Arduino UNO regulator capacity?

I had a hard time deciding which section to put this is. The forum lacks a "General Arduino" section.

I just received a new Arduino UNO REV.3 after blowing up my Duemillanove.

Two things I noticed right away are the strange placement of the reset button and the voltage regulator is about half as large.
I could not find the data sheet. Is the regulator not able to handle as much current in the UNO?

Thanks

the official R3 schematic shows an NCP1117xxx. and a brief Google search suggests 1 amp max. You should check the specific part number on your board just to make sure.

Ok sounds good. Just looks alittle small for dissipation.

This problem comes up constantly. The 1A rating means almost nothing. Don't count
on loading the v.reg to 1A without it getting rather hot.

The smallish SOT-223 v.regs are only good for about 0.3-0.5W dissipation. This means
for Vin=12V, Pd = Vdrop*Iload gives a measly Iload = 0.5W/(12V - 5V) = 71 mA or so.
Slightly better if you factor in the series-diode drop. You can figure Iload out for other
values of Vin.

Larger v.regs, like D-PAK, are somewhat better.

72 mA is pretty poor. That is almost what the Arduino takes alone.

oric_dan:
The smallish SOT-223 v.regs are only good for about 0.3-0.5W dissipation.

Don't forget that the tab on the top is designed to act as a heat sink. So you can actually get more dissipation than 500mW. In fact, we need to go to the datasheet to find the real numbers.

The regulator used on the Uno is a NCP1117. http://www.onsemi.com/pub_link/Collateral/NCP1117-D.PDF

Which according to page 8 of On's datasheet the SOT-223 the maximum power dissipation at 50°C with the minimum sized copper pad is 650mW. The Rj is 85°C/W.

Three things note:

  1. The "minimum pad size" is 3mm x 3mm
  2. The graph is drawn for ambient of 50°C (2X room temperature)
  3. Most arduino projects are run in free air at or near room temperature

Looking at the PCB layout for the Uno, the regulator is attached to a ground plane. The ground plane is ~10mm x 15mm uninterrupted. (It is larger than that, but that's where other traces are running.)

Going back to the On's graph, let's look to see the maximum power dissipation for 10mm, which is conservative. That's nearly 1W. Again, we're talking about an ambient of 50C.

Alternatively we can use Rj from the same graph, to can calculate maximum power dissipation at room temperature:
Pdmax = (Tjmax - Ta) / Rja = (175C - 25C) / 100 = 1.5W

Imax = Pdmax / Vdrop = 1.5W / (12V-5V) = 214mA

Personally I have been able to draw up 250mA through the regulator before it started shutting down, which matches the above math.

This is also why we so many people who post saying: "My project works fine on USB but not on 12V!"

Alternatively we can use Rj from the same graph, to can calculate maximum power dissipation at room temperature:
Pdmax = (Tjmax - Ta) / Rja = (175C - 25C) / 100 = 1.5W

Imax = Pdmax / Vdrop = 1.5W / (12V-5V) = 214mA

Personally I have been able to draw up 250mA through the regulator before it started shutting down, which matches the above math.

Do you know how hot 175 degC is? You're talking about the outer limits here. That's
why I mentioned "getting rather hot". I still think 0.5W is a more reasonable value.

The obvious way to increase loading capability is to use a lower value of Vin. With
Vin = 7V, 250 mA is more realistic.

oric_dan:
Do you know how hot 175 degC is?

Yes I do. It's the temperature when the thermal shutdown starts. I will agree, however, that given every other part of the evaluation was done with a conservative value, that something like 150C is probably a beter value to use.

Instead of using guesses about what the part can do, we should always look to the datasheet to see what the parts can really do.

At 7V in, this regulator on the PCB, will easily handle over 500mA before thermal shutdown starts.

Instead of using guesses about what the part can do, we should always look to the datasheet to see what the parts can really do. ....

Guesses, yeah right!

I've looked at the datasheets MANY times. And I know hot from hot from hot. That's
where the 0.5W comes from.

I seriously doubt that many Arduino people want to run at the ragged limits, or burn their
hands either. Water boils at 100C, somewhat short of 175C, or even 150C.

175c? Ouch. I wish they used a more robust device. There is enough room on the board. It should be comfortably able to provide atleast the same amount as USB at 9v in IMO.

Well, you have your choice, 71 mA and it's not too hot, or 214 mA and you can fry
a small egg on it.

They're great at advertising 1 Amp v.regs for the boards, but not too good at mentioning
practical usage. They could at least say something like: "1-Ampere, Full Commando Mode".

Since using the manufacturers datasheet didn't seem to satisfy some, I decided to make some measurements.

I setup a load which would draw ~400mA. I tried with 200mA, but even up to 16V I wasn't getting a thermal shutdown.

All measurements done with a brand new Uno R3.

With an input of 12V on Vin the regulator did not shut down. I measured ~100C on the regulator package. This is the hotest the package would reach as when I went to 14V the regulator finally started thermal shutdown. This was evident by two things 1) VCC started dropping and 2) the package temperature starting dropping as well.

So 71mA is far (far) below what the regulator can handle at 12V. 200mA is still below what it can handle, but probably a good limit to keep the board from getting hot enough to burn yourself.


Going back to the original poster's question. The change to SOT-223 does have an impact. Same setup using a "R1" with a DPAK, no thermal shutdown occured with the same load up to 16V on Vin.

Reg_Full Power [compressed].png

Reg_Shutting Down [compressed].png