0
\$\begingroup\$

While using a capacitor with 2.3A rated ripple current as a dc link capacitor in the circuit (as shown in the picture below "Cd") I used, I replaced it with one with 2.8A rated ripple current parameter and noticed that the switching element heats up faster. Is this because of the rated ripple current? Or is it caused by a different parameter change in the capacitor? Note: I am using full bridge topology and using IGBT as a switching element.

enter image description here

\$\endgroup\$
4
  • 1
    \$\begingroup\$ I don't see any dc-link capacitor here: a front-end filter \$C_d\$ which endures most of the switching ripple and an output capacitor \$C_p\$ which should see less ripple. Capacitance is one thing but how both old and new equivalent series resistances (ESR) of the replaced capacitor compare? \$\endgroup\$ Commented May 7, 2023 at 20:05
  • 2
    \$\begingroup\$ I would assume Cd is the "DC link" capacitor? \$\endgroup\$ Commented May 7, 2023 at 22:28
  • \$\begingroup\$ Yes, I mean Cd capacitor. I have made the necessary correction. \$\endgroup\$ Commented May 8, 2023 at 4:21
  • 1
    \$\begingroup\$ It should not affect your transistors at all, merely lead to a cooler link capacitor. The effects you observe suggest a problem with the circuit or layout. \$\endgroup\$
    – tobalt
    Commented Jun 1, 2023 at 7:01

1 Answer 1

1
\$\begingroup\$

The term used in the literature is indeed a dc-link capacitor and my comment was misleading. Anyway, as shown in the below picture excerpted from the onsemi website, the model for evaluating switching characteristics includes the dynamic impedance of the high-voltage bus:

enter image description here

This impedance depends on various elements including the dc-link capacitor that is causing problems in your case. I would recommend you do characterize the impedance of both capacitors (the original and the new one) and see how they differ, particularly by looking at the inductive slope and the resonant points. It might indicate that the new one has really different parasitics than the previous one and, indeed, deteriorates the switching loss budget.

\$\endgroup\$

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