How much currrent can Pro Mini's regulator supply?

Hey guys,

I need to know how much current the 5v Pro Mini's regulator can supply. I think I may have made a miscalculation when I was originally determining how much current I had available to power one of my circuits.

Because I'm just about to assemble the circuit, I decided to redo my current calculations to make sure I was running everything within suitable parameters. However, the numbers I've come up with are way lower than what I originally calculated.

Here's my current requirements:

1) The microcontroller itself, from what information I could find on the forums, seems to draw between 27-50mA.

2) I have two green leds which I want to power with 20mA becuase they're kinda dim. They're wired in series, so the total current draw for the two would be 20mA.

3) I have a piezo speaker. I've put a 1000ohm resistor on this to ensure that it only draws 5mA. Though it may be drawing much less than this.

4) I have a red power led, and a green mode LED on illuminated pushbuttons. Each of these will draw 10mA.

5) Lastly, I have a multiplexed 8x8 array of LEDs. I'd planned to allocate 160mA to this array. (20mA per LED per row, one row lit at a time, current sunk by darlington array connected to GND.)

The problem is that last one. It now looks like I may not have enough current to spare to run the array without making the LEDs really dim.

This, I've heard, is the datasheet for the Pro Mini's voltage regulator: http://www.micrel.com/_PDF/mic5205.pdf

With a 9v input, a 5v output, and a 0.7 voltage drop across a diode I believe may be in front of it, the regulator needs to drop 3.3v.

And according to the datasheet, the power dissipation can be calculated as follows: Pd = (125 degrees C - 40 degrees C) / 220 degrees C/W Which gives: Pd = 386mW

(I did the calculation based on an ambient temperature of 105 degrees Farenheit. I figure nobody'll be using the prop on a day which is hotter than 90 or so. Also, it's based on the minumum heat sink.)

So, if we take 386mW, and we divide it by the 3.3 volt drop across the regulator, we arrive at 116mA as the maximum current the regulator can supply if the ambient temperature is 40 degrees celsius.

Even if I assume the ambient temperature is less than that, say, 20 degrees celsius or 70 degrees farenheit, the picture doesn't get a whole lot better.

So... 116mA. That's not much. Subtracting the atmega's current requirements from that leaves me with 66mA. 20mA for the two green leds, and 5mA for the piezo eats up another 25mA leaving me with 41mA. And supplying 10mA to each of the illuminated pushbuttons leaves me with only 21mA.

21mA to drive an array of 64 LEDs. If that's all I've really got, I might as well have not even included the darlington array to sink the current. And I've got a big problem because the bargraphs I planned to use for the array only have 16mcd LEDs.

So what do you think? Are my calculations correct? When I originally designed this thing I thought the Pro Mini reuglator could handle 335mA, but now I wonder if there wasn't a mixup somewhere along the way, and what was actually meant was 335mW.

If I can't get more current out of this thing, my only option may be to buy 64 rectangular LEDs which are way brighter than the ones in these arrays.

Trying to find ways to solve this in case my calculations are correct. So far I've only come up with one potential solution.

It turns out using brighter LEDs may not be an option because there don't seem to be any available. The bargraphs I'm using are rated at 19mcd, and that's as bright as they come. Same goes for discreet LEDs. There don't seem to be many green rectangular leds which are brighter than about 10mcd.

The only other option I've come up with is to use a different voltage regulator. One which can provide the Pro Mini with more current. Or try running it off an unregulated 4.5v. Though I don't know how long 3 AA or AAA batteries would be able to prvide sufficient voltage to keep it powered. I think I calculated recently it could continue to run down to 3.7 volts at 16mhz.

Anyway, even if I use a diffrent voltage reuglator, I'm not sure how much current the Arduino will be able to supply.

I guess I'll pick up a 7805 when I head out to Radio Shack today though.

Done some more calculations. If I assume there's no diode in front of the regulator (and I don't know that there is one) and I take ground current into account (whatever that is... the datasheet references it in its equations) but I assume a best case scenario of running at room temperature, I get a maximum of 108mA continuous from the regulator before it goes into thermal shutdown.

