GB2242750A

GB2242750A - Testing circuit boards for withstand voltage

Info

Publication number: GB2242750A
Application number: GB9104718A
Authority: GB
Inventors: Koji Okawa; Michihiko Yoshioka
Original assignee: Mitsubishi Cable Industries Ltd
Current assignee: Mitsubishi Cable Industries Ltd
Priority date: 1990-03-06
Filing date: 1991-03-06
Publication date: 1991-10-09
Anticipated expiration: 2011-03-06
Also published as: GB2242750B; GB9104718D0; JPH03255966A; JPH0614092B2

Abstract

A method of testing withstand voltage in circuit boards comprises measuring the partial discharge that occurs in the insulation layer of the board when a predetermined voltage is placed between a circuit conductor and the metal base of the board and evaluating the breakdown strength on the basis of the total charge that flows during partial discharge. A good board might pass less than one picocoulomb of charge.

Description

METHOD OF TESTING FOR WITHSTAND VOLTAGE IN PRODUCT INSPECTION OF CIRCUIT BOARDS FOR HYBRID INTEGRATED CIRCUITS This invention relates to a testing method for evaluating the withstand voltage of circuit boards for fabricating integrated circuits (hereinafter referred to as "hybrid ICs").

Common circuit boards for installing densely packed or high-power hybrid ICs comprise a metal base, such as an aluminum base, that serves as a heat sink and which is overlaid with an insulation layer made of an organic insulator such as an epoxy resin or polyimide or an insulation layer containing an organic insulator such as glassepoxy, which insulation layer is further overlaid with circuit conductors. In conventional routine product inspection, such circuit boards are tested for withstand voltage or dielectric breakdown strength almost exclusively by placing an AC voltage of 2 - 5 kV for 1 minute between a circuit conductor and the metal base of the circuit board (JIS C 2110). The dielectric breakdown strength of the product of interest is evaluated by checking whether dielectric breakdown occurs in the test.

With high voltage cables, power capacitors, transformers and other high-voltage electrical products which are required to have long-term reliability extending to 30 or 40 years or more, a method is known that measures the internal discharge or "partial discharge" of insulation layers that takes place on account of voids and other defects in the insulation layers. This method is effective for the purpose of evaluating the long-term reliability of the insulating performance of high-voltage electrical products and is currently used not only in evaluating the long-term reliability of their insulating performance in various situations such as research and development (R & D) activities and approval tests but also for diagnosing the degree of deterioration in the insulating performance of high-voltage electrical products in long-term service.

Taking, for example, power cables, particularly highvoltage ( < 60 kV) cable, partial discharge measurements are also conducted during product inspection prior to delivery and, in this case, too, the measurements are primarily intended to check for the long-term reliabIlity of the insulating performance of the cables under test and their dielectric breakdown strength is evaluated bv other testy inciuding an AC voltage breaRco*-n test and a DC impulse voltage bre ;;åor test.

Packaged devices, i.e., modules or boards having electronic capabilities, are produced by installing semiconductor devices (e.g., transistors), resistors, caacitors and other devices cn circuit boards comprising metal bases overlaid with insulation layers that consist c or contain organic insulators and which in turn are overlaid with circuit conductors. The circuit boards for use in producing such packaged devices have of course passed the withstand voltage test described above, and yet dielectric breakdown will very often occur in the insulation layer of these circuit boards during a subsequent withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture.Even if the conditions of testing circuit boards are made more rigorous by placing a higher voltage between the metal base and a circuit conductor, the frequency of the occurrence of dielectric breakdown in the subsequent withstand voltage test either remains the same or even becomes higher.

This fact means that the above-describe conventional method of withstand voltage testing is incapable of evaluating the withstand voltage or dielectric breakdown strength of a circuit board of interest. It is therefore necessary to establish a testing method by which circuit boards that will eventuallv fail a withstand voltage test to be conducted on packaged devices either in the process, or after the completion, of their manufacture can be definitelv identified and rejected as defective products.

The aim of the present invention is to provide such a method.

