possible, even perhaps into the' bottom-bend area, so
long as the resistance was not'so high as to be shunted
We now have a clearly defined area in which to play.
too much by stray capacitance at the top signal fre­
But we should remember that there may be a section
quency. The working point would be fixed where it
of it on the right that is only allowed for transient
gave equal positive and negative grid swings within
occupation-not for lingering in. That is, if there is
reasonable limits of distortion.
a Va max lower than the Va(b) max- On the other For a power amplifier, on the other hand, one wants
;
hand momentary trespassing across the "overheat­ to get the power out into some external load, such as
ing" boundary is permitted, so long as the working a loudspeaker, not waste it all in a resistance coupling.
point itself is not outside. The coupling is done by a transformer, which has very
If we were aiming at the maximum power output little-perhaps negligible-d.c. resistance, but con­
from this valve we would put the working point
siderable signal-frequency resistance. The usual pro­
actually on the 6W boundary at Va,max, which (let cedure would be to place the working point as already
done in Fig. 7, and then draw from it to the voltage
scale a line representing the d.c. resistance of the trans­
former or choke coupling. Being such a low resistance,
the line would be almost vertical, and the resulting

-IT .(b) 1
Va(b) indicated by where it cut the Va scale would be
--- V
•

only slightly more than the working Va.

The a.c. load line need not touch the Va scale at any
V
40�--+---�--�---t--� . particular point such as V.(b) max; it should be
swung round 0 as a pivot until it indicates the maxi­
mum output. The output power is represented by
� 30�--+---�r,r-r+-��� �____���-J
one-eighth of the area of the rectangle of which the
....
�
... load line is a diagonal. If the load line slopes too little,
..:! 20 this rectangle is too flat to have much area; if the line
slopes too steeply the rectangle is too narrow. The
length of the load line diagonal must be equal in both
directions from its pivot at 0, and must not go beyond
the grid-current or bottom-bend boundaries. The
o 6,2500 line in Fig. 7 is unlikely to give the largest
V". VOLTS area because an input signal limited at its positive
peak by grid current leaves quite a lot of useful space
between its negative peak and the bend boundary. A
Fig. 7. The Fig. 6 curve sheet with load line added. more promising line would be steeper, indicating a
through the working point (encircfed). lower load resistance; drawn, in fact, from the point

WIRELESS WORLD, JULY 1955 341

where "Vg= -30" cuts the bottom-bend boundary. negative feedback. Then, of course, there are pen ­
In practical design there is vastly more to it than todes. Their curves have quite different shapes, but
this; all I have been attempting to do is show what except in detail the methods are the same.
the various lines and things on this kind of diagram At least one whole book* has been written on the
mean, and how it is that they mean them. If I have subject, and the uses included in the Radio Designer's
succeeded in making this clear, then perhaps you Handbook would almost make another book. So there
would like to turn back to last month's treatise and is plenty of scope for follow-up.
note how the ordinary valve curves can be used to
derive another set of much straighter curves that * Graphical Constructions JOT Vacuum Tube Circuits, by A.
represent the behaviour of a valve combined with Preisman. (McGraw Hill.)

Manufacturers' Products: NEW EQUIPMENT AND ACCESSORIES

In Type FRM, which is 4in long, a single wave-wound

Ground-to-Air Transmitter coil at one end, of 165 ILH inductance, covers 180-550
A NEW v.h.f. transmitter for ground-to-air communica­ metres when tuned by a 500-pF variable condenser. The
tions, rated at 20 W output, has recently been introduced Q at 1 Mc/s is stated to be 205.
by Ekco Electronics to replace an earlier model. The Type FRD has an additional winding at the other end
new set, Type CE91, can be operated on any crystal­ of the rod giving a combined inductance of 2.2 mH to
controlled spot frequency in the band 100 to 1S6 Mc/s, cover wavelengths up to 2000 metres. The length of this
channel changing being effected by fitting the appropriate rod is 8in.·
crystal and realigning the circuits. All the controls are Rubber grommets are provided for mounting, and a
readily accessible from the front panel but protected fibre disc, secured to each coil former, facilitates adjust­
against accidental misalignment by easily removable cove: ment when moving the coil on the "Ferroxcube" rod
plates. core.
The price of Type FRM is 8s 9d and of Type FRD
12s 9d.

Commercial Literature
. Marine V.H.F. Radiotelephones, a range of six models
glVlng 10 watts output and covering 40-1S5 Mc/s with 10
or 20 channels. Available for a.m., f.m. or combined
a.m./f.m. Brochure from Redifon, Broomhill Road, London,
S.W.1S.
Soldering Irons by Hydrel of Switzerland with pointed or
hammer-shaped copper bits claimed to withstand oxidation.
Elements from 45 to 500 watts, lengths 12!in to 17tin, weights
7 oz to 2t lb. Leaflet from the sole distributors, A. B. Hobbs
& Co., 214, Hatfield Road, St. Albans, Herts.
Overtone Quartz Crystals, 17 Mc/s to 36 Mc/s, listed in
New Ekco ground-to-air v.h.f. transmitter, Type CE91. a new easy-reference catalogue of Salford crystals from the
General Electric Co., Magnet House, Kingsway, London,
W.C.2. Also a booklet on selenium rectifiers, giving per­
Particular attention has been given to the suppression formance figures and curves for various circuits, and a leaflet
on Gecalloy micropowder permanent magnets.
of spurious emission, a matter of some importance now
that the 200-Mc/s band is likely to become a highly Waveguide Components and test instruments for centimetre
populated one before long. The inclusion of bandpass and millimetre waves, with notes on automatic measuring
instruments suitable for production testing. Illustrated
and lowpass filters in the .circuit contribute, no doubt, to
catal.ogue from Elliott Brothers (London), Century Works,
the "clean" performance claimed for this transmitter.
Lewlsham, London, S.E.B.
The set, including the power supply, weighs 75 lb and
High-voltage Control Valve, triode Type TV501. With
fits into the standard 19-in rack. It is made by Ekco
70 kV on the anode, the anode current (max. 1.5 A) can
Electronics, Ltd., Southend-on-Sea, Essex. be C �lt off to 100 JJ.": by �pplication of -400 V to the grid.
DetaIls and charactenstIcs III a brochure from Solus Electronic
Ferrite Rod Aerials Tubes, 15-1S, Clipstone Street, London, W.1.
Microwave. Frequency Meter, 2,400 .to 10,200 Mc/s, and
TWO directional rod aerials are now available from the other wavegUIde components and test lllstruments described
Teletron Co., Ltd., 266, Nightingale Lane, London, N.9. in an illustrated catalogue from the Narda Corporation' 66,
Main Street, Mineola, N.Y., U.S.A.
Selenium Rectifier Stacks for domestic sound and television
receivers. A booklet with information on ratings, coding and
polarity markings, dimensions, weights, and 45 pages of per­
formance curves. From Standard Telephones and Cables '
Rectifier Division, Edinburgh Way, Harlow, Essex.
Radio Control of Models. Ex-Government equipment for
this and other purposes listed in a new mail order catalogue
(No. 12) from Arthur Sallis Radio Control, 93, North Road'
Brighton, Sussex; price Is 6d including postage.
Mobile Television Units in motor vans for outside broad­
casting, with cameras, control equipment, centimetre-wave
tra���itters, . etc. . Diagrams and photograp�s showing
factlmes avatlable III a booklet from Marcom's Wireless
.. Te/etron .. Type FRD ferrite rod aerial. Telegraph Company, Marconi House, Chelmsford, Essex.

342 WIRELESS WORLD, JULY 1955