Choosing RJ45 with magnetics inside

Question

I know that on the wiki page you recommended some RJ45 transformers. Let’s take, for example, the first one, from Cetus, J1B121ZCCD. It has the following internal schematic. Notice the beginning of the internal windings. We are using the following transformer from Amphenol: LMJ1998824110DL1T39J. Also notice the beginning of the internal windings. The Amphenol RJ45 transformer works, but notice, that even if the TX section is on the same pin numbers (1, 2 and 3), the beginning of the windings, on your recommended transformer is on pin 1 and on our transformer is on pin3. Question is: being in the RF zone (100-125Mhz) is this aspect important? Or can be used any transformer, no matter of the beginning of the windings? Or should TX+ be kept at the beginning of the windings and TX- to the end, meaning in our case to switch W5500 TX+ and TX- on the transformer input? I want to mention here, that for the sake of experiment, we put TX+ and TX- directly and then reversed on the RJ45 transformer input and worked either way. But for the sake of conformity, please tell us how it’s better? Thank you!

alexpescaru · Answer

I hate to push the issue, but someone from the technical department can answer this? We are at the stage of designing the PCBs and we really need the answer to know how to proceed. Many thanks in advance!

alexpescaru · Answer

I’m sorry to find out that after almost a week, even if the Wiznet’s people/support visited and answered to other threads this one remains unanswered. Even if I kindly asked a response because we have thousands of equipments on the field that need to be updated and time is of the essence. And knowing all in the PCB’s design stage is critical. Maybe it’s a stupid question but still… Maybe we made a wrong choice when we choose the Wiznet products… Please, if it’s hard to answer with words, tick one of the below: It’s a stupid question Don’t want to answer Don’t care to answer Don’t know to answer … The same with the “socket.c bugs” thread… For months the users/customers found out some bugs on the sources and told you. How hard and how much time did something like updating some source code could take, when the mistakes are clearly pointed out? The community found out for you some problems, at least give something back to the community. A product without good support is no product at all. We are working with other big Korean companies since 1997 and they were and are great from support point of view. Sorry to see that this is not the case… Sorry, harsh words, but this is how this customer feels when he’s let down by the support team. Thank you!

jeongsk · Answer

I am really sorry. I am W5500 hardware the person in charge. So I have to answer, but I couldn’t do it. I’ve been missing in your question to continue. I’m a big fault. Sorry about the late response. The start point of the Transformer windings can be used without care. A and B are being connected with a LAN connector and you have data sheets related to check it out. And the question for the RF area makes no sense. Explain in detail? Once again I apologize for being late with a question. Thanks.

alexpescaru · Answer

Thank you for the answer. I read it between the lines… Regarding RF area: the pulses on the PHY output are at a frequency of 100-125 MHz for 100Mbps, right? Question was: at that frequency, it’s important the start point of transformer windings or not? For understanding better, which is the hardware configuration of PHY output? Push-pull or open collector?