DC-DC converters

[machine88]

im trying to build two dc-dc converters (buck & boost) using breadboards and the input & output voltages range from around
0-15 V. I have tried using MC34063a but its very inefficient, I want efficient of around 90%. My question is are there any other ways I can improve the efficiency using that same chip (mc34063a) or do you know any other better chips that i can use on breadbroad. Any help you provide will be very appreciated. I want to use the converters in small application like automated meter reading.

[johnwa]

Depending on the circuit you are using it in, you should be able to get reasonable efficiency from this IC. All of the components will affect the efficiency, not just the chip. Are you using the internal switching transistor or an external one?

I find a good way to optimise for efficiency (once you get into a power class over a few watts) is to find which components are heating up. Then ask yourself why they are heating up and how can you stop them from doing so. If the inductor is getting warm, can you use one with a lower DC resistance? Or is it saturating? Losses in the switching transistor could result from Vce_sat/Rds_on being too high - could you use a part with better specifications? If the losses are from switching, rather than static, can you provide a better drive signal? Or perhaps redesign to use a lower switching frequency?

[notsob]

Also have a look at DC DC converters from http://www.monolithicpower.com

you can even bag a handful of MP2307 based assemblies on ebay  & modify them for you own situation -

 http://www.ebay.com/itm/10-pcs-MP2307-DC-DC-step-down-power-supply-module-3A-/270900254413

[Neilm]

You could be having problems dependent on the layout. I have found that even minor changes can have a very large effect on the efficiency of a SMPS.

Neil

[mazurov]

You won't be able to make efficient switcher on a breadboard. You need to actually solder the thing and be very particular about layout. At the very least, solder inductor and shottky directly to pin 1 as close as possible. Then, solder input cap directly to the other side of the inductor and output cap to the other side of the shottky. Then, solder ground sides of input and output caps to pin 2. The directions are given for the boost topology, buck is similar - switch node and power ground returns should be short and fat.

[Jeff1946]

Remember the Schottky diode is dropping 0.3 to 0.5 V, depending on the current.  Unless you go to a circuit using a synchronous rectifier, this will be a significant portion of your wasted power.  You may want to experiment using an external transistor to switch the power.   At higher voltages you can use an FET,  I have used an inverting FET driver such a Claire IXDF604PI to get fast switching.  Slowing down the FET reduces noise at the expense of efficiency.
 

[Psi]

To get better efficiency
- Confirm your switching inductor is at the 'idea' value for your circuit, frequency and load. Also check it's able to handle the load current without loss.
- Select a schottky diode with the lowest Vdrop or put multiple schottkys in parallel (5 or so). Note that they will not share equally so its best to spec each diode so it can handle all the current. With less current through each diode the Vdrop will be less. (Or you can look at using a fet instead of a diode as mentioned above)
- Select a switching fet with the lowest Ron but still adequate for the other specs
- Arrange the system using a star-ground
- Make sure your capacitors are ultra low ESR and not just low-ish.

[titam]

Forget the breadboard.  Get an evaluation kit with an optimized layout and hack into that.