I have a design checklist question for a typical CAN circuit for an automotive application as below : The question is related to the ESD suppressor : "Was a capacitor parallel to the diode considered, respectively carefully proofed the max. power losses in case of high-frequency distortion ? (Critical : overloading of diode through current loading during BCI-test)" My application circuit is something similar to the below circuit, except that there are capacitors on the CAN lines on the other side of the ESD diode(NUP2105L) near the CAN bus, as well. enter image description here

I understand that the diode is for clamping high voltage surges and the capacitors are to smooth out the noise on the CAN lines, and BCI effect on the CAN lines is high frequency noise- I'm imagining this is what the CAN lines see during BCI test(I have actually asked a question regarding this, but unfortunately couldn't get an answer, Bulk Current Injection on CAN bus )

So, thinking of an answer to the question above, a capacitor parallel to the diode is considered(but is my thinking right ? are they asking about the capacitor on the other side of the diode- nearer to the CAN bus), but I don't know how to proof the maximum power losses in case of high frequency. Also how do I quantitatively calculate the overloading of the ESD diode during BCI.

  • \$\begingroup\$ Power losses during high frequency distortion, how do I calculate this ? \$\endgroup\$ – NIDHI Apr 28 '19 at 18:02
  • \$\begingroup\$ I think what they are meaning by saying high frequency distortion is that , during bulk current injection, are the capacitors proofed for the max. power losses, but I don't know how to calculate it. \$\endgroup\$ – NIDHI Apr 28 '19 at 18:08
  • \$\begingroup\$ Hmmmm. I read what a BCI test is and I agree. You would need the datasheets for the ESR and Z frequency curves of the caps you are using as well as the frequencies and magnitudes of the BCI test. \$\endgroup\$ – DKNguyen Apr 28 '19 at 18:16
  • \$\begingroup\$ onsemi.com/pub/Collateral/AND8253-D.PDF FYI \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 28 '19 at 23:30
  • \$\begingroup\$ @SunnyskyguyEE75 that is exactly where I've taken the above example circuit from. \$\endgroup\$ – NIDHI Apr 29 '19 at 17:53

NUP2105L specs.

current injection (BCI). Adding 5.0pF RF ceramic capacitors in parallel with the TVS diodes, as shown in Figure 14, provides a simple solution to attenuate high frequency noise signals. Figure 13. The NUP2105L functions as a capacitor at frequencies below the resonant frequency (fR); however, the inductance term must be included at frequencies greater than fR=616MHz @0Vdc

Thus ultra low ESR (1 Ohm) low ESL, low capacitance diode has a Q (~5) at resonance near 616MHz implying some voltage ringing at this frequency. This Q means the diode conducts more reactive energy than resistive energy at this resonance. Yet at 0Vdc the diode should not be conducting much but due to this stored energy, and now high impedance parallel resonance it is damped by the CANBUS shunt R impedance. So if we add a capacitor with about 50 Ohms impedance at 616MHz then we can expect a well damped response at some lower resonant frequency.

By injecting a bulk cap across this low-capacitance TVS diode , we reduce the circuit resonant frequency and eliminate the current sharing of the diode ESL when it is conducting by diverting the current at these frequencies. But one must ensure the layout has low ESL in the Bulk Cap’s path to not introduce another high Q resonance.

This 5 pF cap is about 50 Ohms at 616 MHz and higher impedance at its new lower fR. CAN bus impedance is 60 Ohms. Matching these approximately, gives a well-damped low-ringing , low EMI result. 2-3pF is also suggested.

The EMI test results are improved in this region for Bulk Current Injection Test’s when this bypass bulk cap is applied correctly at the board interface with the TVS diode to achieve 30kV ESD protection.

  • \$\begingroup\$ "5pF cap is about 50 ohms at 616 MHz and higher impedance at its new lower fR" - could you please tell me how you arrived at this data ? 2-3pF is also suggested - for the same reasons ? Are you referring to the 5pF cap as the bulk cap? \$\endgroup\$ – NIDHI May 31 '19 at 3:20
  • \$\begingroup\$ yes the bulk cap Do need to know Zc(f)? if ground plane adds 2 pF then 2~3pF may be all that is needed \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 31 '19 at 3:40
  • \$\begingroup\$ Yeah , Zc(f) . Any idea how they decided on 5pF ? \$\endgroup\$ – NIDHI May 31 '19 at 5:18
  • \$\begingroup\$ Zc(f)=1/(2pi f C) \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 31 '19 at 14:57

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.