In-Circuit-Testing (ICT) can be defined as a PCB validation system in order to verify correct assembly of components, by presence or value, shorts and/or open tracks, ensuring the correct manufacture of the board. The main challenges of in-circuit test systems are the speed, cost effectiveness, accuracy, repeatability and parallelism of both Devices Under Test (DUT) and components on boards.
"The main challenges of in-circuit test systems are the speed, cost effectiveness, accuracy, repeatability and parallelism of both the Devices Under Test (DUT) and the components on the boards."
Key Componentes in In-Circuit
The minimum system for carrying out the ICT tests consists of: Fixture, measurement system and Switching systems.
Fixture
Fixtures are devices used to consistently test DUTs, in the case of ICT they are composed of a bed of needles that will generate contact with the test points of a PCB, allowing measurements to be carried out. Hence the name "bed of nails tester".
The working principle can be explained in a very simple way: when the cover is lowered, the system is locked and, generally, the fixture is commanded to lower the needles, a process called "engage".
The electrical contacts made by the test points are connected to a measurement system, which can be an analog acquisition board or a DMM - Digital Multimeter. In general, the system will validate the assembly of the components, measuring the nominal values (within a certain tolerance) or by presence in cases where the mesh does not allow the measurement to be carried out. Not infrequently, some ICTs, after measuring the components passively, that is, with the board not powered, perform functional measurements, with the board on, especially in cases where the cost of encapsulating the part is high and detecting a functional failure only after assembly can be very costly.
Test Points
Test points are "locations" within the PCBs that are used to allow "electrical access" to the board so that signals can be measured and/or injected into the board as shown in the images below.
Using Test Points
After the "engage" process is performed and the needles come into contact with the PCBs, test needles will be used in the following format:
The principle is the same as a multimeter, the system will directly measure the resistance between Test point high and test point low. However, in real PCBs the measurement will not be so easy to simply perform the measurement by pointing to two points due to the influence of other components on the board, creating the need to use complementary strategies, such as Test Guards.
The guards will generate electrical current at the peripheral terminals to the VTs selected for measurement.
Below you can see a real example with a DMM board and a DMM-Guard from NI (formerly National Instruments):
In the image above times R1 being measured and the guard system "blocking the leakage" of current through R2 and R3.
Analyzing another scenario, to perform the measurement of capacitor C1 we would use TP2 and TP4, however, resistors R4 and R2 will directly interfere with the measurement. Therefore, we use TP3 and TP1 as guards, that is, we apply a current at these points so that the current from capacitor C1 does not "drip" to the other components.
The same principle would apply to perform the measurement of other components in any such system.
Now, if we observe the resistors R1 and R2, even using the guards, we will not be able to prevent the "leakage" of current, in these cases how to proceed?
There is a relatively simple alternative: take the component out and observe the measurement and check the effect on the measurement compared to the measurement before the component is taken out, what happens between TP2 and TP3 with TP1 on guard when R2 is taken out of the system? Let's assume two scenarios:
Scenario 1:
Let's assume that the measurement before the removal of R2 was 2kOhms and after the removal and measurement it became 1kOhms, this means that we can attest to the presence of R2 by value.
Scenaro 2:
After R2 is removed, the system measures infinite resistance. This means that we will need to study the behavior of the PCB with R2 and without R1. Probably in a scenario like this, the lack of R1 or R2 would indicate a failure, so we can consider measuring it together, in a test step that evaluates the presence of both components.
Measurement and Switching
Practically no measurement system will have sufficient channel density to measure dozens of components, so the solution is to use matrix systems:
Relay arrays, as the name suggests, are a set of relays organized in such a way as to allow you to switch signals between your PCB and your measurement system, for example, your DMM or your analog board. The channel density will be defined by the number of lines and columns of a board, and the types of relays used have a direct influence on the opening and closing speed of the signals, that is, the switching between them, directly intervening in the agility of the measurements. of your system.
In the image below we can see the topology of a relay matrix:
Golden Standard e Measurement Centralization
Given the fact that in a real system the measurements of the components will hardly be the nominal values, during the process of developing the test specifications a board previously validated in a laboratory where the assembly and operation have been validated in an empirical way will be the "gold standard". ", that is, the pattern that the other pieces need to follow. For example, given a resistor of nominal value 1kOhms with 10% tolerance that had an empirical measurement of 700Ohms when mounted on the PCB, what value should be considered for the test?
For scenarios like this the Offset strategy is used. Once it is known, empirically, that the assembly is correct in the post processing of the measurement, it can be centered by adding or subtracting the missing value to center the measurement. Examples:
Nominal value 10kOhms with an empirically measured value on the assembled PCB of 10.5kOhms, we will have an offset of 0.5kOhms
Nominal value of 100kOhms with an empirically measured value on the assembled PCB of 102kOhms, we will have an offset of -2kOhms.
Rated value of 500Ohms with an empirically measured value on the assembled PCB of 380kOhms, we will have an offset of 120Ohms.
However, the use of offsets does not guarantee repeatability of measurements, for example, the part of example 1 above in a DUT can measure 10.05kOhms and in another DUT it can measure 9.01kOhms, considering that even if both are within the tolerance for quality parameters these measurements are far from accurate, despite being within tolerance. This concept can be seen in the image below:
And a good test system needs to be precise and accurate. One way to centralize these measurements is by repeating the tests, that is, when performing a single component instead of measuring it just once, you can perform several measurements and perform an average so that the tester be precise and accurate.
Let's consider a tester measuring a 10nF capacitor. The test TPs and guards were correctly selected and the offset properly calculated. Now consider that to ensure accuracy we will use 5 measurements:
Measure 1 | Measure 2 | Measure 3 | Measure 4 | Measure 5 |
9,9nF | 10,2nF | 10,1nF | 9,8nF | 10,1nF |
The result to be analyzed by the test system will be the average value of these 5 measurements above, that is, 10.02nF, thus making the measurement more accurate and more precise.
Another important factor that will interfere with the measurement is time, as when using a bench multimeter to measure a component it takes time for it to stabilize, that is, the same applies to ICT type testers with typical time around 5-10ms per repeat measurement. The direction of the High and Low TPs influence the stability time. Therefore, during the validation of the gold standard of a PCB board, it is necessary to establish:
The directions of the High and Low TPs, and the position of the TP Guards
Offset - if necessary
Repetition and time between repetitions
To achieve accuracy, precision and repeatability all of the above factors need to be combined at their optimum and not necessarily maximized whenever possible. In the next chapters of our series on automated testing we will address the quality analysis factors, the MSAs.
Did you know that Blue Eyes Systems offers a universal test system for In-Circuit systems? Check the link below:
References:
[01] L. S. Milor, "A tutorial introduction to research on analog and mixed-signal circuit testing," in IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing, vol. 45, no. 10, pp. 1389-1407, Oct. 1998, doi: 10.1109/82.728852.
[02] Bateson, John. In-Circuit Testing. 1.ed. Springer, 1985
[03] National Instruments Corporation, Developing Test Programs with TestStandᵀᴹ Course Manual, Austin, Texas, 2017.
Comments