▼ Menu

How Valves Fail

W T CockingWireless World, November 30, 1934.
Extras ▼


The end of the useful life of a modern valve is not usually set by the breaking of its heater or filament, but more often by a deterioration in the performance of the receiver. In some cases the valve may become microphonic, in others hum may appear. Some of the commoner symptoms shown by valves of various types are discussed in this article.

The rapid growth in the number of valves employed in a modern receiver makes it increasingly difficult to locate a defective specimen. At one time the death of a valve was obvious, for its filament ceased to light, but the introduction of dull-emitter valves made matters more difficult. Although the valve no longer gave out enough light to read by, the glowing filament was easily discernible. Users found, however, that the valve became useless long before the filament burnt out, since after a period the filament ceased to emit electrons in sufficient quantity for the valve to function correctly.

It was at this stage that the use of the milliammeter for valve testing came to the fore, since it provided the only means of determining the state of a valve. The emission of the cathode or filament varies throughout the life of a valve, and the tendency is towards a continual fall. This is reflected by a gradually dropping anode current, so that if a note be kept of the anode current of a new valve it is easy to see when the emission is failing by measuring the anode current periodically. Providing that the voltages applied to the valve are kept constant, a drop in anode current can only be due to failing emission. The precise amount by which the anode current can be allowed to fall depends largely upon how great a deterioration in performance one is prepared to tolerate.

Anode Current V Performance

In cases where the highest quality of reproduction is required the anode current should not be allowed to fall by more than some 25 per cent. below its initial value, and this applies particularly to the output valves. It is hardly ever permissible to operate a valve the anode current of which has dropped to less than one-half its original value if one wishes a receiver to give any pretensions to a good performance.

When testing the valve anode currents it is important to make sure that the voltages are maintained at their correct value. If it be found, for instance, that all the valves in an AC set pass low anode current it is more probable that the rectifier is failing and causing low voltages throughout the set than that all the valves are defective.

At one time the life of a valve was set either by the breakage of its filament or by the loss of emission, and this is still true of most battery valves, and, indeed, of directly heated valves in general. With indirectly heated valves, however, it is often found that the end of the useful life is marked rather by the appearance of defects rather than by falling emission. Noisy or intermittent reception is a common occurrence when valves are failing; the precise nature of the symptoms, however, varies, not only according to the use to which a valve is put but with its make.

Thus, when an indirectly heated triode is used as a detector or LF amplifier it is probable that the end of its useful life will be marked by the appearance of hum if it be a Mullard valve. This hum is usually of quite a high pitch and is intermittent, continuously varying in frequency and intensity. With a Mazda valve, however, it is more probable that a very deep, steady hum will appear, and the valve may become very microphonic. A Cossor valve will often give warning of its approaching decease by becoming noisy and giving a gentle background of crackles to reception, which will often temporarily cease if the valve be gently tapped. With a Marconi or Osram valve, however, the symptoms are less definite, and where low emission does not set the limit to its usefulness the end of its life may be marked by noisy reception whenever the valve is subject to vibration. If the valve be tapped when it is in this state there will usually be a single crash from the loud speaker.

It should be understood that these remarks are based upon the valves sold over a year ago, so that they may apply only to such valves. Changes in construction are continually being made, and it is by no means improbable that the valves now supplied do not exhibit the symptoms just described. It is yet too early to say how this yearbs valves behave, simply because insufficient time has elapsed for them to have reached the end of their lives with normal use.

The Broken Heater

There are other valve failures, however, which are likely to be found at times with any valve. One of these is a broken heater in an indirectly heated valve, and this is not as easy to diagnose as one might suppose. With a broken heater it often happens that the two ends remain in contact, so that no defect is revealed when the heater is tested for continuity. When the set is switched on the heater warms up normally and the set functions. When the heater has reached its full temperature, however, it expands to such a degree that the broken ends part company. This interrupts the current, the heater cools, and reception gradually ceases. When the heater has cooled sufficiently the ends again make contact, and current again flows, so that reception recommences. The cycle repeats itself indefinitely, and the continual interruption of the programme makes the defect particularly irritating.

Insulation Breakdown

Broken-down heater cathode insulation is another defect which can be very puzzling, particularly if it be intermittent. If it occurs in an LF. valve it usually causes poor quality of reproduction, since it short-circuits the bias resistance. It is, however, a rare fault in such a valve, if only because the difference of potential between heater and cathode rarely exceeds about 6 Volts. This applies also to an HF valve in a set fitted with AVC, but when it does occur the symptoms are usually instability and motor-boating.

In cases where volume control is obtained by biasing the cathode positively by some 20 to 40 Volts, and this means in most sets fitted with variable-mu valves and a pre-detector volume control, a breakdown in the heater-cathode insulation will almost invariably render the volume control inoperative. Instability and severe distortion may also appear, and are highly probable in a sensitive receiver. This defect is often intermittent, which naturally increases the difficulties of diagnosis, but a voltmeter connected across the bias supply will usually speedily reveal the fault, for its reading will drop whenever it manifests itself. It should be noted that a test of the heater-cathode insulation with the heater cold is of no value whatever.

Poor heater-cathode insulation, as distinct from a complete breakdown, may cause modulation hum which does not respond to the usual remedies. This is more likely to occur with high-voltage Universal-type valves than with the ordinary AC specimens on account of the greater AC potential between heater and cathode.

The Output Stage

Output valves are usually directly heated, and it is quite rare for them to develop any defect until the emission falls. Occasionally, however, grid emission or some similar defect occurs, and the symptoms are so violent that it is impossible to overlook them. The writer has several times met with valves of the 2.5 Watts output class which caused the loud speaker to emit a sound which can only be described as a scream. At the same time the anode current rose to several times the maximum rating for the valve, and the voltage developed across the bias resistance fell to zero an apparently impossible condition which can only be accounted for by reverse grid current equal in value to the anode current! This effect is rather rare, and it must be two years since the writer has met with a case.

Before concluding, some mention should be made of modern multiple valves. It is not uncommon for one-half of a double QPP pentode or Class B valve to fail before the other. The result is that the receiver continues to function, but very serious distortion appears. It is well, therefore, to test the anode currents of such valves separately. The diode sections of duo-diode-triodes do not usually cause much trouble, but it is as well to remember that a severe overload may cause that portion of the cathode which feeds the diodes to lose its emission, with the result that the efficiency may fall and distortion appear, while AVC may not function correctly.

Use browser back button to return.