Mullard produced a five part series on valve testing. Part one to three is original factory testing and stores testing, part four is returns testing and the final part is testing at dealers premises. We have parts two to five.
In the first article of this series the standard production tests which are applied to every Mullard valve and tube before it leaves the factory were described, and it was explained that the object of these tests is to ensure that the properties and performance of each individual valve comply with the very high standard specified by the Technical Department, which is an organisation entirely independent of the Production Department. This article deals with Quality Control testing.
The Function of Quality Control
In order to understand the function of Quality Control it is necessary to remember that every valve which fails to pass the standard Production Tests is not only rejected but is ruthlessly destroyed. All the material, time and labour spent in making it is thus wasted, and it is obviously an economic necessity to reduce these losses to a minimum. This is one of the main objects of Quality Control.
Quality Control Methods
The method of quality control testing is based on the fact that in mass production tests upper and lower limits are set and, in order to pass the tests the valve must give readings between these limits. If a single reading falls outside the limit the valve is rejected. It is obvious that if the majority of valves are near the limit, the number which fall outside the limit and thus be rejected will be greater than if the majority of valves are near the centre of the limit.
Routine quality control is conducted in the following way. From each production unit a batch of five or ten valves is taken at regular intervals. These are valves which have passed the normal production tests and would, in the ordinary course, arrive on the dealerbs shelf. These valves are then given a series of comprehensive tests, including not only those corresponding to the routine factory tests but also additional tests according to the particular type of valve.
There is, however, one important difference between production testing and quality control. Whereas the production tests merely prove that the performance of the valve is within the specified limits, in the quality control test quantitative measurements are made and the readings are recorded.
By comparing the measurements taken on successive batches of valves, any trend towards the lower limit can at once be detected. The Production Department can then be informed so that they may take any steps necessary to correct the downward trend before it reaches a point at which excessive rejection occurs. These steps may range from a check on the adjustment of production equipment to the introduction of some small modification in construction or method of manufacture, or possibly special chemical or microscopic tests on components or materials.
Among the tests included under quality control are those for inter-electrode insulation, emission, vacuum, mutual conductance, amplification factor and internal resistance. For indirectly-heated valves there is a test on heating time, and for battery valves a series of tests to indicate the performance when, for example, the HT or LT batteries in the set have run down. Inter-electrode capacitances are measured in the case of all HF valves, and there may also be tests of equivalent noise resistance and input damping. In the case of voltage amplifying valves, measurements are made of the gain in a standard circuit; for output valves the output for a given percentage of total distortion is measured. Special microphony tests are taken when considered necessary. Frequency changers also undergo a special series of tests where such properties as conversion conductance and oscillator performance are measured.
A proportion of the valves also undergo life test. In order to save time, valves on life test are operated under over-run conditions so that a run of 600 hours is equivalent to approximately 1,000 hours normal use. At intervals during life test the valves are remeasured, readings being taken of anode current under specified conditions, of mutual conductance, emission and vacuum.
Utilising the Information Gained
The results of the Quality Control tests not only serve to assist the production department to keep up the quality and reduce rejects. The information gained is produced at frequent meetings between representatives of various sections of the organisation - production departments, laboratories, development departments and technical service departments. At these meetings the effect upon performance and quality of any minor modifications in construction or manufacturing methods are revealed; or the desirability of such modifications may be discussed; and the knowledge and experience of all concerned is directed to the one object - the maintenance of the highest possible standard of quality in Mullard valves.
In previous articles it has been shown that, in addition to the production tests which each Mullard valve undergoes before it leaves the factory, a proportion of all the valves is tested for quality control purposes. These tests generally cover 'static' operation of the valves with little reference to the types of circuit in which they may be used. Similar batches of each type of valve are also sent to the Valve Measurement and Application Laboratory. Here they are subject to 'dynamic' tests in order to gain information relating to their performance in actual circuits.
The Valve Measurement and Applications Laboratory. The technician in the foreground is using equipment for measuring Equivalent Noise Resistance. On his right, power output and distortion are being investigated.
For each type of application there is certain information which is essential to the best use of any particular type of valve. For instance the design of audio-frequency amplifiers involves a knowledge of the power output and distortion of the output valves for various levels of input signal and with different values of circuit components such as load resistor and grid resistor.
It is also desirable to know the voltage amplification characteristics and the hum level of the pre-amplifier valve. Similarly, in the design of broadcast receivers, a knowledge of the cross-modulation assists in the correct choice of frequency changer and IF amplifier valves and the design of the automatic gain control circuits. (Cross-modulation effects are noticed when the modulation of an interfering signal appears on the carrier of a desired station. The cross-modulation factor is the ratio of the modulation depth caused by the interfering signal to the original modulation depth; for normal broadcast receivers this factor may rise to 1 per cent without undue interference resulting. The data sheets for Mullard frequency changers and variable-mu pentodes usually contain a graph of the signal which causes 1 per cent cross-modulation plotted against the slope or conversion conductance.