The withstand voltage testing method of the present invention is basically a method that is intended to evaluate the dielectric breakdown strength of a circuit board for a hybrid IC that comprises a metal base overlaid with an insulation layer that either consists of or contains an organic insulator and which in turn is overlaid with a circuit conductor. The method is characterized by measuring partial discharge that occurs in the insulation layer when a predetermined voltage is placed between the circuit conductor and the metal base of the circuit board, and evaluating the dielectric breakdown strength of the circuit board of interest in terms of the magnitude of discharge electricity.

Circuit boards that produce discharge electricity of an intensity higher than a predetermined level on partial discharge are highly likely to experience dielectric breakdown in a subsequent withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture, and these circuit boards should be rejected as defective products that have unacceptably low dielectric breakdown strength. In actual product inspection, it is not always necessary to measure a maximum value of the quantity of discharge electricity that is produced upon discharge and, more conveniently, circuit boards can be checked for their dielectric breakdown strength by examining whether discharge electricity is generated in an amount exceeding a predetermined level upon discharge when a predetermined voltage is applied to the circuit boards.

The term "discharge electricity" as used herein means electric charge that is yielded at a frequency higher than a predetermined value upon application of a predetermined voltage. Examples of the "predetermined value" of frequency are 10 pps, 20 pps, 100 pps and other values but usually 50 pps or 60 pps is conveniently selected in conformity with commercial frequency in cycles.

The term "circuit conductor" as used herein covers both a conductor circuit fabricated by etching a conductor foil on an insulation layer and the conductor foil to be etched.

The Applicants have closely studied circuit boards that passed a withstand voltage test performed by the conventional methods and which yet experienced dielectric breakdown in a subsequent withstand voltage test conducted on packaged devices either in the process, or after the comppletion, of their manufacture. As a result, the present Applicants found that the dielectric breakdown that occurred was due to voids present in the insulation layer of each circuit board. Probably, the withstand voltage test repeated by a packaged-device manufacturer or the application of high voltage to the circuit boar itself during product inspection would have cause dirge in voids in the in insulation layer of the circuit board, leading to deterioration of the organic insulator in the Insulation layer, which combined with the extremely small thickness of the insulation layer, a unique feature of the circuit board, to cause te problem under consideration, i.e., the circuit board passed the withstand voltage test conducted on itself during product inspection but yet experienced dielectric breakdown in the withstand voltage test subsequently conducted on the packaged device either in the process, or after the completion, of its manufacture.This mechanism is in good agreement with the already described observation that even when the conditions of withstand voltage test to be conducted on circuit boards by the conventional methods were made more rigorous by applying a higher voltage, the frequency of the occurrence of dielectric breakdown in the circuit boards during a subsequent withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture remained substantially the same or even became higher. In a test of the type contemplated by the present invention the state of void-induced internal discharge in the insulation layer of a circuit board is examined, and its dielectric breakdown strength is estimated on the basis of the examination.Such a test according to the present invention is believed to be indispensable in order to evaluate the dielectric breakdown strength of a circuit board in a reliable manner and to reject all products that have insufficient dielectric breakdown strength and can potentially experience dielectric breakdown in a subsequent withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture.

The partial discharge measurement is a method that examines internal discharge which occurs in circuit boards on account of such factors as voids in the insulation layer.

While various discharge-related parameters may be measured such as inception voltage, extinction voltage and average corona current, the method of the present invention comprises measuring discharge electricity due to internal discharge that occurs in the insulation layer of a circuit board when a predetermined voltage is placed between the circuit conductor and the metal base, with the dielectric breakdown voltage of the circuit board being evaluated in terms of the quantity of = measured discharge electricity. ore conveniently, the dieec--ic 'c breakdown strength of the circuit board of interest can be evaluated by checking to see whether or not discharge generating discharge electricity exceeding a predetermined level will occur upon application of a preQetermined voltage. This is because of the following two reasons: the greater the quantity of discharge electricity, the more extensive the resulting deterioration of an organic insulator, which leads to dielectric breakdown of the very thin insulation layer of the circuit board and hence is best adapted for the purpose of the present invention; secondly, the convenience of the test procedure is suitable for performing product inspection in routine work.

