5v dc-dc converter

ok smps voodoo doctors and gurus, I need some opinions

my goal is to build an smps that can supply 5v at 2a, ultra-low ripple is not a major concern, the primary goal will be to provide charging current for digital devices with on board charging circuitry.

the source of power for the converter is two or three 2.6 ah lithium cells (rated for high current charge and discharge), ideally in parallel for easier recharging, which means a step-up converter

i don't mind smt, I prefer it actually, but insanely small packages like micro-sop and micro-max are too small for my fabrication abilities, so something in an so, tso or sot package is ideal. looking more for a integrated converter instead of just a controller, but maybe a controller is the only option.

National has a couple of chips that come close, but don't quite fit.

please share your suggestions

I suggest a SEPIC converter with a controller by Linear Technology. I used one to generate a 5V supply for a thermal printer from a camcorder battery. It worked very well but layout is critical on the high current paths.
https://www.linear.com/pc/productDetail.do?navId=H0,C1,C1003,C1037,C1078,C1089,P1517

Use a TL5001.

Linear always has interesting parts:

https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1042,C1031,P1795,D1485

Also, TI has one part with a internal switch that might be able to hit that, but it's a .65mm pitch part, and it's at the edge of it's operational area...TPS61032...

Having 2 batteries in series and doing a buck regulator is probably the easiest way of doing things though. The Lithium batteries are effectively dead at ~3V, so the circuit would have to try to pull ~4A out of the battery near the end of life, which is a bit tricky for parts with internal switches. However, if you happen to want to charge from a 5V power supply, there aren't many choices...

Hi

You could try MC34063A available in 8pin DIP and SO08

Datasheet from www.Alldatasheet.com

mvs sarma said:

Hi

You could try MC34063A available in 8pin DIP and SO08

Datasheet from www.Alldatasheet.com

Click to expand...

We use these parts in in both the buck and the boost configurations at about the 1 W power level. They are robust, reliable and manufacturable. A great part.

Papabravo said:

We use these parts in in both the buck and the boost configurations at about the 1 W power level. They are robust, reliable and manufacturable. A great part.

Click to expand...

Please see the ST datasheet from www.datasheetcatalog.com, where you can see external transistor supporting in darlington mode thereby delivering higher power output.

anyone have experience with the LM2576 'Simple Switcher' from National or the MIC4576 Switcher from Micrel

I really like the low parts count of these units, and the fact they don't require ceramic / tantalum caps. They both come in a friendly TO-263 package (or to-220)

It seems the buck topology makes a lot more sense for the power levels involved (ignoring more complex designs like sepic). So this means two or three cells in series. Next question, whats a simple (and relatively safe) way to charge these cells while they're in series?

I've used a part in the LM2575 series - no real complaints- there aren't many knobs to tweak(in the wrong direction), so they work as advertised.

1) battery protection is a good idea in any case - these will disconnect the battery from the circuit if charge gets too low, or worse, if the cells get too badly unbalanced.
2) charge balancing is needed when you start having cells in series - this can be implemented with a microcontroller + a couple pass transistors. Essentially read the voltages and selectively siphon a couple 10's of mA from batteries that have higher voltages than the others.

I've seen recommendations that ~2 cells is reasonable to use without balancing circuits, but that the more cells you have in series, the more the balancing circuits are desired.

Of course the low-tech, brute-force method is to use a relay to change the battery configuration as needed...