Issue link: https://iconnect007.uberflip.com/i/1530610
DECEMBER 2024 I PCB007 MAGAZINE 19 is current does not flow through the measur- ing bridge itself. is bridge can also be used to measure resistors of the more conventional two-terminal design. e bridge potential con- nections are merely connected as close to the resistor terminals as possible. Any measure- ment will then exclude all circuit resistance not within the two potential connections 1 . Now that we understand the Kelvin bridge, the basics are that using the 4-wire measure- ment we can eliminate the parasitic resistance that is introduced through the leads and con- tact. What is le is the true reading. Utilizing 4-Wire Kelvin (Method 1) e 4-Wire Kelvin test is done by measuring a series of samples on one board or panel and creating a "master" table of values. Once the "master" values have been obtained, subse- quent boards or panels are tested against the master values. Any sample (plated hole or via) that responds with a value outside the mas- ter values' tolerances will be flagged as a fault. For the Kelvin test to be as accurate as possible, the probes should be as close as possible to the target via or microvia that is being tested. Optimally, the probes should be directly probing the two sides of the via directly. is is because as you increase the copper, the resistance total will increase and it will take a larger deviation in the resistance to trigger a fault condition. e testing of complete traces can be done but remember, we are measuring in the milliohm ranges. Depending on what the tolerance values are, it may take a large deviation to trigger the fault alarm. Although this method is widely accepted, there are drawbacks. In this method the resis- tance values are determined by the sample UUT. As with any self-learn process, you are captive to the lot presented. As we know, there can be variations in plating from lot to lot. e results of a saved program from this method may present errors on subsequent lots just due to the minor resistance changes in the plated barrels or bonds. is does not mean that the test is invalid. It just means that when using this method, you should develop master tables for each lot you process. As I stated previously, we are testing in the milliohm ranges and small changes in resistance can cause large swings in the value when compared to a previous lot. It doesn't necessarily mean the new lot is bad. It should just show that, perhaps, there is a bit larger or smaller baseline of copper in the barrel or bond. Utilizing 4-Wire Kelvin (Method 2) In this method, we do not rely on the sig- nature presented by the PCB or panel. As we learned in Method 1, we can learn the values of the sample, save them to a master table, and test subsequent product. It is worth noting that in Method 1, if there is an adjustment necessary to the master values, this can be changed in the system editor. Now, in Method 2 we determine what the resistive values will be when the PCB or panel is presented. is removes the self- learning process and provides a robust guide- line to control plating processes at the manu- facturing plant. To calculate the copper resistance in a plated hole, use the formula: R = (ρ * L) / A where "R" is the resistance, "ρ" is the resistivity of copper (typically around 1.7 x 10 -8 ohm-meters), "L" is the length of the plated hole (i.e., the PCB thickness), and "A" is the cross-sectional area of the plated copper within the hole (calcu- lated as π * (diameter/2) 2 ). Figure 2: Kelvin leads.