Adaptive Measuring

Using the Best Measurment Procedures

written by:
Manfred Propkop

atg® Test Systems

Throughput is the central problem for flying probe test systems.

Generally, increasing test speed represents the greatest challenge for this technology. Test speed can be influenced in various ways. For example, the number of test heads can be increased. The resulting high costs and complicated motion control (i.e., anticollision) do not justify the attainable speed increase. Other possibilities for increasing the speed of the flying probe test system have been investigated. All investigations have led to one conclusion: If you are using a flying probe (direct contact-no other procedure is currently practiced), minimizing the number of measurements (contacts) brings the greatest throughput increase.

Let us examine the currently practiced methods of measuring continuity.

Trying to solve the problem of flying probe throughput by combining the advantages of several test methods.

"The adaptive method of measurement currently represents the fastest measuring procedure."

Ohmic Measurement

The first measurement we will look at is ohmic measurement with current flow. The ohmic measurement is the most precise, simplest, and best-mastered method of measurement. It requires a contact at the end of a track in each case. The resistance measurement can be carried out in a range from Mohms to high Gohms.

There are certain disadvantages. Two simultaneous contacts are always necessary for a continuity measurement. Consequently, the slowest contact (normally the one with the longest way) determines the measuring time. Additionally, collision-free motion must be guaranteed with a flying probe test system.

Capacitive Measurement

In the capacitive continuity test, the capacity is measured at every pad of a net. These measurements can occur at different dates. These values are stored, and if they differ at different points of a calculated net, an open must be assumed.

For the measurement, only one simultaneous contact is necessary in each case. This means that the motion of the individual test probes requires considerably shorter distances. Additionally, the capacity values for every net can be analyzed for the shorts test, saving additional contacting. Also, an open in large nets (large capacity) is recognized if the disconnected conductor track is measured with less capacity. This especially applies to relatively large capacities (i.e., in power/ground nets).

Since accuracy and the measuring tolerance determine the resolution and error recognition, two-pad nets with very small capacities cannot be checked reliably with this test. The ohmic value of a contact can only be determined in a very limited manner. Parallel and secondary capacities falsify the measuring results.

Shorts Test

For the shorts test, three methods of measurement were adopted: ohmic measurement with contact minimization through adjacency, capacity comparison, and electrical field measurement.

1. Ohmic Measurement with Adjacency

It is virtually impossible to check all nets of a PCB against every other net. In order to minimize measurements, a computer analyzes conductor routing on the PCB. Every conductor track is examined along with any track with which it could have a short. Distance and possible shielding are the deciding criteria. A track running between two other tracks is regarded as shielding. Consequently, only nets practically capable of presenting shorts between each other are crossmeasured. This is an ohmic measurement.

Normally, however, one net requires several measurements.

This means repeated pad contact-particularly if a qualitative test statement is desired. Repeated pad contact substantially increases the total test time.

2. Capacity Comparison

The capacity of every net is determined, and all nets are compared with each other in the computer. Two nets of a PCB with the "same" capacity are considered short-circuited. Since the capacity value could already be determined in the continuity test, no further measurement is necessary-considerably reducing test time. This is a reliable test statement, especially for networks with large capacities.

There are limitations to capacity comparison. A good board (golden board) is always required for the computer learning procedure. This procedure is considerably slower than other test options. Limiting factors include the measuring accuracy and measuring tolerances of the test system. With very small test pads and nets, the measuring statement becomes much less reliable. In turn, considerable remeasurements are required, though these might be reduced with adjacency.

Also, the ohmic value of a contact can only be determined in a very limited manner. The accuracy depends on the capacity value. Parallel and secondary capacities falsify the test results.

3. Electrical Field Measurement

The electrical field measurement can be practically compared with a multidimensional measurement. The accuracy of the measurement results, therefore, is considerably better when compared to capacity measurement. The number of so-called retests is drastically reduced, and the insulation value can be determined up into the Mohm-range.

Still, there are concerns. Signals must always be sent for the electrical field measurement. Two to three probes achieve this. However, only one test probe can measure at a time; signals cannot be sent simultaneously. This utilization of test probes for sending signals prevents them from being used for point scanning-strongly reducing the test speed.

The Adaptive Measuring Procedure

In this patented procedure, the previously described advantages of all measuring procedures are applied. Since the application of each procedure is strongly dependent on the board, the test system automatically decides on the appropriate procedure for each net after the first test run. The parameters, preset by the operator, are also taken into account. Low ohmic testing (i.e., greater than twenty Ohms) in the continuity test greatly reduces the number of nets to be investigated or opens via capacity measuring.

For a measurement of ten Ohms, a predetermined minimum capacity is necessary. Normally, only power/ground layers reach such values. In addition, the number of retests can still be decreased through the application of adjacency.

If throughput is the main criterion, the highest possible values are chosen for the continuity test and lowest values for the shorts test. The optimally correct value is determined by the test system itself (adaptively). The operator only presets his parameters.

If the capacity value of a net is less than the measurement error of the test system, ohmic testing and field measurement is carried out. For values greater than the measurement error of the test system, the capacitive measurement is used for checking continuity and shorts.

The adaptive method of measurement currently represents the fastest measuring procedure. The measurements adapt to the board and the test specifications, and are considered optimal in each case. An increase in the throughput by up to 60 percent compared to the current values is to be expected.