Post Go back to editing

Transformer and Transistor

Thread Summary

The user inquires about substituting the PSMN075-100MSEX MOSFET with the BUK9M60-100LX and using a transformer with common mode chokes (S558-5999-T7-F) instead of the recommended Wurth 7490100143. The final answer suggests verifying the MOSFET's SOA against the LTC9102/LTC9103 foldback curves and testing the circuit. Common mode chokes are problematic for the LTC9103 as they can distort low-frequency/DC signaling, leading to signal integrity issues or communication failure. Alternative MOSFETs tested and approved by ADI include the SI7322ADN-T1-GE3 and DMT10H015LFG-7, available at Digikey.
AI Generated Content
Category: Hardware
Product Number: LTC9103

The datasheet strongly suggests PSMN075-100MSEX for the transistor and a 10/100 Mbps transformer without common mode chokes for isolation.  Wurth 7490100143 is given as an example.

Both of these components are difficult to source.

BUK9M60-100LX from Nexperia seems to be very similar in parameters to PSMN075-100MSEX.  Would this be a valid alternative?

For the transformer, S558-5999-T7-F from Belfuse is the same footprint, but it has common mode chokes.  Why are common mode chokes and issue for this application?

  • We cannot test all variations of FETs in an application, so we can't say with any certainty that a particular FET would work in an application.  That said, the specs look close enough.  It's something you'll have to test.

     

    As for the transformer, the short answer is: because the LTC9103 uses low‑frequency/DC signaling over the Ethernet magnetics, and a transformer with integrated common‑mode chokes can distort or attenuate that signaling. The disadvantage of using a transformer with common‑mode chokes is reduced signal integrity, timing errors, or outright communication failure of the controller’s isolated interface.

  • Just to add a bit more information to Tim's answer; you should also verify the MOSFET's SOA. Check the datasheet for LTC9102/LTC9103 foldback curves to get an idea of the required SOA and test in your circuit. Alternatively, the SI7322ADN-T1-GE3 and DMT10H015LFG-7 have been tested by ADI and are acceptable substitutes for PSMN075-100MSE with stock at Digikey. And per the linked Q&A thread, keep the chokes on LTC9103-side of the high-speed bus and it should be OK.

    Best Regards,

    Eric

  • Related question:  With the SOA curves, when the FET is fully on (after classification), I will be Rds limited and most FETs of this relative package size will work.  Is the concern related to when the device is in detection/classification, or inrush current limit modes and the FET is in the linear region?  This is where the FET characteristics seem to vary much more.  The PSMN075-100MSE seems to be an order of magnitude better than a lot of others in this respect.      Thank you for the help with this.   

  • Yes, the concern is in regards to inrush and current limit events.