Hot answers tagged

4

I would go with option 1. The fundamental goal of a differential pair is that any unintended signal which couples on to one wire in the pair also couples on to the other wire in the pair, and since the receiver subtracts the two signals, they cancel. This works better the closer the two wires are coupled. Another argument is the inductive loop. For ...


3

Using exactly the same values as you did, but with another calculator (from EEWeb), I get: Which is different from what you had, and closer to the target. There are several models that can be used to estimate the differential impedance, and different tools may use different underlying computation techniques, so you can get different results. In any case, ...


3

I have just taken apart a low cost IP camera made by Dahau and it appears to be based on a SoC made by Ambarella, part number S2L-M-A1-RH. Details are here. The SoC includes just about everything you need to make an IP camera: ARM A9 processor camera front end video processor AES cryto engine H.264 video encoder Ethernet MAC Most of these IP cameras run a ...


2

This flex cable is designed for 100 ohm differential signals. To maintain that impedance you would want the + / - signals to be adjacent to each other. You could put either one or two grounds between each differential pair. Obviously two grounds will result in less crosstalk between pairs.


2

Maybe it's not a direct answer to your question, but i want to draw your attention on the following possible workarounds: skew rate is controllable both at RGMII PHY and FPGA IC Typically RGMII PHY implements a de-skewing mechanism (e.g. KSZ9021 can absorb skews up to 1.8 ns, very near to that what you need), therefore (if your phy has it, of cause) you ...


2

Short answer: Yes you can use the KSZ9031RN with only 10/100 magnetics. Read the Support datasheet and note section 3.8. So you will have to program the devices to ensure it does not attempt to size 4-pair as opposed to 2-pair connections.


1

Sure, if the width is the same, you won't need a gearbox. But most high performance serializers are going to be a power of two, and 66 is not a power of two, so you're pretty much guaranteed to need a gearbox.


1

No, that would not work. Even if you ignore completely messing up what is basically an RF link by introducing a nonlinear element, it just has not way to work. You have to monitor actual traffic to do what you want to do. An ethernet link is basically negotiated at the physical layer when the link is initiated. If the physical layer sees that the two ...


1

MDIO is not a part of GMII or SGMII. SGMII does have some autonegotiation features, but it does not encapsulate MDIO.


1

I need to route 100ohm differential pair [...] for parallel routing on one layer. I can't find any PCB calculator. That is because a one- or two-layer board without prepregs leads to ridiculous track width. 70mil and wider with 10mil spacing. (I had once seen an USB hub with such a stackup. It worked for a while, then it began destroying random devices, ...


1

1) yes, you can. 1000BASE-T (MDI) is suitable for doing this. but keep in mind that: you does not need a separate ground "under" each 1000BASE-T MDI (= a couple of four pairs matched), one plane "under" all of them may be enough. the main tip here is that plane must be tied to chassis and topologically separated from any local ground despite of such a ...


1

The importante part is that you must set the rule about matching. As your data is diferential lines you must pay attention to the matching of the lines, around 100mils. If you want a more precise information many manufacturers add the routing rules in the datasheet. Also the impedance for physical lines is 100ohms in diferential as you have it. Another ...


Only top voted, non community-wiki answers of a minimum length are eligible