Engineering Research Paper: Capacitors

Production testing of capacitors

Author: D. S. GIRLING, C.Eng., F.I.E.E., F.I.E.R.E.

Introduction

The progress of automatic testing of capacitors is explained and an indication given of future trends.

The most promising method of testing is to make contact to a large number simultaneously and to carry out a sequence of tests by automatic switching under computer control. The best place for this testing is while the components are orientated in manufacturing jigs. At the present time the speed of testing is limited by that of the transfer system

Tests Required

Tests on capacitors are required in three distinct areas:

(i) Production sorting

The production of capacitors, as with most components,is associated with a 'yield'. This normally varies between 80 to 95% overall and is basically due to the fact that the spreads of the parameters exceed those that are acceptable to the customer. It is, therefore, normal practice to carry out 100% testing using  automatic equipment by production operators.

(ii) Acceptance sampling

This consists of the formation of production lots and the taking of samples to determine acceptability. This sampling  is carried out in accordance with BS 9000. Since the manufacturer is required to allow for experimental error in his testing, it pays to use the most accurate methods available.

(iii) Quality assurance

A further requirement of BS 9000 is for lot-by-lot tests  periodic tests  every three months, and  every 3 years. These tests are of environmental nature and are destructive. All of these tests require the results to be recorded both initially and finally. The cost of this testing is an important factor in determining the acceptability of the BS 9000 scheme.

The main measurements or tests required on capacitors are as follows:

(i) voltage proof

(ii) capacitance

(iii) tan 5

(iv) leakage current or insulation resistance (see

Appendix)

(v) impedance (l.f. or h.f.)

  

Testing Methods

·         General Principles

Measurements may be made either on an attributes or a variables basis. The first of these is sometimes referred to as a go/no-go test, although this term strictly applies to dimensional gauging. In the field of variables measurements the problem is more complex. Let us consider the measurement on a manual bridge. This consists of the following operations:

(i) load,

(ii) balance,

(iii) read result,

(iv) decide if in limits (some calculation may be required)

(v) write down (if necessary)

(vi) unload.

·         Manual Equipment

This needs little or no explanation, but it is surprising how much of this equipment is still in use and even sold commercially.

Generally speaking, it is any equipment which requires manual adjustment of controls, visual reading of results, and manual recording if necessary.

·         Comparators

These are widely used for production operations, particularly where some adjustment is required. It consists of a standard of the value required and some means of indicating an out-of-balance between that and the work under adjustment. A typical accuracy is about 1 % but it is possible to expand the scale to give greater accuracy over a limited range provided that the zero can be maintained.

·         Automatic Test Set

One of the most well known of these consists of a wheel into which the capacitors are loaded in contact jaws. They are then carried round so that contact is made in turn to each of a number of test equipments. The results of each test is recorded in a memory so that when the capacitors reach the eject position they are ejected into the appropriate drawer.

·         Measurement of Leakage Current or Insulation Resistance

One of the simplest, and yet the most time consuming, is the measurement of leakage current. This is the conduction current flowing through the capacitor with a direct (either the rated or an arbitrary fixed value)

voltage applied. A schematic diagram of the method of measuring leakage current is shown in Fig.

Fig. Measurement of Leakage Current or Insulation Resistance

The resistance of the circuit combined with the capacitance of the capacitor under test result in a time-constant. When the voltage is applied to the capacitor it is necessary to wait for the charging current to decay before the leakage current can be measured.

Conclusions

(i) Investment in automatic test equipment for large scale component testing can be justified if advantage

is taken of their high inherent speeds.

 (ii) Automatic test methods permit the high accuracy testing of capacitors at speeds up to 3000 per hour for electrolytics and about 10000 per hour for non-electrolytics. The difference is mainly due to the different measuring frequency.

(iii) The main problem is the electrification time before the measurement of leakage current. By the use of

low impedance charging circuits early decisions may be made as follows:

(a) The use of reduced time periods and corrected limits to take accept/reject decisions.

(b) To accept as soon as the limit value has been passed. In all cases it is necessary to charge a number of

capacitors in parallel.

.