ChuckY asked these questions on YouTube. YouTube's comment system is difficult to address multiple points, so I am reposting here.
The too long didn't read version:
- We are talking about current consumption, not regulation.
- Devices draw what they can (or designed to). This statement is fact, not "folklore."
- LEDs do not violate that statement.
Shifting the conversation a little to current regulation: I just got an RPi 3B and added an audio hat. I am using a wall-wart USB charger that is rated 2A nominal, and the power LED is staying steady on. According to Pi getting started info pages, for the 3B that means power supply adequate for that config e.g. it would start to flicker if current starved. (They say that the meaning of the power LED in the 3B (all Pi3s?) is opposite of earlier Pi b boards.)
Sooner or later I'm going to want to add some more gadgets, thus will incur more current draw. The recommended 2.5 A wall wart would be an option, but circuit tinkering begs the larger question: how much current can be safely sourced through the preferred micro USB connector? I have a few good computer PSs with multi-amp 5V leads but hesitate to try them for fear of frying the 3B. I have other smaller 5V power supplies scavenged from junk electronics from hamfests and transfer station discards. Seems a shame to buy something when I have all this good inventory to work with.
So I'm confused about how to think about current regulation, and about folklore wisdom that says devices will only draw the current they need, even when the source supply is well in excess of their needs. Clearly the LED with limiting resistor circuit is an example of folklore being wrong in some cases. I'm reasoning that to be absolutely safe I need a current limiting circuit between any high current 5VDC source and the Pi micro USB. Maybe you could work that into a video, or point me to a tutorial that addresses my question? Also, would you weigh in on back feeding Pi through any of the USB ports < seems risky to me. tnx in advance.
There are a lot of different questions here with some confusing points. So let's break these down.
Shifting the conversation a little to current regulation
This concept is current consumption, not "regulation." Current regulation is like with voltage; you force current to be a certain amount. Generally when you regulate current, you let the voltage swing around. This is how constant current regulators work when powering high-power LEDs.
According to Pi getting started info pages, for the 3B that means power supply adequate
I'm not aware of such a statement. Such a statement is a guideline. If that is in fact on the getting started guide, I would wish they remove it. The power LED on all of the Pis is just that, an LED. It's a good first-order indicator that power is on. However, the only way to know for sure the supply is adequate is by measuring both the 3.3 volt and 5.0 volt rails.
how much current can be safely sourced through the preferred micro USB connector?
This is my concern with the Pi's method of being powered. The USB connectors used aren't designed for much over 2 A of current. I don't think it is advisable to draw more than that. It is possible the Pi foundation is using a special version of the USB micro connector, but considering the price point of the Pi, I find that unlikely.
So I'm confused about how to think about current regulation, and about folklore wisdom that says devices will only draw the current they need
This statement about current consumption is not "fokelore wisdom." It is a fact. Regardless of the amount of current a supply can provide, the device connected will only draw what it can.
Clearly the LED with limiting resistor circuit is an example of folklore being wrong in some cases.
Sorry, but you are entirely off-base in this case. Yes, LEDs do need current limiting. As a device, an LED will draw as much current as it can. But that phrase is the key phrase here "as much as it can." Without a current limiter, an LED will draw around 100 mA, at least, until it burns out. It does NOT matter if the supply you provide can provide 100 mA, 1 A, or even 100 A. The LED, by itself, won't draw more than 100 mA.
The amount of current an end-device draws depends on its design. That is the case for a single LED and is the case for a Raspberry Pi. It has been designed so that no current or thermal runaway situations can occur. Whether you use a supply can that provide 2 Amps or 2,000 Amps, the Pi is only going to draw about 2 Amps. The only exception is in the event a fault condition occurs.
That said, the more current available when a fault occurs, like with your unlimited LED example, the more damage can occur. But that is the essence of design.
I need a current limiting circuit between any high current 5VDC source and the Pi micro USB
Any current limiting you introduce between a supply and a device will create a voltage drop. If you want to add circuit protection, for the case of a fault condition, that is different.
would you weigh in on back feeding Pi through any of the USB ports < seems risky to me
Since the Raspberry Pi foundation doesn't provide actual schematics of their boards, it is difficult to say. Using a multimeter, I found that it appears all of the 5-volt pins are connected. So whether you provide power to the micro USB port or back-power through the USB A connectors or just power via the 5-volt pin, they all appear the same. However, I have no idea if there are any differences in trace widths or power planes since there are no specs published.
My recommendation is: do NOT power high current devices through the Pi. Power them with a separate supply.