12 Flexible Printed Circuit Boards
14.4 Testing of Printed Circuit Boards
The manufacture of printed circuit boards comprises a large number of process steps. In order to produce a quality product, it is necessary that proper inspection and testing are carried out at every step in the process. In spite of conducting in-process quality control measures, final testing is vital to ensure the reliability of the product.
Open circuit Short circuit Warped
Marginal performance – Speed
– Temperature Bare PCBs
Wrong label or value Partially or totally inoperative Received components
Module assembly Component
insertion
Component PCB soldering
Wrong position Wrong orientation Physical damage
Cold-soldered joints Shorts Damaged or open lands
All previous faults Interactive or out-of-tolerance faults
Finished module
Fig. 14.2 Spectrum of module faults: approximately 75 per cent of module faults occur in assembly, 20 per cent are component faults, and 5 per cent PCB faults (Adopted from Coombs, 1988)
Quality, Reliability and Acceptability Aspects 571
From experience and statistics, it has been estimated that approximately 75 per cent of faults in an assembled PCB occur in assembly, 20 per cent are component faults and 5 per cent are the board faults. Figure 14.2 shows some common faults in assembled printed circuit boards.
14.4.1 Automatic Board Testing
With the introduction of surface mount technology, the board packing density has increased manifold.
So, even for boards of modest density and moderate board quantities, automatic board testing is not only essential but economical also. Two approaches have been common in the testing of complex boards: bed of nails method and two probe or flying probe method.
14.4.1.1 Bed of Nails Method
This method consists of a spring loaded pin brought down on each test point on the board. The spring action enables one to achieve a pressure of 100-200 grams to ensure a good contact on each test point. An array of such test pins is called ‘bed-of nails’. Under the control of test software, test points and the signals to be applied to them can be programmed: Figure 14.3 shows a typical bed-of- nails fixture set-up. The tester is provided with all the information about the test points and only those pins which correspond to the required points are actually set-up. While designing the board, it is advisable to keep all the test points on the solder side of the board, though it is possible to conduct tests simultaneously on both sides of the board with the bed-of-nails approach. The equipment is expensive and difficult to maintain. Nails come in different pin head arrangements depending upon their applications.
PCboard
Bed of nails
Test fixture Vacuum port Board under test
Test nail Prode
Spring Socket
Wire-wrap connection
Bed-of-nails fixture
Receiver To tester
Receiver
Fig. 14.3 Bed-of-nails fixture for automatic board testing (redrawn after Coombs, 1988)
A basic universal grid handler consists of a drilled plate populated with pins spaced on 100, 75 or 50 mil centres. The pins act as probes and make direct mechanical contact with electrical connections or nodes on the circuit boards. If the pads on the boards match the test grid, a mylar sheet drilled to specification is placed between the grid and the circuit board to enable design specific probing.
Continuity is tested by accessing the end points of a net, defined as the X-Y co-ordinates of a pad.
Since every net on the board is tested for continuity, an isolation test is thus performed. However, the effectiveness of a bed-of-nails fixtures is limited by pin proximity.
14.4.1.2 Two Probe or Flying Probe Method
The flying probe tester does not rely upon a pattern of pins built onto a fixture or holder. Depending upon the system, two or more probes mounted onto tiny heads move freely in an X-Y plane to test points as directed by CAD/Gerber data. The two probes can come within 4 mil of each other. The probes move independently and there are no real limitations in terms of how close they can get to each other.
A tester with two roving arms can be based on the measurement of capacitance. The board forms one plate of a capacitor by being pressed onto a dielectric layer over a plate. In case of a short-circuit between the tracks, the capacitance will be larger than it should be at a particular point. If there is an open circuit, the capacitance will be lesser.
The speed of testing is an important criterion for the selection of the tester. While a bed-of-nails fixture can test literally thousands of test points at a time, the flying probe is only able to test two or four points at a time. Also, while single-sided testing on a bed-of-nails may take only 20 to 30 seconds, depending upon the complexity of the board, a flying probe requires as much as an hour or more to perform the same evaluation. Shipley (1991) explains that even though the moving probe technology is considered slow by high volume PCB manufacturers, the method does offer benefits to the lower volume producer of complex boards.
For testing of bare boards, special types of fixtures are designed (Lea, 1990). A cost-effective method is to have a universal fixture. Although this type of fixture is initially more expensive than a dedicated fixture, its initial high cost is offset by reducing the individual set-up costs. In a universal grid, the normal grid for leaded component boards and for surface mount devices is 2.5 mm. Here, the test pads should be greater than or equal to 1.3 mm. For grids of 1mm, test pads are designed to be greater than 0.7 mm. In case of smaller grids, the test pins are small and fragile and easily get damaged. So, it is preferable to have the grid greater than 2.5 mm.
Crum (1994b) illustrates that using a combination of universal (standard grid testers) and flying probe testers enables one to provide accuracy and economy in testing high density boards. Another approach suggested by him is the use of conductive rubber fixtures. This technique makes it possible to test off-grid points. However, the varying heights of solder pads created by hot-air solder levelling prevents all the test points from being contacted.
