How to generate less heat in circuits ?

[elementHTTP]

How to generate less heat from 12v source whit regulated output of 5v which draws 3A
lm350 ? LT123A ? pwm ?...etc

Is there any way (technic ) to avoid so much waste heat (Q≈I^2 * R ) and not to require heat-sink ?
thanks :D

 

[vk6kro]

If you use a linear regulator, there will always be 7 volts across it and 3 amps flowing in it, so that is 21 watts. It will get hot.

However you can get switched mode regulators that are a lot more efficient.

Just as an example, have a look at this one:
http://www.ebay.com.au/itm/DC-DC-Step-Down-Adjustable-Power-Supply-Module-/110723383063?pt=AU_B_I_Electrical_Test_Equipment&hash=item19c7a0bf17

This one has a maximum current rating of 3 amps so you could look for one that gives a bit more than this as a safety margin.
Just hunt on Ebay for this:
"Adjustable Step-down Power Supply Module"

I would not suggest you try to build one when you can get these for less than $5 already designed, built and tested. And delivered.

Just a caution, though. Switched mode power supplies can produce radio interference, so you would need to test for this before you use it.

 

[Packocrayons]

They use switching BEC's in the hobby industry as well, you can get them up to 15A
http://www.hobbyking.com/hobbyking/store/__14310__HobbyKing_UBEC_3A_2_6s_LiPO.html

 

[elementHTTP]

Thanks for answers :D

BEC is way to go

what do you think about this circuit ?
http://www.coolcircuit.com/project/bec4a/bec_4a.html

I wonder what power supply dose NASA Mars rover use ?

 

[NascentOxygen]

Thanks for answers :D

BEC is way to go

what do you think about this circuit ?
http://www.coolcircuit.com/project/bec4a/bec_4a.html

That coolcircuit project uses a linear regulator IC. That is not what you want; a linear regulator will get hot, hot, hot. For efficiency you need a switching regulator. The way it works is to rapidly switch the battery in and out, taking only what power is required by the load, and leaving the remainder in the battery for later use.

The type of HobbyKing module pointed to by Packocrayons looks ideal. (Take heed of the caution by vk6kro about the propensity for creating RFI, and include input/output filtering and shielding. Perhaps include it in a closed metal box.)

The modules on ebay that vk6kro pointed out should likewise be suitable.

 

[elementHTTP]

wow answers come so fast :D

Hmm
Can then switching regulators be used in parallel to boost current output ?

 

[NascentOxygen]

Can then switching regulators be used in parallel to boost current output ?

http://imageshack.us/a/img824/5512/glitterno.gif That would not be advisable. Extra complexity would be called for to ensure they shared the load & worked in harmony.

 

[elementHTTP]

That would not be advisable. Extra complexity would be called for to ensure they shared the load & worked in harmony.

Its like trying to juggle while ice skating :D

But what about heat in H-bridge is there better way to go ?
- example circuit 10A
http://www.pyroelectro.com/tutorials/h_bridge_4_transistor/index.html

 

[yungman]

Unless you want to take on the challenge of designing a switch mode regulator, buy an off the shelf one. Don't monkey with parallel or something, that is a very bad idea, most have closed loop feedback inside, you put them in parallel, bad things will happen.

You are going to need heat sink if you use linear regulator for 3A. If you want to get something going fast instead of spending more time talking here, get yourself 6 to 8 big diodes that is 5W, put 7 in series from +12V to drop 0.7X7=4.9V before driving the regulator, so the regulator only drop about 2V and that would be 6W or so. You are going to need heat sink, but at least is only 6W, not 21W. Cheap, dirty and quick, much quicker than building a switching regulator...BUT slower than buying one off the shelf.

Just go buy one off the shelf!

 

[NascentOxygen]

If you want to get something going fast instead of spending more time talking here, get yourself 6 to 8 big diodes that is 5W, put 7 in series from +12V to drop 0.7X7=4.9V before driving the regulator, so the regulator only drop about 2V and that would be 6W or so. You are going to need heat sink, but at least is only 6W, not 21W.

This arrangement will dissipate exactly as much heat as the original wasteful, hot circuit elementHTTP wants to get away from. The only difference being that with added diodes, those diodes waste some of the heat, leaving less for the IC's heatsink to dissipate—though the total amount of heat produced remains unchanged. An enclosure containing the PS electronics will get just as hot; the battery will last not a second longer.

 

[yungman]

Right, just allow to have a smaller heat sink for the regulator. Each of the diode dissipate part of the power and don't need heat sink.

As I said at the beginning and at the end, buy a switcher off the shelf. But OP kept asking to the point I felt he want to build one. Building a switcher is never easy even if you follow the exact application circuit. I know, I had to mob up another designer's mess on two LT switcher circuits that sang like a bird. The circuits were built as the application notes. You have to be careful with application circuits, the closed loop feedback on those switchers were not tamped! This is not exactly for hobbyist.

And yes, for people that think you get one of those switcher IC , build a quick little switching circuit with the application notes. Yes, they even give you the parts list as if you just slapped them together and it'll work...Think again. We had three of those in a video camera and the image had lines on the screen. It finally dropped on my lap after a year and two other engineers working on it. It was just that two little switchers. No body suspected that! One year, must be over $100K wasted for that two little switchers.