To reduce EMI the datasheet recommends using a LP filter as shown in: http://ww1.microchip.com/downloads/en/DeviceDoc/39762f.pdf page 219, Figure 19-2.
The filter is made of 56pF capacitor and 120ohm ferrit.
Why is the 56pF capacitor placed directly in parallel with the TPOUT+/- pins?
I would have placed the ferrit bead in series with the TPOUT+ pin and the capacitor after the ferrit bead (and likewise with the TPOUT- pin).


I finally figured it out.
The TPOUT pins can only sunk current, which is why the transformer center-tap is connected to VCC.
This is why the LP filter seems in inverse order, current is only flowing into the TPOUT pins, never out of the pins. And therefore the LP filter will limit the high-frequency current flowing into the TPOUT pins.

| improve this answer | |
  • 1
    \$\begingroup\$ I'm dubious about that. TPOUT may only sink DC but it generates outgoing AC by modulating the magnetic field in the transformer.It's a genuine AC source (else the ethernet implemented would be non functional.) \$\endgroup\$ – Russell McMahon Jan 11 '12 at 12:57
  • \$\begingroup\$ But the flow of current would still only be INTO the TPOUT pins, or am I misunderstanding something? \$\endgroup\$ – JakobJ Jan 12 '12 at 8:44

It does not seem to make sense by "normal" standards. 'Deep magic' may apply.

Their section 19.1.5 explains the filter considerations. The ferrite cored inductor location is correct.
But, a classic low pass filter would have the capacitor at the inductor + 49.9 ohm, resistor junction.


  • It's a mistake, or

  • The loading effect on the slew rate of the IC is effectively what is being utilised. Unlikey. or

  • Other ?

None of the following useful refs use that type of filter:

Stunning & useful related discussion here - head spins - but not directly relevant to this question.

Relevant and usefl but doesn't answer this question. Common Mode Choke Filtering Improves CMRR in Ethernet Transformer Applications

More SimpliPHY’d TM 10/100/1000BASE-T Magnetics and EMI Control

And more

| improve this answer | |
  • \$\begingroup\$ Hehe, yeah I got me thinking too. But "Deep magic" may apply. ;-) \$\endgroup\$ – JakobJ Jan 11 '12 at 9:56
  • \$\begingroup\$ Thank you, the "SimpliPHY’d TM 10/100/1000BASE-T Magnetics and EMI Control" made me realize why the filter has been designed as it is. \$\endgroup\$ – JakobJ Jan 11 '12 at 12:39
  • \$\begingroup\$ @JakobJ You are more than welcome to submit an answer to your own question. \$\endgroup\$ – Michael Pruitt Feb 11 '12 at 3:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.