3

I'm trying to have redundancy in part of my design.

I want to use two LDOs in parallel before an OR-controller (specifically, an LM66100-Q1.) One LDO is enough to feed the system, so it will not be overloaded.

If one LDO fails, or one LDO has a short circuit, fuse burns and the other LDO can supply the system.

Diagram Block

Does this make sense? Can this work without any problem?

JRE
  • 67,678
  • 8
  • 104
  • 179
Indexer
  • 51
  • 3
  • Without a design spec for FITS or MTBF it makes no sense to use two unreliable parts rather than 1 better part. Pls define environment and MTBF goal. https://parts.jpl.nasa.gov/mmic/4.PDF Remember this only defines a fault-free solder process with perfect parts and a perfect flaw-free design. If this was going into your wireless charged pacemaker, would you trust it? How reliable do you need it? (get my point?) – Tony Stewart EE75 Dec 15 '22 at 07:06
  • 1
    what if one branch fails short to ground? it will fry the second branch too.. Designing redundancy is more complicated than just paralleling units. A better version cuts out the broken unit, connects the replacement and notifies the operator – tobalt Dec 15 '22 at 08:19
  • 1
    If one LDO fails short to ground, fuse burns and it stops being connected to the rest of the system, because of the or-controller (based on mosfet). It isnt? Then, do you think that is better to use only one LDO. – Indexer Dec 15 '22 at 10:06
  • @Indexer, yes you are right. The diode controllers will protect the remaining branch. So it can make sense, but only if the user has a way to known if and which branch is dead to replace it before all branches wear out. – tobalt Dec 18 '22 at 07:42

1 Answers1

0

This would only protect the load from a power failure in the event of an LDO failure. In the event of a short on the load it would not protect. The addition of more components would probably also increase the failure (albeit low) rate.

Another thing you may want to consider is a backup power instead of two supplies in parallel.

But as far as paralleling ldos the ideal or diodes are probably enough impedance between the supplies to mitigate large currents between the supplies from small voltage differences

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
  • The bottom line is the MTBF mean time between failure and redundancy does not protect an imperfect design such as a shorted input and reverse voltage breakdown without reverse diode protection across LDO if Vin >> 5V. So inadequate answer until improved (-1) – Tony Stewart EE75 Dec 15 '22 at 07:23
  • This is a part of a design, but 5V INPUT comes from a 28 to 5V buck converter (also in parallel with OR diodes), with output OV protection, exactly TPS5430-EP. This buck shut-downs if Voutput is over 5.625V, so it protects LDO from input OV. – Indexer Dec 15 '22 at 07:44