Cook Your Chips and PCBs
Read the T&MW Product Survey of Environmental Test Chambers.
This article also ran in Test & Measurement Europe. Read the PDF Version Here.
|A variety of thermal environmental-test products can help you test electronic devices and subassemblies and subject them to temperature extremes. They range from benchtop hot plates to walk-in chambers. In between, you’ll find a plethora of ovens, conductive coolers and heaters, and forced-air sources. (See our product survey, below)
Burn-in is the traditional form of environmental test, in which you subject devices to their rated voltages and temperatures either before or during production test to weed out devices that would quickly fail in the field. An alternative to burn-in—a process that can take hours—is environmental stress screening (ESS), a process that seeks to hasten infant mortalities by rapidly cycling devices under test through high and low temperature extremes while subjecting them to high supply voltages and—optionally—vibration.(1)
If you intend to cycle your devices through hot and cold extremes you’ll need a chamber that cools and heats. If your cooling requirements are modest, you can employ a closed mechanical refrigeration unit or a water-cooled system. For lower temperatures you’ll need to supply liquid CO2 to provide cooling down to –65°C. Below that, you’ll need liquid nitrogen (LN2).
Lab and Bench Test
For thermal testing in the lab you have several choices. If your product uses a heat sink for direct-conduction cooling, you can employ a benchtop thermal platform—a hot (or cold) plate onto which you can clamp your product’s heat sink. Environmental Stress Systems (ESS) makes such products; the company’s T600 Series thermal platforms cover a –65°C (LCO2) or –100°C (LN2) to +200°C range and offer 2x2-in. to 11x22-in. plate areas.
If your product must be convection cooled, you can select a benchtop temperature chamber. An alternative is a benchtop conditioned-air source such as Temptronic’s Model TP04100A ThermoStream system ( Fig. 1), which can focus 2 to 6 SCFM (cubic feet per minute at standard room temperature, pressure, and humidity) of heated or cooled air onto a device under test.
For thermal evaluation of production devices, you can choose from a variety of floor-standing environmental chambers, such as the one in Figure 2. Or you can use a temperature forcing system to deliver conditioned air to your ATE test site.
Depending on your manufacturing setup and on what equipment you already own, you might choose a system that provides thermal-control capabilities as well as other functions, such as pick-and-place device handling (Fig. 3).
To control the burn-in temperatures of production devices, you will probably need a combination of the hot-plate and chamber approaches. The factory-floor analog of the lab hot plate is a clamshell arrangement that brings a conductive heat sink or source into contact with each device on a burn-in board. If your product doesn’t lend itself to conductive heating and cooling, you can choose a simple oven to convectively apply thermal stress to your devices before testing them. Going a step further, you can cycle the devices through thermal extremes at programmable rates of temperature change. The limit of such testing is the thermal shock test, in which your product plunges nearly instantaneously from a hot thermal chamber to a cold one, or vice versa.
Don’t overlook the need for cooling even if your application calls only for burn-in. If you’re burning-in multiple VLSI devices at once with power applied, you’ll find they often generate more than enough heat to keep your burn-in chamber at rated temperature, and the challenge will be to safely channel away the excess heat.
For best results, you’ll want to electrically test your devices while subjecting them to environmental test (see "Control Chip Temperature During VLSI Device Burn-in," Test & Measurement World, April 1999, p. 67). That complicates matters. You have to choose a system that not only meets your environmental requirements but also includes—or is compatible with—the necessary instrumentation. In addition, your temperature control is complicated, because the controller must contend with various levels of device heat dissipation in response to varying input test patterns.
Temptronic doesn’t provide instrumentation. For its ThermoFixture thermal-inducing enclosure, however, it does offer interfaces to ATE from many companies.
Accurate Temperature Control
Precise thermal control and fast thermal response are critical in today’s test environments. Accuracy is paramount if you are characterizing temperature-sensitive components such as crystal oscillators. For such applications, you can employ systems such as Saunders & Associates’ temperature test systems, designed to characterize crystals, capacitors, diodes, resistors, and ECL, TTL, CMOS, and HMOS oscillators. These systems comprise temperature chambers as well as instrumentation. Temperature accuracy is ±0.1°C.
If you’re burning-in VLSI devices you won’t need ±0.1°C accuracy. Accuracy is nevertheless important, especially when testing high-density chip-scale packaging technologies.(2) For example, the low thermal mass of a flip-chip or C4 package can result in the device quickly reaching destructive temperatures in response to changes in dissipation due to varying input test patterns. Schlumberger has addressed this issue with its ETC 1000 (Fig. 5), a thermal controller that can maintain DUT package temperatures to within ±0.1°C.
Bake, Shake, and Sweat
A combination of thermal cycling while applying other stresses (such as vibration, altitude, and humidity) can expedite your environmental test. This ESS approach can accelerate the infant mortalities that might appear only after hours of burn-in. Moreover, by realistically simulating your product’s target environment, you can verify that your product is up to its intended task.
Two other ESS implementations are highly accelerated lift test and highly accelerated stress screening, developed by Qualmark founder Gregg K. Hobbs. HALT and HASS both combine thermal and vibration stress. HALT is a destructive test used to estimate the potential lifetime of a product under development; HASS is a nondestructive test used to qualify production devices.
More Stress Sooner
The future is likely to see environmental testing earlier in the production process. Burn-in at the die and wafer level are two approaches being pursued; these approaches avoid the expense of packaging dice that would fail during a traditional package burn-in process.
In addition to employing environmental test early in production, efforts are underway to shorten burn-in and ESS times. One approach is to subject devices to high temperatures and overvoltages (for instance, 8.5 V on a 5-V part) for a matter of milliseconds instead of rated voltage for a matter of hours. This technique, called accelerated dynamic burn-in (ADBI), shows promise for detecting weak parts when combined with IDDQ testing.3 Though not yet ready for prime time, ADBI shows promise as efforts proceed to provide maximum reliability in minimum time. T&MW
1. Kececioglu, Dimitri B., and Feng-Bin Sun, Environmental Stress Screening, Prentice-Hall, Upper Saddle River, NJ, 1995, ISBN 0-13-324229-3.
2. Malinoski, Mark, et. al., "A Test Site Thermal Control System for At-Speed Manufacturing Testing," Proceedings, International Test Conference 1998, Washington, DC, pp. 119-128, www.itctestweek.org
3. Dudinski, Frank E., "Why Burn-in and Is It Necessary?" Yamaichi Electronics USA, San Jose, CA, www.yeu.com/banner3.html
FOR FURTHER READING
Izumi, Jyuro, "Combined environmental testing for equipment used on automobiles," Technology Report No. 6, September 30, 1998, pp. 9–17, Tabai Espec Corp., Osaka, Japan, www.espec.com/features/tech_report.htm