As will be described hereinafter, the present Applicants conducted an experiment to see whether the results of a withstand voltage test conducted on circuit boards over time would correlate well with the results of partial discharge measurements, and the experiment clearly shows that circuit boards that will experience dielectric breakdown within a short time can be identified and rejected as defective products by the partial discharge measurements-.

It has already been mentioned that a partial discharge measurement is sometimes conducted together with an AC voltage breakdown test or a DC impulse voltage breakdown test during the inspection of electrical products such as high-voltage power cables. However, such high-voltage electrical products are equipped with insulation layers of adequate thickness (e.g., 1 - 2 mm for the 0.6 kV class cables or 4 - 5 mm for the 3 - 6 kV class cables) and those which have passed an AC voltage breakdown test or a DC impulse voltage breakdown test will never experience dielectric breakdown in subsequent service within a short time in the absence of any special phenomena such as external damage and chemical corrosion.Further, the partial discharge measurements on high-voltage electrical products are primarily intended to evaluate the long-term reliability of their insulating performance extending to 30 or 40 years or more.

In contrast, the testing method of the present invention, which is intended to evaluate the dielectric breakdown strength of circuit boards, corresponds to the AC voltage breakdown test or DC impulse voltage breakdown test conducted during the inspection of electrical products such as high voltage power cables, and differs in both object and nature from the partial discharge measurements conducted on those high-voltage electrical products.

With respect to the circuit boards on which the testing method of the present invention can be used, those as disclosed in Electronic Technology, special issue (June, 1985), pp. 92-96, "Metal base wiring board" and Japanese Utility Model Registration Publication No. 46-25756 may be cited.

Specific examples of the circuit boards that can be tested by the method of the present invention are described below with reference to the accompanying drawings.

Figs. 1A - 1E are cross sections of five circuit boards having different structures except that a metal base 1 serving as a heat sink is overlaid with an insulation layer 2 which in turn is overlaid with a copper circuit conductor 3.

In the circuit boards shown in Figs. 1A and 1B, the insulation layer 2 is composed of a polymer film 21 such as polyimide film sandwiched between layers 22 of an adhesive such as an epoxy resin. In the circuit boards shown in Figs.

1C and 1D, the insulation layer 2 itself is made of an adhesive such as an epoxy resin with an inorganic filler such as alumina, silica or boron nitride being often incorporated in order to provide enhanced heat conductivity.

In the circuit board shown in Fig. lE, the insulation layer 2 is made of glass-epoxy. The circuit conductor 3 in the circuit board shown in Fig. 1E is a yet-tobe-etched conductor foil clad on the insulation layer 2. In Figs. 1A - lD, the conductor foil is etched to form a circuit. A seal that is made of an insulating or semiconductive material, and which covers the lateral sides and edges of the circuit conductor etched to form a circuit, is represented as 4 in Figs. 1B and 1D. The seal which is proposed in our Japanese Patent Application No.

61-31396 (corresponding to JP-A-63-164491) has the ability to suppress corona discharge from the edges of the circuit conductor. The term "JP-A" as used herein means an unexamined published Japanese Patent Application.

Guide figures for the dimensions of the circuit boards shown in Figs. 1A - 1E are as follows: the metal base generally has a thickness of 1 - 4 mm, preferably 2 - 3 mm; the circuit conductor generally has a thickness of 35 - 105 Rm; and the insulation layer generally has a thickness of 50 - 200 clam.

If the circuit boards shown in Figs. 1A and 1B use an polyimide film as the insulation layer, the polyimide film generally has a thickness of 20 - 50 Rm, particularly 20 - 25 pm to form the insulation layer having a thickness of 40 - 80 pm, particularly 40 - 50 gm. In the case of the circuit boards as shown in Figs. 1C and 1D using an inorganic filler, the insulation layer generally has a thickness of 50 - 150 Am, particularly 80 - 100 Am.

Fig. 2 is a top view of a conductor circuit formed by etching a conductor foil. For clarity, the circuit conductor 3 is hatched in Fig. 2.

In each of the circuit board shown in Figs 1A lD, voids are prone to form in the adhesive layer 22 or the insulation layer 2 (which is also an adhesive layer) when they are applied to the metal base, copper foil, polymer film, etc. The chance of void formation is particularly great when an inorganic filler is contained in the adhesive material or when the thickness of the adhesive layer formed by one application is fairly great and exceeds 30 ijm after drying.

