0
\$\begingroup\$

I've implemented the presented 80W PFC boost pre-regulator in the following application note without using the applied fuse at the input. Surely I will add the fuse to the finalized tested circuit. But it was recommended to me not to use the fuse for early tests because it doesn't protect the semiconductor devices from failure and if you use the fuse, besides losing the main components, now you also lose the fuse without any advantage. Is it correct?

The reason for most failures is due to the "gate staying ON for a long time" and resulting in high current flow, core saturation, short circuit of the input terminal, and burning of the diode bridge, MOSFET, CS resistors, and IC.

Does the fuse protect the circuit from these failures and explosions during initial tests or do you recommend continuing testing without a fuse?

https://www.st.com/resource/en/application_note/cd00195944-solution-for-designing-a-transition-mode-pfc-preregulator-with-the-l6562a-stmicroelectronics.pdf enter image description here

\$\endgroup\$
4
  • 1
    \$\begingroup\$ That fuse is important. It is fusing the power. \$\endgroup\$
    – Perry Webb
    Commented Jun 14, 2022 at 16:38
  • \$\begingroup\$ I sometimes replace the fuse with a toaster. If the toast smells good, the circuit has a problem, but the damage potential is low. \$\endgroup\$
    – Jens
    Commented Jun 14, 2022 at 18:53
  • \$\begingroup\$ @Jens, really? :)))) making electronics sweeter \$\endgroup\$
    – WeTech
    Commented Jun 14, 2022 at 19:43
  • 1
    \$\begingroup\$ @KMoradi Yes, sometimes a circuit escalates very fast and the components are dead before the source of the problem can be detected (e.g. cross conduction in a H bridge caused by a software bug, a bad inductive path etc.). The toaster or a similar 1 kW resistor allows to take measurements of the unwanted behavior without immediate component failure. Also good while repairing unknown circuits. \$\endgroup\$
    – Jens
    Commented Jun 14, 2022 at 20:12

2 Answers 2

3
\$\begingroup\$

Does the fuse protect the circuit from these failures (...?)

Would the fuse protect the switching element? Probably not - this is consistent with my experiences as well.

do you recommend continuing testing without a fuse?

Oh I've done this before and it doesn't end well. Essentially your circuit becomes the fuse, and does it's protection by burning until a circuit is no longer complete, and even 88VAC can draw a reasonable arc once the air ionizes. Do recall that mosfets like to blow short as well.

Even if the fuse wouldn't protect the switching elements, the difference in energy between a unfused failure and an fused failure could save the PCB, or adjacent capacitors and inductors/transformers, perhaps even the heatsink or enclosure. A fuse would avoid putting a nasty scorch mark on the workbench anti-static mat that all your coworkers could point to and say "remember that time you set fire to the bench?".

Digikey has 4A fuses for 30 cents a piece in qty 10 packs in my country, which is cheap insurance in my opinion.

\$\endgroup\$
5
\$\begingroup\$

Why would you ever not use a fuse? It's basic protection.

Sometimes a tungsten filament bulb or several in parallel (in series with the fuse) can help during early testing. It has relatively low resistance when the filament is cold, and it limits current when the filament is hot.

\$\endgroup\$
7
  • \$\begingroup\$ surely I will add the fuse to the finalized tested circuit. but it was recommended to me not to use the fuse for early tests because it doesn't protect the semiconductor devices from failure and if you use the fuse, besides losing the last components, now you also lose the fuse withot any advantage. is it correct? \$\endgroup\$
    – WeTech
    Commented Jun 14, 2022 at 16:22
  • \$\begingroup\$ @KMoradi You already have the answer to that. How did it work out for you? \$\endgroup\$
    – winny
    Commented Jun 14, 2022 at 16:30
  • 2
    \$\begingroup\$ @KMoradi even if the fuse wouldn't protect the switching elements, the difference in energy between a unfused failure and an fused failure could save the PCB, or adjacent capacitors and inductors, perhaps even the heatsink or enclosure or avoid putting a nasty scorch mark on the workbench anti-static mat. Digikey has 4A fuses for 30 cents a piece in qty 10 packs. \$\endgroup\$
    – Bryan
    Commented Jun 14, 2022 at 16:38
  • 1
    \$\begingroup\$ Cheap fuses are not designed to protect semiconductors, they're there to prevent the "explosion" you mention which could cause injury, not just destruction of a few pennies worth of semiconductors. The light bulb trick I suggest can (but might not always) protect semiconductors during testing. Fuses that are fast enough (it's more than fast, so I am loosely speaking) to protect semiconductors are seldom used in such circuits as yours and tend to be quite expensive- see I^2t fuses- they are sometimes used to protect thyristors. \$\endgroup\$ Commented Jun 14, 2022 at 17:29
  • 1
    \$\begingroup\$ @SpehroPefhany I used headlight bulbs the same way after re-wiring cars or trucks - after a fire it’s a mess!! But the bulb can stop a lot of damage - nice 5A current limiter : worked a treat. Gave the tip to one guy who said it was a waste of time. He just connected the battery and found out he had made a mistake with one cable… got to do most of his work again - this time with a bulb. \$\endgroup\$
    – Solar Mike
    Commented Jun 14, 2022 at 17:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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