Overheat onboard V-regulator if drawing from Vin?


I'm using the arduino Mega (and hopefully the Uno if I can simplify the design) to turn on and off several very long 12V LED strips. I'm using the I/O pins to turn on/off MOSFETS which pass 12V from the Vin pin (with a 12V power adapter plugged into the barrel jack) to the LEDs.

My issue is that I think I blew the voltage regulator- it's always very hot and now doesn't work. I know there are limits about how much power the onboard regulator can dissipate, but what I can't find through researching is: does that apply when I'm drawing from Vin, NOT the regulated 5V and 3.3V supplies? If so- is the best option to put a barrel jack on the PCB w/ a higher V Regulator and then passing the regulated 5V into the arduino 5V pin to power it? If I do that through the external pin, will the Arduino's V-Regulator also regulate that down to 3.3V or will I have to provide that pin's power from the VReg on my board as well?

Otherwise, perhaps it's from connecting 11 digital I/O pins straight to the MOSFET gates w/o 1K resistor, which I know is unadvisable (it sources current- right?), but there's not a lot of room on the board. Are those resistors really neccessary- it works find without it, but it might be using too much current?


Are those resistors really neccessary- it works find without it, but it might be using too much current?


Eight pages of discussion so far.

The Vin pin is on the other side of the protection diode, plus 12V is on the high side for the input socket.

I'm not that familiar with the arduino hardware, specifically the protection diode. By "other side of the protection diode", you mean all the current I'm drawing for my LEDs is flowing through some of the arduino harware, not just out of the wall socket?

Does it make sense then for me to mount a jack onto my PCB and power the arduino using Vin?

The power-in socket goes to Vin via a M7 diode (D1). If you are powering a lot of LEDs from Vin (a couple of amps, say) then you are therefore dropping a couple of watts through D1. The M7 is rated at a maximum average forward current of 1 amp, so it is probably the diode that is getting hot.

It is probably a good idea to power the LEDs directly (connect the grounds of course) and run a wire across to the input jack or Vin. The input jack is probably better because then you still get the benefit of the reverse polarity protection.

It helps to download the schematic for your board, so you can tell what you're hooking to what.

The only REAL caution about Vin is that it's on the wrong side of a series protection diode. If you make a mistake with the polarity to Vin there's no protection diode to prevent damage to PC1 a 47uF cap on the input to the regulator. The regulator data sheet doesn't indicate that dire things will happen if the input is reversed and does say that it (the capacitor) is required for stable operation from batteries where it actually makes the battery effectively last longer. So be careful. I use a 7.5V supply for DC operation when I'm not using the USB power and anything external to the Arduino like displays or sensors are powered separately from a 5V supply. A 5V 1A Switcher used originally as a cellular battery charger and this has proven adequate for all I've done in the past.


Thanks guys that helps a lot:


Everything will still work (5V and 3.3V) if you power directly from Vin, just make sure the polarity isn't reversed. Since I'm using a standard 2.5mm, I don't think this will ever be a problem

Alternative solution?: I've seen a couple people use thermal paste and a heat sink to cool down the voltage regulator. If I put on one the VReg AND the Series protection diode, do you think the ~1A limit on current through the diode would be larger since it's shedding heat much faster?

Diode D1, marked M7, gives no clue as to what the real part number might be to look up its characeristics.
I know we've discussed it as having 1A capability, which would support the 800mA rated regulator.
I don't know if it will support >1A even with cooling.
What I don't understand is why the regulator failed, even with controlling MOSFETs. I can't see that drawing a significant amount of 5V current

Mine seems to be marked GW 4007, and a search for "M7 diode" seems to reveal that is in fact its part number:

Forward current: 1 ampere.

Mine seem to be marked:
Duemilanove: M7 DIC
Uno: HY M7

Have you measured for a short, or at least a low resistance across the output of the Regulator and ground? A possibility is that one of the parts is damaged somehow and shorting out the regulator, or at least drawing lots of current through it.