Testing is usually carried out at the following three levels:
a Bare Board Testing;
a In-circuit Testing; and a Functional Testing.
Quality, Reliability and Acceptability Aspects 573
The testers can be universal type that may be useful for a range of board styles and types or they may be dedicated for a specific application.
14.4.2 Bare Board Testing (BBT)
The increasing track density and increasing number of via holes make it important to test the PCB before the assembly operation begins. An earlier PCB on an average had about 400 through-holes out of which about 25 per cent were via holes and an average of 200 networks. As the density has increased, a typical PCB is likely to have about 2000 through-holes, 40 per cent of which are via holes and a network connection of about 600. These highly populated PCBs have increased the failure rate, which may at times be even upto 20 per cent.
Some extensive failure mechanisms have been identified, which can lead to a total circuit failure at the advanced level. These failures may prove to be extremely expensive in the case of high density and multi-player PCBs, which have a track separation of less than 0.01 inch (0.25 mm).
It is also observed that under extreme humid conditions, there will be an electrochemical growth on the copper which can lead to some shorts. Improper applications of dry film solder mask leads to voids appearing on the board, especially between closely spaced conductors. The voids can entrap moisture and dust particles which can lead to shorts. These types of failures occur because of shorts, opens, cuts, leakage and contamination. Hence, the boards must be tested before PCB assembly.
With the increasing track density and number of through-holes, it has become necessary to test the printed circuit board before assembly. It has been observed that the failure rate in highly populated printed circuits may be as high as 20 per cent. If the boards are not tested at the pre-assembly stage, the failures at a later stage may prove to be extremely expensive in the case of high density and multi-layer boards. Before populating a board with expensive devices such as application-specific ICs and microprocessors, it is cost-effective to first check whether the bare board meets expected quality standards. Bare board testing is thus becoming mandatory for the PCB manufacturers.
14.4.2.1 Causes of Boards’ Failure
Normally the reliable functioning of a PCB is taken for granted unless a serious field failure occurs.
A PCB is expected to perform perfectly well with respect to the application for which it is designed.
It should be able to withstand the environmental effects depending upon the mechanical, electrical and chemical nature of the product. Generally, the performance and reliability of the PCB to a large extent, depends upon the electrical, mechanical and chemical characteristics of the copper clad laminate used in the manufacture of the PCB.
The presence of dust or any other form of contamination on the board can result in electrical leakage, corrosion and failure of electronic components which can lead to the failure of the total equipment. These contamination tests can be performed on a bare board level and some of the faults that are generally found on a single-sided and double-sided PCB, such as opens, shorts, silvers, etc.
can be easily corrected whereas the same faults occurring on the inner layers of multi-layer boards are difficult to eliminate.
14.4.2.2 Testing Techniques
As the PCBs are becoming more and more dense, the testing of a bare board is becoming mandatory for the PCB manufacturers. There are fully automatic equipments available which can test anywhere between 10,000 to 50,000 nodes and networks. The electrical tests must be performed at very high voltages to check for contaminations.
The commonly used methodologies for bare board testing are automated optical inspection (AOI) and electrical test. AOI is an in-process tool to check on the inner layers but is limited in its ability to verify, for example, the electrical continuity of plated through-holes. Similarly, AOI cannot identify poor multi-layer interconnections after lamination (Dytrych, 1993b). The electrical tests, on the other hand, help to identify potential defects like ionic contamination and hairline cracks, which are difficult to detect through AOI. In case of small geometries, most fabricators are doing a combination of electrical tests and automated optical inspection on final product. Straw (1992) details the solutions to some of the fine pitch bare board electrical test challenges.
Bare board testing generally checks for short-circuits between tracks and continuity of tracks.
These tests can be performed by fully automatic machines which can test upto 50,000 nodes and networks.
The cuts and shorts in case of single-sided and double-sided PCBs can also be detected by the age-old method of visual inspection by optical means. A more sophisticated method of inspection of multi-layers is by the use of X-ray scanning, which gives an in-depth inside view of the board structure, thereby enabling one to detect the faults quickly.
14.4.2.3 Electrical Tests
Isolation Resistance: For performing this test, initially the test voltages are accurately determined in order to ascertain the resistance. Thus, the test is performed from one network to all other networks.
Once the leakage is established, the tests can be repeated from network to network in order to determine the exact leakage path and the same could be re-worked. Generally, the isolation resistance is performed by applying 100 to 250 V and the failure threshold will be typically in the range of 10 to 100 Mohm.
Breakdown: The voltage breakdown tests are dependent upon time and the voltage applied. Generally these tests are performed for 10 ms during which a consistent result can be achieved. However, during this time, the energy discharged should be controlled in order to avoid damage to the PCB under test. This breakdown test may not be necessary in case of boards where their operating voltage is low.
Continuity: This is one of the commonest and most widely used tests performed on any PCB. These tests are generally conducted between two nodes on the same network in order to ensure accurate results. Continuity measurements are usually made in a 10 ohm to 500 ohm range.