And once again, the ATMega328 itself draws between 30-50mA, so that leaves me with only 58-78mA to play with.

That's less than half what I oriignally planned to put through the LED array alone.

Reading through your calculations, I think they look just about right except for the AtMega which seems high. According to the datasheet this should be approximately 12mA at 16Mhz and 5V. If you estimate this to 20mA (accouting for the two pro-mini leds, on-chip peripheral circuitry, reset pullup and margin) you should still be ok.

Max total current (vcc and/or gnd junction) for the AtMega is specified as 200mA which may also be a limiting factor for your project.

You may very well need both an external regulator and external circuitry to drive your leds and so the mini-pro may perhaps no longer be the best fit given your requirements?

Problem is, the circuit boards have already been made. :'(

I'll find some way to make it work. Shave some mA off here and there. Maybe improve the multiplexing. Give up pulse width modulation, and light one LED at a time instead of a whole row. Feed it 40mA for a few microseconds... You can trick the eye with multiplexing. An led flicking quickly, but brightly looks brighter than an LED lit dimly, but constantly. Was already using that trick, but with less current and leaving the leds on longer.

An led flicking quickly, but brightly looks brighter than an LED lit dimly, but constantly.

That is a common misconception, it is wrong. An LED with 20mA through it flashing with an on / off ratio of 50% looks almost the same as one on continuously with 10mA flowing through it. Try it and see, I have.

Not according to some researchers in Japan: http://www.ledsmagazine.com/news/5/5/11

Also, I have tried it, and have seen results with my 8x8 matrix when driving the leds with much less than 10mA.

It may come down to how many mA and how bright the leds are in the first place. I'm using 10 segment bargraphs with LEDs that are fairly dim to begin with.

Hm... you didn't happen to do your test with a red led did you? The article says that:

Based on an evaluation test using subjects, the group reported that a blue LED looks 1.5-1.9 times brighter while green and red LEDs look 2.0-2.2 and 1.0-1.3 times brighter, respectively.

So if you were using a red led, at most it would only appear maybe 30% brighter. I'm using green leds, so I would have gotten the most difference.

Reading the article some more...

When a short-cycle pulse voltage with a frequency of approximately 60Hz is applied to an LED at a duty ratio of about 5%

So what exactly are they saying to do here?

Are they saying if I have an led I might normally drive at 5mA then I drive it at 100mA, and turning it on and off 60 times a second, or once every 16 milliseconds, and when I have it on, I should leave it on for 5% of that period, or .833 milliseconds? Ie, 833 microseconds?

In oither words, are they saying I should turn the led on for 833 microseconds while putting 100mA through it (or whatever it's designed to handle with very short duty cycles) and then off for 16 milliseconds, roughly?

I guess it won't hurt to give it a shot. I'm updating my display fast, but I think I was simply updating at 480hz, and lighting a whole row at a time. Not trying to force as much current through an led as I could in as short a period as possible.

I use a blue LED when I tried it and comparing the light output / current requirements for this project:- http://www.thebox.myzen.co.uk/Hardware/Econo_Monome.html

Also in my previous company before I arrived the engineers had tried to save current on an RFID reader by flashing the LEDs but it didn't work.

I read the link and I thing the relevant part is:-

As a result, the group discovered that, when a pulse voltage with a frequency of approximately 60Hz is applied at a duty ratio of about 5%, the impact by the Broca-Sulzer effect becomes greater than that of the Talbot-Plateau effect, so that the light emitted from the LED looks brighter to human eyes.

Note they are working at the 5% brightness which is not very bright at all. You then can't go on to extrapolate that result at higher brightness levels like the article did. At those sorts of levels I have seen LEDs glow dimly when pulsed when you would not see them at all with the equivalent continuous current.

There is also a lot of rubbish on that link about discharge tubes:-

"The Talbot-Plateau effect is a principle derived in the days when fluorescent mercury lamps and other light sources driven by a power supply with a longer voltage cycle of about several hundred milliseconds were used," Jinno said.

