According to the normal thinking logic, the line layer of the high-speed signal is generally 0.5oz or 1oz. If you see a high-speed signal on the thick copper, will you be very surprised?

Why do high speed signals usually go on the 0.5oz signal layer? Regardless of performance requirements, from the point of view of processing, 0.5 oz of etching layer to go line, PP plastic flow is relatively stable, and for the design, the distance between the lines to walk and walk also won’t too strict, so in general processing of impedance can be stable, also too won’t big. But what happens if the wire, for whatever reason, goes on top of a thick layer of copper? Today we will share with you such a “one in a hundred” case.

This is a customer designed test fixture plate for high speed connectors. As the name suggests, it is used for high speed connector testing. The customer’s target rate for the self-developed connector is 25Gbps, so the fixture performance is required to support at least 20GHz. In fact, the design, processing and crimping of this project are all completed by the customer. Originally, Our Mr. High-speed could not see this case, but the performance of this fixture is very problematic, so the customer came to our Mr. High-speed for help, hoping to help locate the problem. Below is the test fixture board designed by the customer. We see that half of the signal routing is on the surface, connected to the high speed connector through the SMA connector.

The principle and design were simple, but problems began immediately after the customer returned to the board for testing. That is the impedance of the surface line is incredibly low!! The customer conducted an impedance test and found that the single-end cable of the surface layer was generally low, even some were only about 42 ohms!



This puzzled the customer a lot. In principle, the link is relatively simple, and there is no crosstalk in the distant adjacent positions, and the line width of the line is not very thin. The 6mil line width is easy to process. After this case was submitted to Mr. Gao Hua, we also conducted impedance test and found that the impedance of the surface line was really as low as the customer’s test! At this time, we calculated the impedance according to the customer’s design of laminating and wiring, and found that according to the design of laminating and wiring width, the impedance could indeed be calculated to 50 ohms.



So what’s wrong with the line? It’s definitely not visible to the naked eye. Fortunately, our company has its own board factory, so we can think of the method is to let the board factory to slice, we see the actual structure of the line through the microscope!

After a lot of tossing and turning, we finally finished slicing up the test board, and sure enough, here we found the answer. We made cross-sectional measurements of these surface lines and found that the copper thickness of the surface lines was far beyond our imagination, it was 3.6mil!!



We are plating the surface line with 0.5oz base copper (as we often call it with 0.5oz+plating) and the copper thickness is usually around 1.6mil when it is plated. But this board customer board factory actually live made 3.6mil, Mr. High-speed had not witnessed the test results, it is unbelievable!

What’s more, the line width was changed from the designed 6.5mil to 9mil. If it wasn’t for the thickness of the plate and medium, it would still be right. We all think that the board factory took this project as another project to do.

Then we calculate the impedance again based on the measured data of the slice, and the calculated results are very close to the actual impedance test results.



In this case, it is actually difficult to understand how the plate factory the customer found processed. We preliminarily judged that the plate factory used THE BASE copper of 2OZ for electroplating to obtain the thickness of 3.6mil copper through the huge difference in copper thickness. In addition, we can see that the wiring was also etched very unevenly. Anyway, this case once again proves that there are many uncertainties in the processing of microstrip lines, and its impedance control is certainly more difficult than that of strip lines. Therefore, we should pay special attention when using microstrip lines to design projects with high speed and high precision.