Automotive Electronics Environment Increases Need for Fast, Accurate Tests

Automotive Electronics Environment Increases Need for Fast, Accurate Tests
Today’s automotive electronics have expanded far beyond bread-and-butter convenience items, such as climate control and audio systems. Now, electronics control all critical aspects of car performance, from air bags to transmissions. This expansion is accompanied by competitive pressures and governmental regulations that place stringent demands on cost, quality and performance.

These pressures drive the demand for more and better testing on all types of automotive electronic components and assemblies. Some examples include:

Qualification testing during product development.

High-speed component testing.

Test and adjustment of assemblies.

Increased sampling rates and QA testing.

Better testing often means increasing measurement accuracy and speed. A good example is the electrical connector. Connectors constitute a carmaker’s #1 warranty problem in terms of occurrences and numbers of claims.

Connector failures often are contact-related problems such as corrosion or insulating films. Correcting these problems requires highly accurate low-resistance testing which may need to be performed in an environmental chamber to simulate service conditions.

Low-Resistance Measurements

Low-resistance measurements are required on many types of automotive electronics assemblies. For example, to ensure deployment, the contact resistance of connectors used in an air-bag assembly must stay below a critical value for the life of the car. You can use resistance testing to gauge the reliability of those contacts. Similarly, resistance testing can check the integrity of the air-bag igniter squib. In both cases, low levels of current and voltage must be used.

Contact resistance normally ranges from microohms to a few milliohms. The measurement of resistance in this range requires special techniques and instrumentation to ensure accurate, repeatable readings. Typically, the four-wire method is used. Current is forced through the device under test (DUT) through two of the leads, and voltage across the DUT is measured with the other two leads (R=V/I).

Insulating films, resulting from surface contamination, can adversely affect the contact resistance. Films range in thickness from a few angstroms to several hundred angstroms and vary in composition, but they typically consist of metallic oxides, sulfides and halides. They can add series resistance ranging from a few milliohms to a few ohms.

Using excessive test voltage can result in erroneous contact-resistance measurements because it can break down surface films. Voltage sufficient to cause film breakdown typically ranges from 30 mV to 100 mV. For this reason, testing usually is conducted under dry-circuit conditions, commonly defined as an open-circuit voltage below 20 mV and source current below 100 mA.

Selecting/Configuring Test Equipment

Automotive-electronics suppliers no longer have to rely on standard instruments and less-than-optimum work-around solutions. For example, there are digital multimeters (DMMs) available with a dry-circuit voltage clamp that limits test voltage to 20 mV for contact-resistance measurements.

By comparison, open-circuit voltages on many standard DMMs can be as high as 5 V. Because a lower-source current produces a lower voltage across a measured resistance, DMMs designed for dry-circuit testing also have greater measurement sensitivity.

Another mode found on some DMMs is a separate low-power ohms test which minimizes device self-heating by limiting the test current to about 100 µA. This is an order of magnitude lower than most DMMs and makes the instrument suitable for air-bag igniter-squib resistance testing as well as end-of-life contact testing per ASTM B539-90.

Throughput and Accuracy Issues

For high volume production, instruments must process a large number of DUTs in a short amount of time. Automating the measurement system with computer-controlled instruments decreases test time. Moreover, these instruments facilitate complex testing, such as when connectors and assemblies are cycled and measurements are taken in environmental chambers.

Many instruments can be computer controlled via IEEE 488 or RS-232 interfaces, but you need to find those that have all the other features needed to ensure accuracy as well as throughput. A programmable scanner is recommended to accommodate contact resistance measurements in multiterminal connectors and to perform production testing of multiple components. Scanners greatly reduce measurement time by switching one set of test instruments to multiple contacts on a DUT or test fixture. The scanner should use four-pole type switches to facilitate four-wire measurements (Figure 1).

Also, the relays that switch voltage-measuring lines should have a very small inherent voltage potential across each set of contacts when closed. A contact potential of less than 0.5 µV is recommended to avoid voltage offsets.

Measurement accuracy and test throughput are closely related, particularly when making rapid multiple measurements of low-level voltage, current or resistance. System accuracy should be at least an order of magnitude better than the specified tolerance of the DUT. Within that constraint, you should have the capability to make trade-offs between accuracy and speed.

The most useful instruments allow you to program the delay between the time when a measurement is initiated and when a reading is actually taken. For example, you might choose between 1,000 readings/s at 4-1/2 digits of resolution or 500 reading/s at 5-1/2 digits.

One settling-time issue is relay contact bounce. Instrument manufacturers can reduce this time by using dry-reed or solid-state relays instead of conventional mechanical relays. This reduces the settling time from about 3 ms to less than 0.5 ms and improves reliability.

