THS4520: how to use THS4520 to drive 100Mbps ethernet pulse transformer?

您好，已经收到了您的案例，调查需要些时间，感谢您的耐心等待。

Hi，

Great question!  I have not spend much time in the lab working with the THS4520 with inductive loads.  Do you have an example of the magnetics port/transformer combo or discrete pieces you would be considering?  If you do not have a part number to share, an estimate of inductance and other transformer values would be appreciated.

We do have line driver devices, such as THS6222 and OPA2675, which can drive differential outputs into heavy line loads.  You can consider these parts also, which have demonstrated application uses for powerline communication where the load includes a transformer to the line.

For factors to consider, it is important to examine if the FDA circuit (or line driver circuit) has the necessary voltage and current output capability to drive the load.  There is also consideration for stability depending on the frequency of operation and load.

I will do more research to support you; please let me know if you can share details on the 100Mbps Ethernet pulse transformer.

The transformer is HX1188NL from Pulse

Thanks!

HX1188NL.pdf

Hello,

I will take a look at our FDA and the ability to drive the selected transformer.

Thanks

Waiting for your reply.

Hello ,

We have used transformers with similar inductance (~350uH) in the past for our line driver tests & validation.  This includes our THS6222 line driver.

For the THS4520, we can extend the circuit ideas from THS6222 testing to the FDA.  The potential concern would be if there was a fast change in voltage (dV/dt for voltage rate of change) which could stress the THS4520 when driving the transformers (HX1188NL).

A design recommendation would be to use isolation resistors (Riso) at the output of the FDA to assist with driving the transformer.  Riso (or Rseries, Rs) are commonly used with FDA and line driver circuits.

The design choices include selecting your output impedance network( Riso, Rload to amplifier, transformer, etc) and understand the characteristics of the ethernet signal.  A fast voltage change could stress the FDA beyond its ability to drive the output signal cleanly (without distortion).

Ideally you want to have the amplifier, THS4520, see an optimal load from its output, while the transformer sees a matched load to the transformer impedance.  One way to do this is use Riso resistors in series on the outputs of the FDA.  You would then need a pi-attenuator network to tune the impedance seen from the amplifier and seen from the transformer into the same output network.  A simple way to think about it: add some passive output components to allow for tuning the output loading to be reasonable for the FDA.  As this is an ethernet transformer, you may start by recommending a test with the FDA driving Riso resistors into a parallel Rload, which is followed by the ethernet transformer.  You can always suggest removing Rload or changing the values of Rload and Riso.  

Is the customer driving a Base-100T ethernet signal, or are they driving a different signal type and using the HX1188NL for another signal?

The short answer: yes you can do this, but there are considerations .