Column Control DTX

VTEP Goes Head-to-Head with Keysight TestJet

Case Studies

Introduction

In the early 1990s, quad-flatpacks and accessible SMT components were the norm on most printed circuit board assemblies (PCBAs). At that time, Keysight Technologies, Inc.’s TestJet was a real breakthrough for increasing coverage at in-circuit test. But today’s PCBAs are inadequately covered by capacitive probe measurement techniques, including Keysight TestJet, opening the door to Keysight Vectorless Test EP (VTEP) technology. In beta tests, VTEP proved its abilities to improve in-circuit test coverage by over 80 percent compared to Keysight TestJet, especially on boards with hard-to-test packages such as BGAs, micro-BGAs, and SMT edge connectors. This paper summarizes some of the results of early tests conducted at customer sites, and also includes results of controlled tests conducted in Keysight’s R&D lab.

Historical Package Types Accommodated by TestJet

At the introduction of Keysight TestJet in the early 1990s, most devices on PCBAs had lead frames and large enough geometries to ensure reliable ICT measurements. Most test engineers will recall that the TestJet sensor plate and amplifier board combined with a lead frame to create a capacitor that was approximately 100 fF. If the lead is open, the measurement is approximately 0, creating an easy-to-spot differential.

In production, the measured values of passing and failing pins fall into a Gaussian distribution. The spread of these measurements is characterized by their standard deviation, or “sigma.” Sigma is a function of measurement noise in this case, variations of package styles, misalignment of sensor plates, variations is TestJet mux boards and probes, and even electromagnetic noise (like fluorescent lights )in the tester environment.

The TestJet hardware and measurement algorithms at the time had roughly a standard deviation (sigma) of 1.7 fF. To get good pass/fail discrimination, you’d like a threshold set six sigma above a failing pin and six sigma below the passing pin. Setting the threshold to 20 fF was enough to guarantee a good pass/fail measurement in nearly all cases, covering most or all pins on a device. While TestJet allowed per-pin programmability, most debugging was done by setting a single threshold per device since most pins measured well above the 20 fF threshold.

In short, TestJet was ideal for its time, allowing robust measurements on the most popular package styles. It was easy to program and debug, and had only a minor impact on fixturing. It became the de facto standard in ICT, and was adopted by many users as well as suppliers of ICT equipment.

Modern Package Types Accommodated by VTEP

Times have changed. QFPs and BGAs have become a common package type and present some challenges for TestJet. This diagram illustrates the problem: everything to the left of the red line is untestable or unstable with TestJet.

Today’s BGAs are especially problematic. They vary in pitch, ball placements, and the presence of thermal plates and shielding. Plus, there are no lead frames, so the wire bonds and traces measure less than 20 fF, creating a problem for TestJet. Since sigma is 1.7 fF, TestJet cannot guarantee good discrimination for a reliable pass/fail measurement on BGAs.

Furthermore, while many pins on a BGA measure low, many others measure high. In this cross-section illustration, note the presence of copper traces outside the chip edge for some pins. This means that a single threshold is not appropriate. While TestJet supports per-pin thresholds, manually changing them for hundreds of pins in a TestJet test can take several hours. Plus, there are no statistical validation tools, so these thresholds have to be qualified by real production. The overall result is a loss of coverage and many commented pins in a TestJet test for BGAs. VTEP addresses these issues with the addition of an auto-debug feature that automatically sets thresholds on a per-pin basis.

Three beta sites participated in a study designed to compare the performance characteristics of VTEP versus TestJet on select boards and devices. Two customer sites and Keysight’s R&D lab participated in the study. Each of the beta sites identified boards that were previously untestable or difficult to test with TestJet, making them good candidates for the study.

×

Please have a salesperson contact me.

*Indicates required field

Preferred method of communication? *Required Field
Preferred method of communication? Change email?
Preferred method of communication?

By clicking the button, you are providing Keysight with your personal data. See the Keysight Privacy Statement for information on how we use this data.

Thank you.

A sales representative will contact you soon.

Column Control DTX