Regulator Failure on Load Enable – Thermal Stress or Inrush Condition?
We perform a test on this device where the regulator input power is enabled with an 8A load connected to the output, while the device itself remains disabled. We then drive the EN pin high to enable the device under load.
In one of our boards, this resulted in the regulator overheating and showing visible damage (burned at one corner). Subsequently, we measured a 7-ohm short to GND on the input.
Is this behavior expected under such conditions? I understand that 8A is a significant load, but I’d appreciate any guidance before proceeding with further testing, as we've only evaluated one unit so far and want to avoid damaging others.
Any input would be greatly appreciated.
1 answer
The following users marked this post as Works for me:
| User | Comment | Date |
|---|---|---|
| ReliablePotion | (no comment) | Sep 16, 2025 at 17:02 |
There are many reasons a regulator can fail under heavy load.
Perhaps the most common is: The regulator datasheet misleadingly claims it can supply a certain amount of current (under certain conditions), and some other device claims it pulls some slightly less current (at certain times), but testing shows they aren't really compatible under normal conditions and times. Nearly all devices more complicated than a simple resistor will briefly pull far more than their rated current on power-on. Even if you measured that your load pulls 8 A under steady-state conditions, I suspect that your load pulls far more than that when first turned on. Also, nearly all transistors and other power components can't actually continuously supply the current claimed on the first page of the datasheet. A careful reading to p. 26 of the datasheet indicates that the "8 A" advertised only applies at ambient temperatures of +15'C (59'F) or below, which I find pretty chilly.
Another common reason for voltage regulator failure is rapid overheating -- even regulators (like this one) designed to automatically shut down to protect themselves from overheating can still fail if it heats up so fast that it fails before the thermal protection kicks in. Is the current limiting resistor R_CL properly attached and less than 2.21 kOhm (see Section 5.3 of the datasheet)? Is there sufficient heat sinking -- are there thermal vias underneath the chip to carry heat into the other copper planes and to the heat sink on the opposite side of the board from the chip, and plenty of copper to spread out the heat, as shown in the "MIC28517 Evaluation Board User’s Guide"?
"Underestimating Complexity of Power Supply Design" is a common problem.
Good luck!
0 comment threads
2 comment threads