14.4.2.4 Fixtures Used for Bare Board Testing
Even though most of the bed-of-nails fixtures used for BBT look alike, several factors need to be considered in order to maximize the use of the fixture and to make them cost-effective. For this
Quality, Reliability and Acceptability Aspects 575
reason, most of the grids are wired upto 50,000 test points and used as a general universal fixture even though individual test fixtures could be designed or purchased.
The Standard Grid Pattern: The basic idea that has to be borne in mind while selecting any grid fixtures is to minimize the nodal test points but at the same time to be able to test maximum area.
One of the simplest ways of doing this is to etch a number of smaller circuits on one bigger panel which saves lot of handling time and consequently the cost involved in checking. The smaller boards can thus be cut out of the bigger panel even after the assembly time. Testing these small boards will be more time-consuming and on the long run will work out to be more expensive. But the testing of these small boards as part of a bigger panel will not only reduce the time and cost of testing but eventually result in a very high output.
Another important feature that determines the purchase of a grid system is its cost. The purchase of a dedicated grid system helps to reduce the initial set-up cost. But universal test grid, though it may initially work out to be more expensive, will offset its initial cost by reducing the individual set-up costs. Also, dedicated fixtures require lot of storage area as compared to a universal grid pattern. But one of the major problems faced with universal grid fixtures is their inability to accommodate off grid test probes.
Even with CAD designs, it is often found that some of the test points will not be exactly on the grid points but will be between two such grid locations. With dedicated grids, these points could be easily accommodated. It needs to be mentioned here, that with the increasing track density of the boards, the spacing between the tracks becomes much less than the diameter of the nail. Hence there is a question mark in the industry about the bed-of-nails method for high density PCBs.
The latest generation of PCB technology comprises quad flat packs (QFPs) with centres of 0.020 in. and less, ball grid arrays (BGAs) with centres of 0.050 in. and less, and other devices with small pad geometries. These components have dictated that new fixture designs be developed to accommodate the new technology. Hallee (1996) suggests that in certain fixture designs, longer pins with greater deflection can virtually eliminate the need for expensive double-density grid electronics. It is also pointed out that if reliable, repeatable mechanical contact with the PCB fails to occur, the testing process is compromised, regardless of the quality of the electronic measurement system involved.
14.4.2.5 Hold-down Systems in BBT Equipments
There are generally two types of hold-down systems available for bare board testing. It could be either a vacuum hold-down or a pneumatic hold-down. The dedicated fixtures are mostly designed to be used with a vacuum hold-down system. Vacuum hold-down is much less expensive than pneumatic systems. In this system, the board to be tested is placed over the test probe and a rubber blanket is placed on the board. Air is drawn out of the fixture by using a pump, thereby giving a contact between the probes and the test points. In the case of pneumatic hold-down system, air cylinders are used to compress and hold the board firmly against the bed of nails. Many universal grid fixtures cannot operate properly with a vacuum hold-down system because the pressure applied on the probes will not be sufficient to establish a perfect control.
Some universal grid fixtures cannot operate satisfactorily with a vacuum hold-down system because the pressure applied on the probes will not be sufficient to establish a perfect control. For example, a universal pattern with about 20,000 test points and each test point requiring a 100 gram pressure, would need 2000 kg or 2 ton of air pressure to keep all the pins in good contact with test pins on the board. If the pressure is not adequate, the test consistency is not satisfactory. On the other hand, with pneumatic systems, the pressure from air cylinders could be much more to give a high throughput.
Cronin (1995) explains the various methods used to generate bare board test programmes which provide different levels of test coverage.
Probe Assembly: In a universal grid fix- ture, when the centre-to-centre spacing is 0.075 inch (1.875 mm) or more, the probes (Figure 14.4) are interchangeable. Hence they are mounted on individual sleeves or sockets. The plunger head should be so de- signed as to make perfect contact with the test point. In order to assure positive elec- trical contact, the probe should exert a pres- sure of 4 to 8 oz of force during normal en- gagement. In the universal grid system, the probe should have a minimum traverse of 0.2 inch (5 mm) to accommodate the use of the through-hole mask. Internal resistance of the probe should be less than 10 Mohm for proper testing.
The testing of the boards could be done with the already existing programme, which is generated during the design stage itself, if the PCB design is done by using a com- puter. Most of the bare board testing ma- chines have the self-learn mode. In such cases, the already established “good” board is placed as a reference to obtain the neces- sary test parameters. With these parameters as the base, the other boards could be tested.
M/s MicroCraft have introduced a bare board tester which is a moving probe fixtureless tester and performs continuity and isolation tests simultaneous on both sides of a PCB (www.vikingtest.com).
The design includes four moving probes, high speed closed loop servo motors, firmware and Windows interface. The four moving probes (two front, two rear) move in X, Y and Z directions and test all necessary pad locations for continuity and isolation. The continuity test involves high precision
Tubular housing
Spring Contact plunger Circuit board node-plunger contact junction
Primary junction between plunger and housing in PLP1600 and P2500series probes Various secondary probe housing- receptacle tube junction
Wirewrap pin-wire junction Various possible
junctions between plunger and housing Plunger-spring junction
Various spring-probe housing junctions Primary probe housing-receptacle tube junction
Receptacle tube
Fig. 14.4 Spring contact system for probe assembly