In the teachings of many signal integrity theories we have accepted, there is a principle on crosstalk to pull the strip line to open for 7h, and the microstrip line opens 7h, which is said to be able to avoid more than 90% of crosstalk. Today, Mr. Expressway will verify the theory of microstrip line 7h, see if it is really so stable!
We started studying this new test board. In the case of the package and the unbumped, we have the same impedance of the two single-ended lines (red), the GAP distance of the two lines is 7 h, and h will not need to say more to the high speed. Then before, we shorten the distance of the privately foramen from 400mil in the video to 50mil, in which case let's take a look at the performance of the package and the unbacked performance.
50miL a via hole is actually a very good design, according to the description of the video, in the case of 400 mil via distance, our package effect will present a refrigerated effect in some frequency bands. In fact, this is due to the theory of 4 sizes of wavelength impedance transformation. Then theoretically we make the ground aperture more dense, will it push the refrigerated frequency band to 50GHz? Let's first look at the distal crosstalk results of the package, as shown below, does not have a fluctuation point of the anti-resonant, indicating that 50 mIL can guarantee the anti-resonant point of the impedance transformation in 50GHz, this is complementary The relationship between the via distance and the anti-harmonic point is actually calculated. Mr. there is room for empty high speed can be expanded in detail, here is not over, we still return to the topic.
So now you must know, if just pull the crosstalk that is not packaged, is it? Ok, this high-speed Mr. does not sell Guanzi, is directly compared.
It can be seen that if the flying hole design is good, the effect of the package comes out, and the crosstalk of the package is better than that of the 50GHz band. The article is here, which has already understood a very good package, which is very close, in the same spacing, and there is still a significant benefit. Mr. Expressway feels that these two words can be "obvious", you can not believe, it is not a few DBs of the far-end crosstalk, why is it very obvious? Mr. Express decided to continue.
Cross disturbance is just one of its appearances, and our signal transmission is actually the most concerned or insertion loss, that is, how much energy can go to the receiving end. From this indicator, everyone may feel very powerful, as shown below:
In the case of not incorporated, the insertion loss is obvious after 10 GHz, and when the last goes to 50GHz, the loss is almost 50% more than the package. In fact, it is also very easy to understand, because the energy is conservation, the energy of crosstalk is much, in the case where the impedance is consistent, the energy of the reflected back source is basically equal, that is, insertion loss will definitely becomerable. To put it bluntly, the energy lost, I went to the crowd.
According to this theory, we will compare the following things, everyone will understand, if only single line, how will the insertion loss in the case of crosstalk?
Not much to say, directly on the results!
It can be seen that the results of the loss will become further in the case where the single line is not crossted. This more indicates that the impact of crosstalk on loss is very large. In other words, crosstalk consumption is still very large, especially when crosstalk is in -15dB, basically insertion loss will reduce one. Very large proportion.
The amount of information in this period is still relatively large, so Mr. Express is finally a summary of this article, let everyone know more clearly, this article wants to express.
1. For single-ended microstrip lines, 7h is actually unstable in the case of not being packaged. If the end of the inch length is in the INCH level, the influence of crosstalk for loss is still very large;
2. A well-designed package is actually able to improve crosstalk, there will be positive help;
3. Whether it is packaged or not packaged, the loss will deteriorate due to crosstalk, and the loss is proportional to the size of crosstalk.
4. The case where the microprozel like a single-ended microstrip can go to the high frequency band, the first thought is that the radiofrequency trace is like a millimeter wave of the current highest 77GHz, so the package is still very very very. necessary. In terms of digital signals, the general high-speed digital signal is not single-ended, so it is not considered within this range.
Edit: hfy, read full text
Our other product:
