If you don't understand, just ask, how do you calculate the routing loss on a mixed-voltage board?

yvonneGan

If you don't understand, just ask, how do you calculate the routing loss on a mixed-voltage board? [Copy link]

Member of Mr. High Speed - Huang Gang

A very simple theory about transmission line loss is that the smaller the loss factor (commonly known as DF) of the board used, the smaller the loss of the transmission line. We usually classify boards according to DF. For example, boards with DF greater than 0.02 are generally called ordinary boards, such as ordinary FR4 boards; boards with DF around 0.01 are called low-loss boards; and boards with DF of 0.005 are called ultra-low-loss boards. There is no way, as board manufacturers are updating faster and faster, and DF has gradually dropped from 0.005 to 0.002, so we have come up with various "extreme" names, such as ultra-ultra-low-loss boards.

(, downloads: 0)

The source of this article is also quite accidental. It comes from a technical question about a specific project asked by our design department colleague Owen to Mr. Chris, the high-speed engineer. His question is the same as the abstract: if the surface routing to the reference layer passes through two media, the high-speed plate and the ordinary plate, then after finally referring to the ordinary plate on the L3 layer, does the loss become the loss of the ordinary plate?

(, downloads: 0)

I don’t know what the fans think. Owen’s worries are indeed reasonable. The surface routing refers to the L3 layer, and when it finally reaches the reference plane after L3, it finally contacts the ordinary board. In this way, will the loss be mainly presented according to the ordinary board?

(, downloads: 0)

Some people may agree with this point of view, but Chris has his own opinion, so he continues the conversation with Owen.

(, downloads: 0)

Yes, Chris's point of view is that it has an advantage over the loss of ordinary boards, but it certainly cannot reach the performance of Rogers' ultra-low loss boards. Of course, Chris didn't want to talk about it, and then promised Owen to quantify it when he had time!

(, downloads: 0)

So how does Chris verify this? It’s actually quite simple. Just compare these three cases.

(, downloads: 0)

Chris started to work immediately, so he opened the 3D modeling software to model the above three cases. The transmission line length is about 500mil, and the premise is that the line width must be the same, otherwise there will be extra loss differences caused by the line width. The first step is to quantify the loss difference between the two cases of pure ordinary FR4 board and pure Rogers board.

(, downloads: 0)

After the above two models were built, they were run out quickly and the loss results were obtained quickly. The difference is as follows! We mark the loss of a 10GHz frequency point and can see that the advantage of pure Rogers board is still very obvious, less than half of the ordinary FR4 board.

(, downloads: 0)

Then here comes the model that Owen is worried about. If the L1 to L2 layer is Rogers board, and the L2 to L3 layer is ordinary FR4 board, what will the loss of the same trace be?

Chris then built a model for the third case, as shown below:

(, downloads: 0)

This is a model of two types of plates being mixed and pressed. Let's take a look at the loss of this case. What is the loss of this case like? This time, I won't force you to add suspense. I will directly run the results and compare them with the losses of the above two cases. It's like this!

(, downloads: 0)

Wow! Chris guessed it right again. After the two types of boards are mixed, the loss of the transmission line will be between the two. The reason is very simple. The combination of the surface wiring and the L3 layer reference plane, the signal ground, is actually maintained by the electromagnetic field running inside. So the path of the electromagnetic field is through both the ordinary board area and the high-frequency board area, so the natural loss is between the two. Don't worry, the high-speed board is not used in vain, it's just a fracture! Chris didn't have time to tell Owen, and the article was sent out. I hope he can see it!