Special Test Equipment
In general, a special piece of equipment has been constructed for each of these types of measurement in which standard frequency or voltage levels are required for comparison purposes and in which easily switchable values of circuit components and input voltages are applied to the valve being tested. This enables the valves to be compared more readily under identical operating conditions with less likelihood of error due to the use of different components, generators, etc., at different times. Each equipment is designed so that as far as possible the information needed can be read direct from meters calibrated in the units required without complicated calculations from simple current or voltage measurements.
In the Valve Measurement and Application Laboratory there are such pieces of equipment for the measurement of equivalent noise resistance, hum level, cross-modulation, conversion conductance and other frequency changer properties, and for measuring power output and distortion.
Valve Performance Tests
The results of these basic dynamic measurements can be used at several stages in the history of a particular valve type. During development and pre-production, samples are tested to ensure that the performance of the new valve satisfies the target requirements. Similar measurements on normal production valves provide information for publication in valve data sheets and application reports. When the valve has been in production for some time, checks are made on batches of valves to ensure that the quality is maintained. These tests are, of course, supplementary to the routine quality control tests described previously.
Performance in Specific Equipment
In addition to these general measurements for which special measuring apparatus is constructed, there are many aspects concerning the application of a particular type of valve which need to be investigated individually.
In such a case it is necessary to design and construct the special circuit so that the overall performance of the valve can be studied in detail. This type of investigation may concern the use of a range of valves in a complete piece of equipment such as a broadcast receiver, an amplifier or a hearing-aid. Alternatively, it may be limited to a particular type of operation. An example of this is the design of a frequency changer stage for television reception, this work involving the measurement of input damping and noise factor at various frequencies. Other examples are the investigation of problems concerning the dissipation of heat from valves in compact equipment and the considerations involved in the choice of suitable valves for high stage gain at low frequencies.
This article has dealt with only that section of the Valve Measurement and Application Laboratory which investigates receiving valves. Other sections deal in a similar manner with cathode ray tubes; gas-filled devices such as thyratrons, voltage stabilisers and cold-cathode tubes; transmitting valves including UHF devices such as magnetrons and klystrons.
Information for the Valve user
Reference has already been made to the ultimate destination of the information obtained in this laboratory. In general, the results of all the measurements performed are passed to the valve user in one form or another. In most cases, the user is the circuit designer employed by an equipment manufacturer using Mullard valves. For him the information is contained in application reports or technical articles. At the same time, the data on specific valves is published together with other information supplied by the Technical Department in the Mullard Technical Handbook. Extracts from many of these reports and data sheets are published in other forms such as wall-charts, pocket-books, filmstrips, the Mullard Outlook and in advertisements. In this way the information reaches the dealers, service engineers, educational establishments and amateur constructors.
(A description of some of the equipment used in the Valve Measurement and Application Laboratory was published in the Journal of the British Institution of Radio Engineers, Vol. 12, No. 1, January 1952, pp. 63-68, 'Dynamic Measurements on Receiving Valves' by A I Heins.)
Dealers will already be familiar with the procedure for returning to the manufacturer valves which have proved unsatisfactory during their guaranteed life. This article takes the reader 'behind the scenes' and describes the tests made in the Service Department on returned valves.
For the sake of simplicity it may be useful to follow the fortunes of an imaginary valve returned in this way.
The valve, having been carefully unpacked and listed, arrives at the testing bench attached to a numbered report form on which the results of the tests will be recorded. At this stage the report shows the valve type and manufacturing code marks, the state of the parcel in which it was received, and the condition of the valve judged from visual inspection.
Section of the valve testing equipment at the Valve Service Department, Waddon.
In the Service Department a series of static tests is made with the aid of a permanently mounted instrument panel. The usual test board includes multi-range micro-ammeters, milliammeters and volt-meters, and by means of plugs and switches the instruments can be connected in the circuits as required. Controls are provided for the operator to set the electrode voltages at the specified values and the current readings are taken one by one for comparison with the values given in the factory specification. The controls have to be re-set for each individual type of valve. Should the readings fall outside the acceptable limits, the exact figure is entered in the report form. This method of working calls for a high level of skill in the operators.
Returned valves are normally tested to standards similar to those set for the tests which were conducted during manufacture.
The static tests are made in a definite order. Failure to meet the conditions at any one stage is recorded and no further tests made. While all triodes and pentodes are tested in the way about to be described, there are some slight differences in detail in the case of such valves as octodes and triode-hexodes.
The first static test is designed to check that the filament or heater circuit is continuous. The filament pins are connected to the rated supply voltage for the type of valve, and the current read from a milliammeter in the circuit.
The next test, of insulation resistance, is made with the cathode at a positive potential with respect to all other electrodes. Measurements are made of the leakage current from the cathode to various groupings of the other electrodes and from these readings any low insulation can be traced to the particular electrode responsible. During these tests the valve is tapped with a rubber-covered hammer to reveal intermittent faults.
A vacuum test is conducted with the valve operating under typical working conditions. If the valve has become soft, reverse grid current will flow. A sensitive micro-ammeter in series with the grid measures this reverse grid current, and thus indicates the condition of the vacuum.
The valve is also tested for cathode emission, for which it is connected as a diode, all the electrodes apart from the cathode being joined together as a composite anode. A standard AC supply is applied to this 'anode' and the total rectified current compared with the expected value. The last of the static tests is a check of at least three points on the anode current-grid volts characteristic curve. With the anode operated at its normal working voltage, the anode current is measured for several distinct values of the control grid voltage.
If the valve has passed all these tests, it is operated for a short time in a suitable receiver. This provides a direct practical check on the behaviour of the valve when in an actual circuit and shows to what extent it may be responsible for noise and microphony.
In appropriate cases, valves are dismantled and examined to ascertain the cause of failure.
The routine described consists only of the basic tests applied to the returned valve. A more elaborate procedure can be used to find the effect of peculiar operating conditions upon the valve when the dealer has mentioned such conditions on his BVA form. Any such information which the dealer can give is always helpful.
The results of the tests are passed to a separate department who decide, on this evidence, whether the valve can be replaced under the terms of the guarantee.
Using the Test Reports
While the proportion of valves returned from the field is extremely small, the work of control and development continues to reduce their number. A summary is made of the results of all service test reports, which is carefully studied at all levels to find whether any particular type of valve is developing a special weakness. This information is passed to the factory as a control on the quality of production. If the fault is not a routine one, sample valves are broken open in the laboratory for microscopic or chemical examination. The improvements in design which are brought about in this way continue to focus interest on valves returned 'from the field'.
The Mullard High-Speed Valve Tester in use at a typical dealer's service department.
Requirements for Service Testing
It will be appreciated that there are fundamental differences between the requirements for acceptance tests on newly manufactured valves and for tests intended to ascertain if a valve that has been in service has in any respect deteriorated to a condition which makes it unsuitable for further use.
The designer of equipment utilising valves has to take the necessary precautions to ensure that the equipment will continue to operate satisfactorily for a reasonable period before valve replacement becomes necessary, and he therefore allows for considerably greater tolerances in valve characteristics than those imposed by the valve manufacturer in his specifications for new valves.
Dealer Test Equipment
Dealer valve test equipment should therefore be designed to allow the engineer to make a check as to whether each characteristic of the valve remains within these wider tolerances, the lower limits of which are known as 'service limits'. On the Mullard High-Speed Tester the service limit is indicated by the top of the red scale.
In order to ensure accurate results in these tests, it is important that the valves are tested under the conditions laid down by the factory, and the apparatus must be so designed that these conditions can be applied with considerable accuracy.
The actual tests which are necessary are similar to those which are applied in the Mullard Valve Service Department and described in our last issue, and owners of the Mullard High-Speed Tester will have noted the similarity of the range of tests made.
Purpose of Valve Tester
The purpose of the valve tester is to enable the service engineer to check that a valve remains within its service limits for each test as enumerated on the test equipment. From these results he is then able to judge whether it is advisable to replace the valve. He can also judge with a reasonable degree of accuracy whether a particular defect in a receiver being serviced is due to the valve or to components in its associated circuit.
The Time Factor
In the early days of servicing, dealer valve test equipment was apt to be complicated in operation and to take up a considerable amount of time in the carrying out of the required tests. As a consequence, in many cases the dealer dispensed with the valve test equipment and used a trial and error method whereby a suspected valve was replaced by a new one. This method of test by substitution proved dangerous, as the use of a new valve might often mask an inherent defect in the receiver under test, with an early recurrence of the trouble and in some cases damage to the new valve.
There must always remain a few cases in which the engineerbs judgement is the final arbiter as to whether a valve should or should not be replaced. For instance, it might reasonably be anticipated that the RF valves in a television set used in a fringe area might require earlier replacement than if the same receiver were used in a situation where much less RF gain was required.
The complete testing of a valve in a dealerbs service department, by modern equipment such as the Mullard High-Speed Tester, enables the engineer to make all the necessary tests in very little more time than that taken by the valve cathode to heat up, and the apparatus has been so simplified that the actual testing can be delegated to a junior member of the staff.