Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Don Y <blockedofcourse@foo.invalid>
Newsgroups: sci.electronics.design
Subject: Re: Power from another source fed to 1117 regulator output pin
Date: Mon, 30 Sep 2024 13:16:11 -0700
Organization: A noiseless patient Spider
Lines: 57
Message-ID: <vdf0ue$2c9fr$1@dont-email.me>
References: <B3wKO.8728$ALMc.3588@fx01.ams1> <vde54a$27tr9$1@dont-email.me>
 <AOyKO.63936$l6I9.52273@fx13.ams1>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Mon, 30 Sep 2024 22:16:15 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="e26c28db990f411086565032b28e36d0";
	logging-data="2500091"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+cE4H4IMkgEoz6t0fJgblv"
User-Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.2.2
Cancel-Lock: sha1:ORmQFbxlefbcrDCkydycIt6r+vk=
In-Reply-To: <AOyKO.63936$l6I9.52273@fx13.ams1>
Content-Language: en-US
Bytes: 4163

On 9/30/2024 7:43 AM, Pimpom wrote:
> On 30-09-2024 05:51 pm, Don Y wrote:
>> On 9/30/2024 4:37 AM, Pimpom wrote:
>>> I've been looking at the schematic of a popular product that has two options 
>>> as the power source. One source is a Vin point (pin or DC jack) that goes to 
>>> a 1117-5V regulator, the other is a USB port. The USB +5V line joins the 
>>> 1117 output via a Schottky diode.
>>>
>>> Does this mean that the 1117 has no problem with this arrangement - that is, 
>>> with power coming from another source to its output pin?
>>
>> Short the LDO's input to GND and consider how it will (mis)behave.
>>
> Sure, this issue is often brought up while discussing regulators. Apparently 
> the makers of the product (Arduino, various versions) seem to feel that the 
> input of the LDO getting shorted while being powered via USB is not likely 
> enough to bother with.
> 
> Anyway, what I'm thinking about is how the LDO's internal circuit reacts to 
> being reverse fed a voltage at the output pin without having an input but 
> *without* the input being shorted to ground.

If the input is powered *or* left "open" (assuming you don't have something
"ridiculous" on the input that can represent a short even for a short time),
it will likely be comfortable in that state indefinitely.

If *designing* in this sort of configuration, I tend to be more paranoid
about the types of "stuff" that will inevitably get connected to the
"input connector" -- especially if you opt for the ubiquitous "barrel
connectors".  There, it seems that the typical user just hunts for *an*
adapter (wall wart) that "fits", mechanically.  Consider that many such
devices are generic and don't even declare the device to which they
are *intended* to be mated (People's Wall Wart Company #773), the
user has very little by way of guidance.

As far as what the adapter (even if powered off!) might look like
from the LDO's point of view:  the polarity could be reversed;
there could be some bulk cap intended to make full-wave DC "look"
more like a regulated supply;  it could be a bare transformer winding
(i.e., AC); the wrong voltage/amperage (excessive ripple); etc.

I've not found a robust solution that doesn't quickly start to
look like overkill.  And, a customer will never admit to having
plugged the "wrong" adapter into it (before *possibly* finding
the right one to include with his RMA).  So, you see a tiny
insignificant part of your design proving to be the Achilles
heel of your *product* (is it worth disassembling to replace the
LDO or just scrapping the thing, whole?)

My current approach is NOT to allow external power sources (all PoE)
AND to use a non-standard connector to ensure the PSEs and PDs
aren't subjected to arbitrary device matings (there are many
things that call themselves "PoE" without being truly compliant).

And, I *still* expect to find some really clever idiot who will
manage to subvert all of my precautions!  <frown>