Hi everyone!
I have a very simple question.
I've just read the following sentence on the Arduino UNO page:
Supplying voltage via the 5V or 3.3V pins bypasses the regulator, and can damage your board. We don't advise it.
What does this mean?
-Supplying voltage to Arduino via the 5V or 3.3V pins from external power source [...] OR
-Supplying voltage to anything(leds,modules,etc) from Arduino via the 5V or 3.3V pins [...]
Thank you!
If external power is supplied to the 5V header pin while the barrel jack is unpowered, the 5V regulator can be reverse driven & damaged. To avoid damage, connect a diode from 5V (anode ) to Vin (cathode).
The 3.3V regulator is powered from the 5V regulator, I don't know of anyone connecting external 3.3V to the 3.3V header.
@Crossroads
Should this also be done on a Pro Mini also?
Wouldn't hurt.
Promini & duemilanove from 2009 designs had hardier regulators - I have been running 14 promini's with external 5V since Jan 2010 with no issue.
It was the change for Uno & Mega when this started occurring.
Since many clones are coming out now, wouldn't hurt to read the spec for the regulator used if you can determine what it is. Or just play it safe and add a diode.
CrossRoads:
If external power is supplied to the 5V header pin while the barrel jack is unpowered, the 5V regulator can be reverse driven & damaged. To avoid damage, connect a diode from 5V (anode ) to Vin (cathode).
That makes no sense.
For a start, the USB port does precisely that - supplies power to the 5V header pin and thereby the output of the regulator.
See Protection Diodes, page 10
Barrel jack may be unpowered and a turned off wallwart could have the effect of being a short to Gnd. Being unconnected would be different - no path to Gnd.
Hmm, one would think the reverse polarity diode would prevent that tho.
I have the thread bookmarked at home where bricking of 2560 boards was happening when folks were powering the board from 5V. The regulator used on Duemilanove boards and Promini boards was not subject to this problem.
CrossRoads:
I have the thread bookmarked at home where bricking of 2560 boards was happening when folks were powering the board from 5V.
Which "2560 boards"? The Arduino Mega has the exact same power circuitry as the Uno.
Paul__B:
CrossRoads:
If external power is supplied to the 5V header pin while the barrel jack is unpowered, the 5V regulator can be reverse driven & damaged.That makes no sense.
For a start, the USB port does precisely that - supplies power to the 5V header pin and thereby the output of the regulator.
Yep. USB power goes directly to the internal 5V rail (and therefore to the output of the 5V regulator NCP1 1 17ST50T3G)
I'll post the thread when I get home.
Paul__B:
CrossRoads:
If external power is supplied to the 5V header pin while the barrel jack is unpowered, the 5V regulator can be reverse driven & damaged. To avoid damage, connect a diode from 5V (anode ) to Vin (cathode).That makes no sense.
For a start, the USB port does precisely that - supplies power to the 5V header pin and thereby the output of the regulator.
Yes it makes perfect sense - regulators are susceptible to damage
if reverse driven, since this reverse biases the pass transistor EB junction
(the easiest way to blow a transistor). Individual regulators have idiosyncrasies
too, especially LDO ones.
The Uno and Mega have power-switching circuits that detect Vin
and switch over to USB only if Vin is low.
Yep. USB power goes directly to the internal 5V rail (and therefore to the output of the 5V regulator NCP1 1 17ST50T3G)
Nope, not directly, but rather the USB voltage first goes to a 500 ma thermofuse and then onto a MOSFET that is under control of the infamous arduino auto voltage switching function. If Vin has greater then proper minimum voltage then the board's 5V bus comes from only the on-board voltage regulator even if the USB is plugged into a PC.
MarkT:
The Uno and Mega have power-switching circuits that detect Vin
and switch over to USB only if Vin is low.
That's the case we're talking about.
When Vin is low the 5V from USB goes directly to the output of the 5V regulator.
The regulator has no input, the output is connected to 5V. Every Arduino in the world that's running on USB power is doing that as we speak.
fungus:
MarkT:
The Uno and Mega have power-switching circuits that detect Vin
and switch over to USB only if Vin is low.That's the case we're talking about.
When Vin is low the 5V from USB goes directly to the output of the 5V regulator.
The regulator has no input, the output is connected to 5V. Every Arduino in the world that's running on USB power is doing that as we speak.
It's the word 'directly' that I sought to correct. Also the power management for an arduino board has some history. The very first design had no 'auto voltage switching' but rather the user had to move a jumper clip to choose from using USB voltage or on-board voltage regulator. Then they went with the current design but with an older different model 5 volt regulator. It was only after the (I think Uno rev1 board) change to a different model 5 volt regulator did a few people report damage to the on-board voltage regulator when they applied an external regulated +5vdc to the 5V pin. This lead to speculation that the new voltage regulator was more vulnerable to having external voltage applied to it's output. It seemed to happen to enough people that only then did the Arduino product page add the recommendation to not apply external voltage to the 5V pin. Most all linear regulator datasheets caution against this practice and many recommend in their datasheet to adding a protection diode across the input and output pins of the regulator.to protect the regulator.
