Thermal images of a Mega

Some interesting thermal images of an Arduino Mega that seem to back up the "VIN should be less than 7 volts" advice we often give.

http://australianrobotics.com.au/news/fuelling-your-arduino-why-you-should-use-an-external-power-breakout


Rob

Impressive and instructive! And a funny project between the lines (Arduino as handwarmer ;)

The main arduino specs site should link to these pics imho.

Very interesting, thank you!

When I first got my Uno I wondered how big a power supply (what voltage) to put into the power jack and thought "well, the more power, the better, right?".

It's now somewhat clearer that 7V is really the "sweet point" which is enough to operate the voltage regulator properly, but not so much the voltage regulator gets hot.

Here's a question (I think I know the answer, but asking in public documents...):

I have in many cases built devices that use external power of one sort or another, some of them 5v, but other voltages (especially 12v) end up being switched by the Arduino to do useful things. In these cases, I have simply added a 5v regulator (7805 with a heatsink) to feed the Arduino board and connected the produced 5v directly to the 5v pin on the Arduino, with a common ground of course. Though I don't always do it, I also sometimes will add an electrolytic capacitor (something halfway reasonable, 100uf or more) after the regulator to smooth out the power a little bit. I figure it can't hurt when the supply power might be seeing jumps due to motors or whatever is being driven, and may be unregulated power in many cases.. in thinking about it, it would probably not be a bad idea to add another before the 7805. Caps are cheap, clean power is worth it. When I start running into problems with odd random behaviour and resets, power noise is a common reason the ATMEGA decides to flip out.

I would think that in some ways this is a superior power solution than using the onboard regulator, but the nagging question remains in my mind as to whether the traces to the leads on the board are okay with this method of providing power, and if there is any issues I hadn't otherwise considered?

Really great writeup, concise and a quick read. Going to check out some of his other posts, I always enjoy a well-written article like that..

I wished that there are cheap, rechargeable, common, high capacity 5V batteries so that I can bypass the regulator on the Arduino and save power!

focalist:
I would think that in some ways this is a superior power solution than using the onboard regulator, but the nagging question remains in my mind as to whether the traces to the leads on the board are okay with this method of providing power, and if there is any issues I hadn’t otherwise considered?

I don’t see why that wouldn’t be a good solution (always use the caps, though - it prevents oscillation of the regulator); in fact, it might be better, since you could add your own heatsink, or (if you wanted), add a bypass transistor to bump up the current output.

dkl65: I wished that there are cheap, rechargeable, common, high capacity 5V batteries so that I can bypass the regulator on the Arduino and save power!

Four rechargeable 1.2 volt cells (NiMH or NiCd generally) in series will give you 4.8 volts, which many people have used without a problem (YMMV)...

Why does the processor run warmer when the supply voltage is increased? Does the regulator's output voltage increase as the regulator warms?

Why does the processor run warmer when the supply voltage is increased? Does the regulator's output voltage increase as the regulator warms?

Notice the bottom of the colour scale goes up from 9.5 - 12.9° so the false colouring is shifted a bit ... so it might be an illusion?

...and the AVR is quite warm compared to other voltages.

The author claims the processor is warmer.

[quote author=Coding Badly link=topic=106232.msg797779#msg797779 date=1337322669]

Why does the processor run warmer when the supply voltage is increased? [/quote]

It's happier.

I wonder if he let it run long enough for the processor to reach its max temp when he started it with USB power.

Hi, I think it’s time we can/should go away from linear regulators, even on these “Little guy” boards.

I have been trying to find a source for 7.5v 1A low-cost Wall Wart supplies, but all I see that is low cost is 5,9,12V

Here’s an example of high-function Arduino-compatibles with mini switching power supplies. This supply design is used on 3 IteadStudio products: Iteaduino2.2, Iteaduino2560 (Mega), and Iboard (328 type with onboard W5100 Ethernet)… This uses an MP2307DN switching regulator chip for 5 Volt output from 7 to 27 ! Volts input. There is a separate 1117 3.3V linear regulator and the processor / I/O can be run from that as well. Here’s the schematic, and a closeup view of the power part of a board:


I wish the UNO had kept the FTDI USB interface which is less troublesome than the 8U32, gone to a switching supply for 5V, and added a good 3.3V regulator like this…

I’m trying to get a good wholesale price on these so I can supply them at a decent price for those who don’t want the absolutely lowest price and want decent power supplies.

Yes I think these switchers are so cheap and easy to use these days it's time to start letting go of linear regulators for the "front end".

As for the FTDI chip, it's very simple to use and reliable, I like it. I just wish they had a smaller version, after all you only need about 6 pins.


Rob

Graynomad: As for the FTDI chip, it's very simple to use and reliable, I like it. I just wish they had a smaller version, after all you only need about 6 pins.

Rob,

Take a look at the new FT230 and FT231 from FTDI. 16 and 20 pins, respectively, and less than half the cost of the FT232. I think I'm going to like these even better!

Jon

[quote author=Nick Gammon link=topic=106232.msg797810#msg797810 date=1337326159] [quote author=Coding Badly link=topic=106232.msg797779#msg797779 date=1337322669]

Why does the processor run warmer when the supply voltage is increased? [/quote]

It's happier. [/quote]

I'll guess thermal transfer through the die. (Case?)

terryking228: Hi, I think it's time we can/should go away from linear regulators, even on these "Little guy" boards.

A couple months ago or so, somebody mentioned some DC/DC converters which were very close in size, and with the same pinout, as a the usual TO-220 regulator. I think it was something by Murata, but I'm not seeing specifically what I remember being mentioned.

http://www.murata-ps.com/en/products/dc-dc-converters.html

Graynomad: Yes I think these switchers are so cheap and easy to use these days it's time to start letting go of linear regulators for the "front end".

Is part of the reason that although the IC is cheap, you also need to buy a coil, and matching those to your requirements is (so I've read) a bit tricky? Those Murata switching regulators cost more, but I think they include all the parts.

some DC/DC converters which were very close in size, and with the same pinout, as a the usual TO-220 regulator.

Traco also make them and a few others, IIRC about $12 for a 1A PSU with inputs up to about 30V.

you also need to buy a coil, and matching those to your requirements is (so I've read) a bit tricky?

Possibly, although many of the datasheets specify a recommended inductor.

Take a look at the new FT230 and FT231 from FTDI.

Thanks for that, just the thing and I see they now have SPI and I2C version as well. I have an application right now that would love one of those SOIC16 USB chips.

I think I'm going to like these even better!

Likewise. I know you can get USB in the processor these days but I don't know how to drive those on-board peripherals whereas I do know that these FTDI chips just work out of the box.


Rob