Crosstalk from fixture wires
Did you know that how you layout the wires in the fixture is very important?
In order to reduce cross talk between wires in a fixture, ideally you should run the wires perpendicular to each other.
Fixture wires running in a bunch, parallel to each other, may look neat and pleasing to the eye, but it is devastating for signal quality. This is because signals running on one wire is inductively coupled to the next parallel wire. The more surface area that the two wires are in parallel, the stronger the coupling, the stronger the coupling, the more likely the signal from one wire can affect the other, this manifests as noise on the other wire.
To minimize crosstalk interference, the wires in the fixture should not be bundled together and should be wired directly from one Pin to another in the shortest path so that the wires intersect each other instead of run in parallel. This reduces the surface area of the wires that are in parallel and thus reduces the effect of crosstalk.
It is actually easier to wire from pin to pin. This small step could help you reduce lots of time debugging a noisy intermittent test. Try it.
Best VTEP sensor plate sizing
With BGAs being very commonly found on PCBA these days, it is important to know the best way to size your VTEP sensor plates that are used to test the BGAs to achieve the best results.
For ICs with leads (gull-wings, J-leads, bat-wings, etc), we generally size the VTEP sensor plate larger than the component body to overlap over the leads. This is so that the capacitance coupling covers a wider surface area (the wire-bond, internal IC traces, and leads) resulting in a larger capacitance detected. Consequently, it would be easier to determine a well defined PASS/FAIL result.
For BGAs, the layout of the balls is different from ICs with leads. The balls are found under the component body instead of at the perimeter of the component. Furthermore, the balls are much smaller in surface area than the leads. A much more sensitive measurement method like VTEP is required compared to testing leaded ICs. However, with this added sensitivity, additional care is required to remove potential “noise” sources in order to get the best test results.
For best results, VTEP sensor plate should be sized smaller than the BGA. This reduces the coupling of surrounding components and PCB traces and therefore reduces the parasitic capacitance from these interferences. The illustration shows a sensor plate that is too large for the BGA. You can see that stray capacitance is introduced from a nearby component (C Component) on the left and from the trace (C Trace) running at the right side of the BGA under test. The final measurement of this VTEP is the sum of the BGA measurement (C Measurement), the nearby component (C Component) and the trace (C Trace).
If the VTEP sensor plate is reduced in size to about 80% of the BGA size, it will be large enough to cover just the BGA balls (or even slightly smaller). This drastically reduces the stray capacitance from the nearby component (C Component) and from the trace (C Trace), leaving only what you desire to measure (C Measurement).
Try this simple adjustment to the sensor plates on your BGAs and enjoy a more consistent VTEP test for BGAs.
Strain Gauge Testing on BGAs
On today’s products, we do see a lot more BGAs used. Because of the mechanical properties of the balls, the BGAs, are less flexible than ICs with leads.
In the ICT fixture, the board under test, with the many BGAs assembled on it, will be subjected to the many different forces on the fixture when the board is forced to contact with the probes. Most fixtures employ press-rods that fit into the spaces on the PCB between the components to press the board onto the bed of nails.
Since it is impossible to place press-rods directly above probes all the time, there will be a torque created when the press-rods press down against the upward force of the probes. These forces act on the components on the board at the solder joints (See Fig1). If the force is too great, they can cause stress fractures or cracks in the solder. This would result in a reliability issue for the product.
Sometimes, these cracks or fractures close upon release of the fixture forces causing electrical contact. Thus, the ICT tester or functional tester may not detect these failures. These micro-cracks may open again during the function of the product as the solder joints or PCB and components expand, causing the product to fail in the field.
We need to address this fault by a preventive measure using strain gauges (See Fig2) to measure the force applied to the product during testing so that any excessive force is reduced by adjusting the press-rod heights or placement to reduce the torque or changing the fixture design. After the forces on the fixture are normalized, it is prudent to continue checking forces on the board under test regularly during mass production to ensure that the forces are kept within the limits. This is a “destructive” test because the strain gauges need to be glued securely on the PCBA. It is normal to reuse the same test PCBA during the regular checks.
The placement of the strain gauges on the board is very important. Strain gauges are directional, depending on the space on the board, try to check as many directions of the force (stress) as possible by placing the strain gauges in different directions. The strain gauges are normally placed along the longest axes of the components, to catch the largest forces (See Fig3). These are usually at the corners of rectangular shaped BGAs.