Which is the most efficient way to get 5v

[limbo]

Hello community.

I need your help in the following project:
I need to create a high efficiency power 5v supply circuit, or better DC to DC converted. The problem is that i'm going to use as source a 12V Lead Acid battery and I want to have the maximum possible efficiency in order not to loose energy.
So using simple solutions such 7805 regulators is out of question!
The output current needed is up to 1000 mA (1A) which is not so much.

Any ideas?

[Hideki]

If you intend to build to circuit yourself, use a simple switching regulator like the LM2675 or one of the many alternatives.
http://www.national.com/pf/LM/LM2675.html#Overview

Try the webench tools to find something suitable.

[Psi]

yeah, there are plenty of switchmode ICs available.

Maxim sell some super high efficiency ones if you need that.

The key things to be aware of when making a switchmode supply are.

• Use low ESR caps.
• When selecting an inductor make sure its value was tested at near the frequency you will be running it at.
  eg, if you need a 100uH inductor for your 200khz switchmode circuit you want to check that it is 100uH @ ~200khz. Some inductors for sale will be intended for other uses and may only be 100uH at 1khz etc..
• The diode needs to be a fast schottky type. The correct diodes will normally say in their datasheet that they are intended for switch mode power supplies. And the datasheet for your switchmode IC will often give recommendations of diodes that will work.
• PCB/wiring layout needs to be what's called 'star ground', basically it means that all ground tracks should come from one central point (the input capacitor). This means each area of your circuit which needs a connection to ground gets its own dedicated path.
  If components shared a single ground track and one area of the circuit used lots of current (which happens in dcdc) then there would be a temporary voltage drop across that pcb track. This is bad if anything else is using that same ground track since it sees the voltage drop too. It's important that the feedback system doesn't see voltage drops generated from switching currents.

[KJ6EAD]

Dave has done a tutorial on the design of a switch mode converter. I'd give you the link but it's too hard to find in the video blogs episode list.

[johnboxall]

Dave has done a tutorial on the design of a switch mode converter. I'd give you the link but it's too hard to find in the video blogs episode list.

http://www.eevblog.com/episodes/

http://www.eevblog.com/2010/09/10/eevblog-110-lets-design-a-dc-to-dc-switchmode-converter/

[Psi]

Dave has done a tutorial on the design of a switch mode converter. I'd give you the link but it's too hard to find in the video blogs episode list.

http://www.eevblog.com/episodes/

http://www.eevblog.com/2010/09/10/eevblog-110-lets-design-a-dc-to-dc-switchmode-converter/

Yeah

Although, the IC he uses isn't exactly the easiest to use, it's extremely common and cheap though.

limbo,
If you want to learn about dcdc and how to use a chip that's very universal then that video is excellent. But if you want to wire up a quick regulator without doing any math then it would be easier to get a fixed stepdown 5V 1-2A dcdc IC which has the oscillator freq and output voltage built in.

[AlsInd]

you are correct that the normal regulators are going to be a poor choice for efficiency. what you are looking for is called a buck dc-dc regulator. if you search buck regulators you will find a lot of info on them. there are many chips available. one of the blog videos specifically addresses dc-dc converters (#110) and gives good general information. generally, the higher the frequency of the reg chip, the higher your efficiency will be. i think you will find, however, that 1A is not considered all that low power when it comes to a big chunk of regs out there. you have to understand that many of the chips are designed for hand held products like phones, calcs, etc, which do not have a very large draw. since you have 1A requirement, i would include this in the parametric search first. i have done quite a bit with dc-dc power applications and approach it a little differently from Dave in the video. i only use distributors to find my chips if i am unfamiliar with which company to go to. so for power applications i have favorites in the industry. in order my favorites are: http://www.linear.com/index.php *** http://www.ti.com *** http://www.maxim-ic.com ***  http://www.national.com

there is no easy way to do this... you have to go through the chips and look at their efficiencies at your voltage and current levels and see what works for your application. keep in mind that higher frequency will allow for smaller inductor which can be a large component at low freq. levels. generally, if you have a choice opt for the higher freq. chip.

also expect to get poor results from your first prototype. it is unlikely that you will get efficiencies advertised by chip manufacturers (usually just BS). switching psu design is somewhat of an art and thus takes patience and perseverance to learn. designing a really great dc-dc power system is not easy which is why some of us get the big bucks for designing custom power systems for major companies.
cheers :-)
al...

[Neilm]

also expect to get poor results from your first prototype. it is unlikely that you will get efficiencies advertised by chip manufacturers (usually just BS). switching psu design is somewhat of an art and thus takes patience and perseverance to learn.

I have usually been able to get close to the efficiencies claimed by manufacturers (within 1% or so) provided you do the tests under similar conditions.

The parts that really cause the problems is the layout. Always review the app notes, design guides etc to find the recommended layouts. It only takes a relatively minor change to the layout to have a dramatic effect on the efficiency.

Yours

Neil

[Zero999]

A simple circuit can be made from a couple of transistors, see the site linked below. You'll obviously need to use beefier components to get 1A. The one shown in the schematic is only really good for 0.5A but it shouldn't be hard to design for 1A. You could also replace Q1 with a P MOSFET, providing other changes are made and the input-output voltage differential is large enough to ensure an adequate gate voltage.

http://www.romanblack.com/smps/smps.htm