If that were true then the lights would be perceived to be flickering. 60Hz is just under 17mS at 33 mS they start to flicker. Several hundred milliseconds Lets say 4 of them would see the light blinking at 2.5 times a second, which is clearly rubbish.

Just because someone took the time to create an article on the web... does not mean it's a 100% true fact.

I'm no expert, but I'm falling into the Grumpy Mike camp here. Contrary to popular belief... LED's are not magical devices with unknown properties only Merlin the magician could fathom...

If you want an LED to be brighter, it needs more "current". You CAN exceed the current ratings on an LED if you drive it with very short duration, higher current pulses. Caution: failure to do this right and you ask the LED to live fast and die young.

Well, 5% brightness is 1mA if normally you're putting 20mA through the leds.

And with my 8x8 array, I was lighting one whole row at a time, supplying each led with 20mA.

Divide that by 8, since each row was only on 1/8th of the time and effectively I was supplying each led with 2.5mA, and running them at a 12.5% duty cycle.

So, I'm not far off from where they were running their leds at.

Also just for kicks I looked up the green leds they used: http://www.all-tronic.com/pdf/NSPG510S.pdf

Look like they might be 5000mcd. So they could get away with that 5% duty cycle and still have a fairly bright led. Me, I'm stuck with 32mcd leds because they don't make bargraphs with green leds which are any brighter, and they don't seem to make rectangular green leds which are any brighter either. :/

(Least none I could find on Mouser, Digikey, or Newark.)

Anyway. can't hurt to try it and see if I get any results. :-) It's not like I have many other options.

I'm no expert, but I'm falling into the Grumpy Mike camp here. Contrary to popular belief... LED's are not magical devices with unknown properties only Merlin the magician could fathom...

Of course not. Nobody suggested they do.

It's your visual system which has unknown properties which only merlin the magician can fathom:

http://upload.wikimedia.org/wikipedia/commons/thumb/9/9a/Optical_grey_squares_orange_brown.svg/500px-Optical_grey_squares_orange_brown.svg.png http://en.wikipedia.org/wiki/Optical_illusion

In this illusion, the colored regions appear rather different, roughly orange and brown. In fact they are the same color, and in identical immediate surrounds, but the brain changes its assumption about colour due to the global interpretation of the surrounding image. Also, the white tiles that are shadowed are the same color as the grey tiles outside of the shadow.

Your eyes perform lots of tricks to make the world look the way it does. I've researched this in great detail because last year I was working on a graphics application where I was trying to extract shape and color information from photographs in order to create bumpmaps and color maps for a texture making application I wrote. Since no algorithm exists which can do this well, I was trying to invent one and I was studying how the brain interpets visual stimuli to see if I could copy some of its tricks.

I don't know much about how this led trick is supposed to work, but off the top of my head, I can think of several possible explanations for it. One would be that doing this prevents your pupil from dialating as much, and so it gathers more light. Another is it could be tied to the response times of your rods or cones. Or it could be how your brain interprets the input.

An led doesn't have to actually be brighter to appear brighter.

It's your visual system which has unknown properties which only merlin the magician can fathom:

Very true - here are two inexplicable examples

Heh. :-)

This was an older thread, but here's a more recent one with a schematic: http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1277880350

The problem is I have an 8x8 array of leds which are multiplexed. I can't simply connect the LEDs to the raw supply. If I did that, how would I index the rows and columns? (Keeping in mind, wiring up a ton of transistors and making a mess of the board isn't an option.)

Anyway, I've already found a solution to the problem as outlined in the other thread. Cut the trace leading to the Raw pin, run a jumper from a +9v PTH to a Vcc PTH, and then attach a voltage regulator to my 9v battery pack and connect that to the 9v pins on my board. Then I'll have a 5v regulated source of power going directly to the Vcc pins, and I can still switch it with my illuminated pushbutton switch connected to the points on the board where it was supposed to go.

I'm glad I added all those extra PTHs just in case. :-)