There also is settling time associated with A/D converters and amplifier circuitry in test instruments. A well-designed servo amplifier front end provides low offset drift and autozeroing to speed up measurements.

An instrument with a wide dynamic measurement range can help improve throughput. With this instrument, range change delays and range shift errors are minimized. Also, a large buffer memory can help the instrument cope with high data- acquisition rates. Look carefully at these specs when choosing an instrument for throughput and accuracy.

Programmable Instrument Features

Besides scanners, programmable instrument features can shorten the test cycle. In the past, you programmed the host computer to control a test sequence.

The basic idea in new instrument designs off-loads some of the host’s tasks onto the measurement instrument, particularly tasks requiring data communications between the two. Some of the functions now built into instruments that help accomplish this are:

Programmable pass/fail (compliance) checking.

Calculation and conversion functions.

Automatic switching of source/measure/compare functions.

ICs with greater functionality have allowed instrument manufacturers to expand internal memories and increase logic functions. Now the instrument controls many test parameters that formerly required computer intervention.

For example, 100 or more measurement configurations can be stored in the memory of some instruments, containing parameters such as source setting, measurement setting, math calculations and pass/fail criteria. Start-of-Test, End-of-Test and Pass/Fail results often can be exchanged directly between the DUT handler and the instrument. This reduces system overhead and speeds up the test sequence by eliminating computer control of these functions.

The use of pass/fail criteria based on compliance settings in the instrument is particularly valuable when performing QC tests. In many cases, you do not need to know the actual value of a voltage, current or resistance; you simply need to know that it falls within an acceptable range. Taking an actual reading slows down the test sequence compared to a compliance check, which might be performed in less than 500 µs.

Component Manufacturing Trends Affecting Testability

There is a growing trend in automotive electronics to develop integrated assemblies of passive components manufactured with thin-film technology. This shrinks the size of the components and provides more complex functions in a single assembly.

While this lowers assembly costs, it makes electrical measurement of individual components difficult and may increase test costs. For example, in resistive networks, measurement errors can occur due to parallel resistance paths.

The problem can be illustrated by a dual-terminator network frequently used as an impedance-matching circuit (Figure 2a). In this network, each resistor is connected to a series resistor and to several series resistor loops in parallel.

The electrical equivalent is a group of delta circuits (Figure 2b). Because of shunting resistances, the apparent value of adjacent resistors will be less than that of each individual component if conventional measurements are taken.

A six-wire ohms measurement technique eliminates the effects of shunt resistance. This method, which builds on the industry-standard four-wire ohms test method, requires the addition of a low-impedance guard buffer with sufficient drive current for a four-wire ohmmeter.

As shown in Figure 2b, the electrical equivalent circuit can be characterized as a set of three resistive elements, each connected in a triangular fashion. Figure 3 shows the electrical equivalent circuit when measuring resistor element R1 of the network shown in Figure 2a.

The solution to this problem is to electrically guard out the parallel resistance by forcing the voltage on the middle point of the two bridging resistors (R₂ and R_L in Figure 3) to the same potential as Source HI at the test instrument.

The six-wire ohmmeter guard buffer accomplishes this by forcing current through R_L, which represents the equivalent parallel resistance shunting R₁ and R₂. The guard buffer is a unity-gain amplifier (op-amp) with a shunt resistor across its input. The voltage between its inputs and, therefore, across the resistor (R₂ in Figure 3), is nearly 0 V. In the technical literature, this methodology is sometimes referred to as a voltage- follower technique.

As a practical matter, this technique is greatly simplified and improved when the measurement instrument is designed for it in the first place. Such instruments have built-in Guard Sense and Guard Output terminals in addition to the usual Source HI, Source LO, Sense HI and Sense LO terminals.

If a scanner is used, it must have a contact configuration that allows six-wire ohms measurement. Also, the Guard Sense terminal, essentially the inverting input of the amplifier, has to be connected to the same point as Guard Output to compensate for cable, contact and switch relay resistance in a component handler’s test head.¹

Reference

1. “Configuring a Resistor Network, Production Test System With the Model 2400 Digital SourceMeter™,” Keithley Application Note #802, 1996.

About the Author

Mark Cejer has been employed by Keithley Instruments for six years, where he is strategic marketing manager for the automotive and component markets. He has held various product management positions with the company since receiving a B.S.E.E. degree from the University of Akron and an M.B.A. degree from the Case Weatherhead School of Management. Keithley Instruments, 28775 Aurora Rd., Solon, OH 44139, (216) 498-2873.

March 1997