Issue link: https://iconnect007.uberflip.com/i/461203
40 The PCB Design Magazine • February 2015 being built and measured. Now, can somebody please tell me what controlled impedance is? Bob Tarzwell, can you help me out here? BOB: Sure, Dan. And to make this really use- ful (I'm never sure if you really understand my answers to your eternal questions anyway), I'll direct my answer to those guys who really care, the PCB designers. As a designer, your project may require a specific imped- ance of, say, 52 ohms, plus or minus 7% percent. The big question is: Does the fabrica- tor give you what you ask for? Well, maybe, and maybe not. Here is why. The standardized coupon the PCB manufacturer inserts into the panel is designed to reproduce the same im- pedance effects your circuit should see, but because it is not inside the circuit it's only a close approximation test. When PCB manufacturers set up the panel, of course they use as much of the panel as possible, which means the impedance coupon will be on the outside of the produc - tion 18x24" panel. The ques- tion would be, is there any difference between the cen- ter of the circuit board and the very outside where the impedance coupon is located? Unfortunately many times the answer is yes. The edge of the PCB does have slightly dif- ferent impedance characteristics than the cen- ter; the edge can have less resin than the center due to resin flowing out at the edge of the lami- nation book. The resin/glass percentage across the entire panel determines the impedance; the more resin, the higher the impedance. Why do the edges have different resin/glass percentages? As the layup book in the hydrau- lic press heats up, the different layers do not al- ways heat at the same rate—the outside panels will heat quicker than the inside ones. The rate of change in the temperature through a critical point near the Tg of the resin system will af- fect how much resin will flow out of the pack- age. Typically we can see a wide range of resin squeeze out along the edge of the lamination package, from a small dark brown flow indicat- ing very little resin loss, to a larger white flow with a lot of air bubbles equating to a lot of resin movement out of the lamination package. The problems associated with the lamination cycle and differences between measur- ing methods and machines can mean four different test coupons on four different ma- chines, which can result in 16 different measurements. Which one is right? Well, per- haps all of them; impedance is an elusive, theoretical number, and the standard PCB coupon's impedance does not always represent the exact impedance inside the shipped PCB. The PCB manufacturer also has a few disadvantages to work around to achieve the 5 or 7% impedance measure- ments you require. The pre- preg used to laminate cores together only comes in a few various thicknesses, and each thickness has a different Dk due to its different resin con- tent, and it also has relatively wide variations in that resin/ glass percentage. Also, the cores they buy direct from the manufacturer will have a slightly different Dk than the same prepreg if they tried to copy the core directly. The core laminate manufacturers use fresher prepreg and much larger lamination presses than the PCB manufacturer. To add to the confusion, the impedance calculators are not always correct; all impedance calculating equations are approximations. Each PCB man- ufacturer will add or subtract a bit (5–10%) to account for their particular lamination press COnTROLLED IMPEDAnCE: A REAL-WORLD LOOK AT THE PCB SIDE continues bob and me the problems associated with the lamination cycle and differences between measuring methods and machines can mean four different test coupons on four different machines, which can result in 16 different measurements. Which one is right? Well, perhaps all of them; impedance is an elusive, theoretical number, and the standard PCB coupon's impedance does not always represent the exact impedance inside the shipped PCB. " "