Hence, circuit boards of the types shown in Figs.

iC and 1D and in which the insulation layer 2 is an adhesive layer incorporating an inorganic filler are very likely to form voids in the insulation layer and are most susceptible to dielectric breakdown in the insulation layer in a withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture in spite of the fact that those circuit boards passed conven tional withstand voltage tests. The testing method of the present invention will prove most effective in product inspec tion of such types of circuit boards.

The partial discharge measurement itself is well known and the testing method of the present invention permits the use of commercially-available instruments for partial discharge measurement, e.g., QM-3 manufactured by Mitsubishi Cable Industries, Ltd.

The partial discharge measurement which involves the measurement of a small voltage due to partial discharge that occurs upon application of a high voltage is highly susceptible to the influences of various kinds of noise.

Therefore, the testing circuit is designed to have a noise preventing capability either by isolating a noise source or blocking the path of its ingress or by eliminating the input noise. Such anti-noise provisions are also necessary in the circuit board to be tested in order to ensure that the intended measurement will not be interfered with by the noise of unwanted phenomena such as corona discharge from the edges of the circuit conductor. To meet this requirement, voltage applying electrodes are rounded or designed to have other structures and shapes that reduce the change of the occurrence of corona discharge and, at the same time, the circuit board under test is submerged in an insulation oil, together with other provisions such as the use of guardelectrodes.

Advantageous for these purposes are circuit boards of the types shown in Figs. 1B and 1D that have the lateral sides and edges of the circuit conductor sealed with an insulating or semiconductive material. With such circuit boards, corona discharge from the edges of the circuit conductor is effectively inhibited by the seal 4, so that testing can be performed in air without any cumbersome steps such as submersion in an insulation oil and its removal from the circuit board after the test.

Examples of the insulating material include soldering resist materials composed of thermosetting epoxy resins or ultraviolet curing resins and as commercially-available forms thereof, S-222 and UVR-150G from Taiyo Ink Mfg. Co., Ltd. may be cited. As the semiconductive materials, those obtained by adding carbon fillers to the above-described insulating materials may be used. The details are disclosed in JP-A-63-164491.

When testing a circuit board of the type shown in Fig. 2 which has several circuit conductors insulated from one other, voltage applying electrodes are brought into contact with each conductor and collectively connected to the testing circuit. Further, the testing method of the present invention is intended to be used in product inspection, so multiple products, say 30 circuit boards, are desirably tested at a time, with the voltage applying electrodes on all circuit boards to be tested being collectively connected to the testing circuit.

The voltage to be placed between each circuit conductor and the metal base is suitably selected from the range of AC 1 - 3 kV in the case of the present invention, with a voltage of AC 2 kV being typically applied. The time for applying the voltage is preferably within one minute.

In the method of the present invention, the amount of discharge electricity determined by partial discharge measurement is used to evaluate the dielectric breakdown strength of the circuit board under test. The amount of discharge electricity used as the criterion is by no means a fixed value since it varies with factors such as the thickness of the insulation layer, its nature (e.g., how soon it will deteriorate upon internal discharge) and the operating voltage on the packaged device that uses the circuit board of interest. As a guide figure, a value of 10 - 20 pC (picocoulomb) is typically selected. Strictly speaking, the amount of discharge electricity is preferably determined by such methods as comparison with the results of a prolonged voltage application test.

An experiment was conducted in the following manner to compare the results of a prolonged withstand voltage test with partial discharge measurement conducted in accordance with the testing method of the present invention.

