The probe must have an impact on the impedance test?

yvonneGan

The probe must have an impact on the impedance test? [Copy link]

Author: Zhou Wei, a member of Yibo Technology Expressway Media

Why can't the impedance test range be the entire line segment but the 30% (50%) ~ 70% segment? Last time, Lei Bao was too embarrassed to ask the master directly about this question, but he did not stop looking for the answer. Since it was not easy to ask the master, he could ask his senior brother Fei Ge next to him! So he shamelessly asked his senior brother again.

Senior brother Fei Ge was straightforward and began to patiently popularize the basic knowledge of impedance testing to Lei Bao: "Xiao Lei, I have asked the master this question before, and now I still have some impression. It's a good opportunity to tell you about it and strengthen it. If there are some omissions, you can check the information later. Impedance testing is generally a product verification indicator used by PCB manufacturers when leaving the factory. Because it involves multiple batches of data, it is more critical for the repeatability and accuracy of the results. Because it is a curve reflected back according to time and length, it is impossible to fix the reading at a certain point, so the result depends on the value range. Generally, the impedance in this value range is required to be as stable as possible, so as to reflect the true impedance value of the line. Now you know why it can't be Look at the result of the entire curve segment. "

Lei Bao seemed to have realized something and answered hurriedly: "The place where the impedance range is stable should be the middle section of the line, because the front end is the contact between the probe and the board. If the board is a coupon strip, the contact point is the via. If it is a line in the test board, then the general contact point is the pad, etc. This contact point will inevitably cause impedance discontinuity, and the end is generally an open circuit. We all know that the open circuit impedance is infinite, so the impedance curve we see is approximately vertically upward. This open circuit impedance will also have an upward effect on the impedance of the end line. In order to make the test results closer to reality, it is more scientific to take the impedance according to the 30% (50%) ~ 70% segment of the line. Is that right? "

"Yes, you actually mentioned a very critical influence just now, that is the probe. Different probes have a great influence on the front-end impedance, which may also affect the impedance of the entire line segment. Let's discuss the differences in the results tested by different probes. "Senior brother Fei Ge added, silently drank a sip of water to moisten his throat, and continued to popularize it to Lei Bao.
"Our conventional probes are usually handheld probes. This type of probe is inductive. The longer the interface ground sheet (the farther the ground pin is from the signal pin), the greater the inductance, and the higher the impedance when in contact with the object to be tested. The following are some common handheld probes.

Figure 1. Common handheld probes

The results of these probe tests are easily affected by the loop length from the ground pin to the signal pin (also called the ground loop), as shown in Figure 2 below.

Figure 2. Effect of probe ground loop

Take the single-ended probe in our hand to test the impedance. When the probe is placed in different positions, we can see that the impedance difference at the front end is relatively large as shown in Figure 3 below. The influence of the probe lasts to the 30% range, and it will also have some impact on the entire impedance. Therefore, it makes sense to take the range where the impedance fluctuation is relatively stable when taking the value. If the ground loop of the probe is larger, it will affect the range exceeding 30%. Therefore, the 50%~70% value range came out later, which is still used by most board manufacturers.

Figure 3. Impedance test results of actual handheld probe

In addition to handheld inductive probes, many clients use SMA probes for direct measurement. Our commonly used SMA probes are shown in Figure 4 below.

Figure 4. Common high-frequency SMA connectors

The above are all high-frequency SMA heads, which can be used in different occasions. SMA heads No. 1 and 2 need to be welded, and the latter two are fixed with screws and can be reused many times. We all know that welding requires tin. The performance of welding by different people or processes is also different. The amount of tin retained will eventually lead to differences in impedance. Generally, most SMA probes No. 1 are capacitive after welding, and the latter ones show different impedance values depending on the optimization during design. We once had a customer who used this SMA head No. 1, which caused a large deviation between the test impedance and the impedance when the board was shipped from the factory. At one time, we suspected that there was a problem with the board manufacturing. Please see Figure 5 below. The impedance of the front end is so low that it makes people doubt their lives.

Figure 5. SMA head test results

The test results of the same PCB factory using a handheld probe are shown in Figure 6 below.

Figure 6. Handheld probe test results

Of course, whether it is an SMA probe or a handheld probe, direct point measurement has a relatively large impact on the impedance of the front end, so there are also micro probes with relatively high precision on the market. This type of probe can greatly reduce the impact of inductance and capacitance. The disadvantage is that the test points need to have a certain spacing requirement (fixed Pitch). Since the probe is very small, some even need to be operated under a microscope, which is relatively inconvenient to use. Another major disadvantage is that the price is very expensive. Bosses who are not wealthy should quietly avoid it. Anyway, I haven’t used it yet, but I have to say that the test results of the following probe are still relatively ideal compared with the results of the SMA probe, as shown in Figure 7 below.

Figure 7. Microprobe test results

Even without this kind of micro-probe, it will not affect our good design and optimization of impedance, especially now that we use a detachable high-frequency SMA head. After many test board optimizations, our current high-frequency SMA probe can completely match the impedance at the contact point with the board. Figure 8 below shows the impedance of the fixture board that we tested with SMA No. 3 and No. 4 in Figure 4. It can be seen that there is basically no fluctuation in the impedance at the front-end probe. This is the most critical technology for TRL de-embedding test boards or fixture boards, and it is also a feature that directly reflects the performance of the fixture, because most high-speed signal fixture interfaces now use high-frequency SMA heads, and the impedance of the contact point between this SMA head and the PCB board (also called Launch impedance) is the most critical factor. Often most of our work is to optimize the impedance of this contact point. "

Figure 8. Optimized SMA head impedance test results

I didn’t expect that the impedance test, which seemed so simple before, turned out to be so complicated. Lei Bao was deeply impressed by the explanation just given by his senior brother and fell into deep thought. It seems that he still has a lot to learn in the future. Keep up the hard work, young man!

Jacktang

The welding performance of different people or processes is also different, and the amount of tin retained ultimately leads to differences in impedance.

How to solve this problem?

There is no consistent reference.