내후성 및 신뢰성 측정 연구실

By 칼 슈람, 신뢰성 테스트 연구소장 / RECOM Development & Trading GmbH & Co KG.

High product lifespan and reliability are a must for products that are intended for high end applications. But how can such a product be recognized? Every manufacturer takes pride in calling themselves a high quality manufacturer, but how can this be proved? For this reason, RECOM has created a new Environmental and Reliability Test Lab in Gmunden, Austria, so as to be able to give customers hard facts and documentary evidence to prove their claims of exceptional reliability and product lifetimes.

The sooner a manufacturing defect is recognized, the lower its consequences and its associated costs. If a design weakness or an inadequate component quality is discovered in the pre-production stage, it is usually a minor problem that can be remedied through a redesign or a simple alteration in the bill of materials with low impact on the overall project costs. However, if a product fails during use anywhere in the world and the end customer must discard the entire production of a high quality product on the basis of a defect in a single part, then the problem can be serious. Then, the cost to the manufacturer of the defective part ranges from a free-of-charge exchange to paying the customer compensation for the entire cancelled production and for loss of business reputation.

How can this nightmare scenario be avoided? Surely, there is no way of knowing in advance the life expectancy of a new product?

However, this possibility exists! A simple method used in the past and still often used today is the Burn-In test. The electronic products are run at full load for several hours at elevated ambient temperatures to weed out unreliable products before they are delivered to the customer. But Burn-In only cannot look into the long-term future reliability of the product as this test finds only the premature failures due to defective materials or unstable production processes.
Therefore, RECOM have invested in HALT (Highly Accelerated Life testing) and ESS (Elevated Stress Screening) methods in order to can cast a more profound glance into the future of its products.

HALT - testing

The difference between Burn-In and HALT testing can be best described as Burn-In should preferably result in no failures or defective products but that the goal of HALT testing is to generate failed products. The consequent FMA (Failure Mode Analysis) will then be carried out to discover the reason why the product failed and a CA (Corrective Action) process can then be carried out to improve the product so that such failures no longer appear in the future. The HALT-tests are then repeated until no more failures appear under the defined condition limits. At the start of a HALT analysis, the minimum and maximum operational limits and the minimum and maximum destruction limits are explored under miscellaneous external stress factors. Stress factors such as very low and very high temperatures, fast temperature changes, random vibration over 6 axes and up to 50 G acceleration, stepped load changes on the DC/DC converter outputs and rapid power switch-on and switch-off cycles can be used. These tests are carried out within previously determined limits, but can be used singly or combined to increase the stress on the product.

The goal of these tests is to discover if any design, process or material weaknesses could cause the product to fail before the planned lifetime. The subsequent product improvements would then block these failures before they occurred.
The benefits resulting from these costly and time-consuming tests can be found in the reduced frequency of product defects in the field, the consequent reduction in guarantee costs and the improvement of the image of the product.
HALT - testing is not a standardized procedure as each manufacturer must formulate their own in-house standards in order to win the in-depth knowledge over their own products. In addition, products always differ in construction and material composition and the interactions can result in different failure modes which require different tests protocols.

	Picture 1: QualMark Typhoon 1.5 Test Chamber	For the stress tests, RECOM have chosen a QualMark Typhoon 1.5 Test Chamber, with which the products can be either single factor or multiple-factor stressed. Technical data:
		Temperature Range Rate of Temperature Change Vibration		-100 to +200 °C >60K/min 50 G, 6 DOF
	Since RECOM products are exposed in use to a wide variety of environmental factors such as temperature, moisture, vibration, etc., appropriate tests are also necessary in these areas. The CTS climatically controlled chamber permits such tests as low-temperature start-up at temperatures below -40°C , long term soak tests at temperatures exceeding +85°C and high humidity (95% rHs) or low humidity, and temperature cycling (-55 °C to +125 °C). Technical data:		Picture 2: CTS climatically controlled chamber
	Temperature Range Rate of Temperature Change Vibration	-70°C to +180°C 3.5 K/min 10% to 95% RH

RECOM DC/DC converters are usually supplied as fully potted or injection moulded products. The moulded case and cured epoxy is not easy to remove in order to carry out failure analysis. If the material is removed chemically or mechanically further damage that has nothing to do with the original defect can be inflicted, thus destroying the very evidence the analysis is trying to uncover. Therefore, an x-ray inspection machine of the type Phoenix|x-ray pcb|analyser 160 has been purchased to permit non-destructive product analysis. The resolution is high enough to allow microscopic examination of even the internal structure of ICs or transistors within the converter without disturbing the case.

Picture 3: Phoenix/X-ray pcba|analyser	Technical data:
	Enlargement Resolution Max Acceleration Voltage Maximum Beam Current Sample table adjustment Radiation protected cubicle	x12800 Max 1 µms 160 kV 500 µA 3 DOF