stepping up a 3.3 supply to 5V, I<50ma

Hi - I just discovered on a set of boards that I'm working on that the microcontroller that I'm using (Atmel atmega168) requires a 5V supply for full speed operation. I had planned on giving it 3.3V so now I have to figure out how to step up the 3.3v supply to 5V. 3.3V line is very low noise, as it was originally designed to be driving this microcontroller. Right now - the best solution I've been able to come up with is the Microchip MCP1252/MCP1253:

It steps up 3.3v to 5V up to 150ma, which is more than enough for my needs. It is an 8-MSOP package and needs three ceramic capacitors. It is specced for 2.5% accuracy and 45mv ripple. It's not perfect, but I think it'll probablay work.

My question is this: Are there any better solutions out there? Cost is not a worry, as this is a very low volume application. Size is my biggest worry as I was already running out of space on this board before I discovered that I need a 5V supply.

Thanks,

-Mike

From what unregulated supply are you deriving the 3.3V? You say the "3.3V line is very low noise" and you are considering hanging a boost switcher onto it? I'd wonder why not pick off from the same source that your 3.3V uses as input and start from there to derive a 5V.

Jon

"M. Noone" wrote in news:1142375332.943151.128550 @i40g2000cwc.googlegroups.com:

Please ignore the above quoted post (as it was accidentally posted twice). Google Groups sucks. I'll be sticking with ol usenet from now on.

-Mike

Jonathan Kirwan wrote in news: snipped-for-privacy@4ax.com:

have

and

H Jon - the board has coming into it two supplies: 6V and 3.3V. The 6V and 3.3V lines are coming from a main power/control board which has two large DC/DC switchers on it that are connected to a li-poly battery pack. But these boards that I'm working on now do not have access to that battery pack, only to the 3.3v and 6v lines.

The reason I didn't mention the 6V line is because my expectation is for it to be very noisy, as it is powering 18 motors. Also, I really would like to keep the logic seperate from this 6V supply as I like to be able to turn off all the motors without interrupting any of the logic circuits. Thanks,

-Mike

Hello Mike,

Does it matter if the 3.3V line is not so low noise anymore after adding a charge pump? These pumps can generate quite some current spikes on the input side. But if you absolutely can't free up the space for an inductor maybe you are between a rock and a hard spot.

If you can afford an inductor TI, National and others offer switcher chips in SOT23-5.

Nice, sometimes I wish that was so on my designs :-)

Regards, Joerg

Joerg wrote in news:%VKRf.53969 $ snipped-for-privacy@newssvr13.news.prodigy.com:

adding

the

Using an inductor is not an issue - but in my experience inductors are normally quite large. Is this not the case in low current situations like this? Can you reccomend a chip that might be good for this application? Thanks,

-Mike

Hello Mike,

If it gets switched off while your logic must remain running that rules out the 6V. But noise? I had pretty sensitive TTL stuff running in my car back in college. This car had a 6V electrical system, a brush cummutated alternator with a temper and no (!) battery. 6V batteries had become rare and it would have cost me the equivalent of three crates of beer so I decided for the beer. Oh yeah, and of course this car had to be hand-cranked which meant that it could send the occasional big spike when the engine kicked back. Then there was the trucker's horn with a huge air compressor that sucked 20 amps or so, only to be used above

Regards, Joerg

Hello Mike,

They don't have to be large. In a boost converter you only have to store enough energy in it to add (Vout-Vin)*Iout. The higher the frequency the smaller the inductor and nowadays well over 0.5MHz isn't a big deal anymore. The TI TPS61070 series is one example. There are tons of them and competition is quite fierce.

I haven't used these yet since I usually roll my own for cost reasons. That cuts about 50c out of the BOM cost but requires external SOT-23 FET and diode. The TPS series generally doesn't.

Regards, Joerg

For this type of applications, I used to use a simple switcher from ST: ST5R50 It's low cost and SOT23. A small inductor, for example from Bourns:SDR0604-470KL, two ceramic X5R 22uF capacitors, and a fast schottky diode (SS14 or smaller) complete the design. No transistor needed. You could find also pretty small low cost switchers in On Semiconductor (NCP1450A with shutdown).

Regards.

Yes, pretty much every semi mfg has this line. It's almost like bread and butter at the supermarket. Best is to pick one that is form-fit-function compatible with one or two competitor parts. That might take some data sheet studying to find.

Regards, Joerg

How do you get the 3.3V?

Because:

If you already have an on-board step-down switching supply you might be able to add a fly-back winding on the filter inductor. The auxilliary 5V voltage will track the 3.3V but not so well; it depends on the ratio of power between the 3.3V and the 5V.

Obviously, the 3.3V should be the dominant one.

This would be fantastic. Tell me if you manage to find two compatible circuits. It all seems like a competition between manufacturers to find new original ways to use 5 pin in SOT23.