EXAMPLE A 35 Am thick copper foil was bonded to a 2.0 mm thick aluminum base, with an adhesive layer of epoxy resin (composed of a phenol novolak type epoxy resin and a bisphenol type epoxy resin) containing an inorganic filler (alumina, average particle diameter 0.6 pm) being interposed to a thickness of 100 pm. The copper foil was then etched to form a circuit conductor, whereby samples of circuit board for a hybrid IC were fabricated. These samples were submerged in an insulating oil (No. 2 prescribed by JIS C 2320) and an AC voltage of 2 kV was applied between the circuit conductor and the aluminum base, with maximum discharge electricity at 60 pps being measured with a partial discharge meter Q-3 of Mitsubishi Cable Industries, Ltd.Ten of the samples, five producing maximum values of discharge electricity of 20 pC or more and the other five producing values smaller than 1 pC, were subjected to a prolonged withstand voltage test, using a life tester for continuous applIcatIon of voltage manufactured by Totoku Toryc S.K.,nn which an AC voltage of 2 kv was placed between the circuit conductor and the aluminum base in the air. The maximum discharge electricity produced by each sample, the time of voltage application until breakdown and the site of breakdown were as showngin Table 1 below.

Table 1 Time to Maximum discharge dielectric Site of Sample No. electricitv breakdown breakdown (pC) 1 25 15 min. under copper foil 2 20 20 min. do.

3 50 4 min. do.

4 42 5 min. do.

5 30 10 min. do.

6 less than 1 1,600 hr. under the edges of circuit conductor 7 less than 1 2,000 hr. do.

8 less than 1 2,500 hr. do.

9 less than 1 2,900 hr. do.

10 less than 1 2,500 hr. do.

As is clear from Table 1, Sample Nos. 1 to 5 which produced large discharge electricity experienced dielectric breakdown within a very sort tie and the failure occurred in the insulation layer under t copper foil formIng the circuit conductor. On the other hand, Sample Nos. 6 to 10 which produced small discharge electricity experienced dielectric breakdown under the edges of the circuit conductor.These results show that the samples producing large discharge electricity failed early because of defects in the insulation layer, whereas the samples producing small discharge electricity had no defects in the insulation layer and experienced dielectric breakdown only after prolonged voltage application when corona discharge from the edges of the conductor circuit caused deterioration of the insulation layer. Therefore, the results given in Table 1 clearly shows that by evaluating the dielectric breakdown strength of circuit boards for hybrid ICs in terms of the magnitude of discharge electricity in accordance with the present invention, products that pass the conventional 1-min withstand voltage test but which experience dielectric breakdown within a short time can be rejected with reasonable precision.

The conventional withstand voltage test is unable to evaluate the dielectric breakdown strength of circuit boards in a precise manner and even circuit boards that pass the test will often experience dielectric breakdown of the insulation layer in a subsequent withstand voltage test conducted on packaged devices either in the process, or after the completion, of their manufacture using those circuit board. However, in accordance with the testing method or the present invention, circuit boards that have insufficient dielectric breakdown strength to avoid such as accident can be identified definitely and rejected from the products to be delivered.

There is another advantage of evaluating the dielectric breakdown strength of circuit boards by the testing method of the present invention rather than the conventional withstand voltage test. That is, in many cases, the voltage to be applied can be made lower than in the conventional withstand voltage test, so the chance of deterioration in the insulation layer due to the application of high testing voltage is reduced.

Claims

1. A method of testing for withstand voltage in product inspection of a circuit board for a hybrid integrated circuit, which comprises a metal base overlaid with an insulation layer comprising an organic insulator, and which in turn is overlaid with a circuit conductor, which method comprises measuring the partial discharge that occurs in the insulation layer when a predetermined voltage is applied between the circuit conductor and the metal base of the circuit board, and evaluating the dielectric breakdown strength of the circuit board of interest on the basis of the magnitude of discharge electricity.

2. A method according to claim 1, wherein the insulation layer in the circuit board comprises adhesive incorporating an inorganic filler.

3. A method according to claim 1 or 2, wherein the sides and edges of the circuit conductor are sealed with an electrically insulating or semiconductive material.

4. A method according to claim 1, 2 or 3, wherein the voltage applied between the circuit conductor and the metal base of the circuit board is 2 kV.

5. A testing method as claimed in claim 1 and substantially as herein described.

6. A method of testing for withstand voltage substantially as herein described with reference to the accompanying drawing or to the foregoing Example.

7. A circuit board which has passed a test as claimed in any preceding claim.

8. The features as herein disclosed, or their equivalents, in any novel patentable selection.