Overheating in POE with TPS23752

Which supplier is used for T2? Leakage inductance in T2 can cause conduction overlap between Q1 and Q2 leading to overheating. We are investigating a transformer design change to improve this.

Sorry, HA3803-ALB of Coilcraft.

We also qualified a second T2 supplier (E&E, Elec and Eltek PN 835-01042FC). I have recently received samples of these and they worked fine. Again, we are working a design change on the HA3803 now. If you want to escalate analysis of the design change, please contact Coilcraft and/or your local TI sales office.

http://www.google.com/url?q=http://mplweb.eleceltek.com/mpl/lib/download.jsp%3Ffile%3Ddatasheet/Power/PX236(03).PDF&sa=U&ei=gwFXTtblEKHLsQK65oy4DA&ved=0CA0QFjAA&usg=AFQjCNG5DRUdjCznnatNNdvIKV7WofftLw

Urgently need to buy 40 transformers. You recommend that I use the option of E & E?. The company has distributors in the U.S.? Thank you.

Due to time I will use the pre-production option transformer Coilcraft. What would you recommend to reduce overheating without changing the design?. It could, for example, instead of the IRF7855, a diode fast?. Thanks.

Yes, you can replace the sync FET with a diode such as the MBRD360 from ON. Or generally any fast, 60V, 3A diode.

Because of the problems I had with the design of the TPS23753EVM extracted, which buy 2 boards, you deem best to use the circuit shown in slua477?

Please use the applicable portions of the circuit in SLUA477.

En este circuito el transformador utilizado es de Coilcraft, No hay problemas de diseño en esta version?

No problems with this transformer, Silvio.

I have also recently introduced a design using the HA3803-ALB as described above. I have had recently assembled units brought to me with overheating of the FET device and also excessive heat produced by the transformer itself. Would the hot transformer be typical of the conduction overlap issues as described above? I don't see this issue on all units, I presumed that production variability of the transformer is the principle issue.

Could you supply a little more info on the FET/diode swap described. Is this intended as a solution for the issues, or an up-rating of the temperature rating of that specific part only?

Thank you very much for your support.

Yes, the hot FET/XFMR can be caused by primary/secondary FET conduction overlap due to mfg variability of the transformer. The solution above changes the design from a self driven (by the transformer) synchronous rectifier to a simple diode rectifier. This eliminates conduction overlap but will generally not be as efficient (over a properly working self driven synchronous design).

We have recently initiated a design change to the HA3803-ALB transformer and this addresses the issues above. We are in the process of updating the EVM user guide to reflect the new PN. The new PN will be HA3803-BL and you should be able to request some test samples from Coilcraft.

Thanks for the response, Eric.

Possibly of interest to anyone else reading this thread: Coilcraft now have a replacement part, the HA3803-BL_ should hopefully address these issues.