So that's the history, somewhat unclear what advice to give at times. Obviously the Arduino company is taking the most conservative stand by recommending to not apply external regulated +5vdc to the 5V pin, but so many of us do it all the time, one way or another. I apply +5vdc to the board via it's ICSP connector pins when I burn bootloaders or upload sketches using programmer option.
However if I had my way and was designing a board I would eliminate the whole auto voltage selection function and revert back to using manual jumper clips.
Lefty
retrolefty:
It's the word 'directly' that I sought to correct.
OK, it's not "direct" if we're being pedantic, it goes through a fuse and a MOSFET as part of the voltage selection circuit (directly above where it says "Rev3" in the schematic).
Electrically though, it's the same thing. The output of that MOSFET feeds directly into the 5V rail, and therefore the output of the 5V regulator.
retrolefty:
a few people report damage to the on-board voltage regulator when they applied an external regulated +5vdc to the 5V pin
The Arduino Uno Rev1 seems to use the same power switching circuit:
I don't see how connecting a 5V supply to the 5V pin is any different from using USB power.
@ Lefty: Ditto. The jumper pins are the most foolproof method, though not as convenient, and convenience / ease of use is an understandable reason for the Arduino hardware team to choose the path they did. I used that approach for home-made boards for a long time, pretty foolproof and up to 4 external power sources can be selected at will using a "+"-shaped pin header + jumper shunt, with the center pin being the one that supplies the board and the ends of each arm being a potential power source.
Another approach is to simply dump all external power via Schottky diodes into a common power rail. That works as long as every component is OK with the incoming power varying somewhat (4.7-5V). The Arduino can certainly handle that and many 5V logic chips too but there are potential issues with Analog measurements unless you implement an external reference of one kind or the other (Zener diode + resistor into AREF, for example).
Some of my home-made boards are 3.3V only, making a 4.7-5V common-bus approach very easy to implement - i.e. the downstream LDO doesn't mind one bit if the incoming voltage drops to 4.7V, that means less heat to dissipate.
Constantin:
The jumper pins are the most foolproof method, though not as convenient, and convenience / ease of use is an understandable reason for the Arduino hardware team to choose the path they did.
I'm not sure...I don't think many people will be constantly switching between power sources.
Even if they are, they still have to physically plug/unplug the external power supply (or throw a switch, or whatever).
Electrically though, it's the same thing. The output of that MOSFET feeds directly into the 5V rail, and therefore the output of the 5V regulator.
I grant you that from the on-board regulator's perspective there is little difference if the board is being USB powered or powered via an external regulated +5vdc applied to the shield 5V pin, in both cases an 'external voltage' is being applied to the on board regulator, other then the USB path has fuse proteciton.
The root cause of failures some were seeing early on with the Unos (and I bet that the arduino co had some returns to verify bad regulators were happening somehow) may not have been nailed down conclusively. As I said I believe the arduino folks 'official policy advice' to not connect external +5vdc was just being cautious and perhaps to try and cut down on warranty returns.
Perhaps it was just an early bad batch of voltage regulators. I do know that it is a bad engineering practice to 'hardwire' two voltage sources together as in the case if one was powering the board via it's on-board regulator or an external regulated +5vdc applied to the 5V pin, which is the case if the board was then plugged into a PC via USB, hence the reason they installed the MOSFET controlled switch to isolate USB voltage in that case.
Again I am not a fan of the design of the existing auto-voltage selection function if for no other reason is that it takes up a lot of board space and doesn't give me complete control. as in the case if I wanted the board powered from USB even if I have a valid Vin voltage applied. Jumper clips are what I would use.
But the real difficulty is what advice do you give a beginners if he asks about the arduino policy advice to not wire external +5vdc to the 5V pin? just tell them it's wrong advice and to ignore it, or try and explain the historical reason the product page states what it states. I tend to normally not answer such questions unless I'm in the mood to go through all this once again, with no real consensus reached. :~
CrossRoads:
See Protection Diodes, page 10
http://www.onsemi.com/pub_link/Collateral/NCP1117-D.PDF
Yep, have looked at that before.
Only when Vin is shorted or crowbarred to ground and Cout is greater than 50 µF, it becomes possible for device damage to occur.
Now according to the UNO R3 schematic, Cin is 47 µF so I suppose that could be equivalent to a dead short if you connect a "live" 5V supply with a large output capacitor to Vcc. Powering up a system once properly connected should on the other hand, be no problem as the rise time of the supply would be limited.
CrossRoads:
Barrel jack may be unpowered and a turned off wallwart could have the effect of being a short to Gnd. Being unconnected would be different - no path to Gnd.
Obviously, if you are going to power via Vcc, you should not have anything connected to the Vin pin. The barrel jack is of course, isolated by a diode.
I suppose in summary, the regulator may well be a bit fragile, but susceptible only to "hot connect" under certain "experimental" connections, should be quite safe for a properly implemented system.
If you have power on the barrel jack then connecting a 5V supply to the 5V pin ISN'T the same as connecting USB power. I could see how damage could occur if you did that.
When there's nothing on the barrel jack though, I don't see how...