Dc-DC Converter required

Thread Starter

aamirali

Joined Feb 2, 2012
412
I need a Dc-dc converter with:

1 input voltage : 6.6-8.2V
output voltage : 6V.
O/p current - 8A. (max i need)

2 input voltage : 3.3-4.2V
output voltage : 6V.
O/p current - 8A.
 

steveb

Joined Jul 3, 2008
2,436
I"m going to infer/suggest that you want to avoid a linear regulator with currents up to 8A... unless of course you want to have this circuit double as a space heater.
That's probably good advice, and he seems to be asking for that anyway. Still, it's not a clear choice if other system requirements are factored in.

We would be remiss if we didn't point out that 6.6 V-8.2 V converted to 6 V with a linear regulator (low-drop out type) implies an efficiency range of 73-90 %, while a good DC/DC converter might give the 90% number over the whole range (it might also give less if care is not taken). But, 73 % is still not all that bad and the 17% additional loss is only another 10W or so of heat dissipation, which is managable technically (need heat sink, but it's not quite a space heater) and not expensive (0.2 cents per hour of operation extra cost). Also, a relevant question is how often will the operation be at the maximum voltage and current of 8.2 V and 8 A. So many times we've seen systems with maxium specifications that are never realized in actuallity.
 

Stuntman

Joined Mar 28, 2011
222
steveb,

Yes, you are right, he is not looking for a large voltage drop and definitely could go the LDO route (for 1). My intention was to be humerous with the space heater.

But more importantly, I was under the impression many buck regulators were well over 90% efficient? Perhaps this isn't true. That said, I've only seen a few linear regulators that handle that kind of current, which usually implies a meaning, hence my suggestion.

I did notice, however, my link did not include Linear's boost converters for the #2 requirements.

http://parametric.linear.com/switching_regulator
 

bountyhunter

Joined Sep 7, 2009
2,512
I need a Dc-dc converter with:



2 input voltage : 3.3-4.2V
output voltage : 6V.
O/p current - 8A.
That one one will be very difficult to build: About 60 - 70W drawn off a 3V source? Huge peak currents on the input devices. One of those cases where it's theoretically possible but ridiculously difficult to build. It's not a feasible design, you need to get a higher source voltage to work from.
 
Last edited:

bountyhunter

Joined Sep 7, 2009
2,512
I need a Dc-dc converter with:

1 input voltage : 6.6-8.2V
output voltage : 6V.
O/p current - 8A. (max i need)
You definitely want to use a linear reg on this one, a discrete LDO using parallel N-FET devices to get the on resistance low enough. Max power dissipation is about 18W (8.2V in) which is manageable with a decent heatsink. You will need to use some kind of charge pump circuit to generate a bias rail of maybe 9-10V to run the control circuitry and be able to bias up the N-FET gates.
 

bountyhunter

Joined Sep 7, 2009
2,512
steveb,

Yes, you are right, he is not looking for a large voltage drop and definitely could go the LDO route (for 1). My intention was to be humerous with the space heater.

But more importantly, I was under the impression many buck regulators were well over 90% efficient? Perhaps this isn't true.
Buck reg efficiency goes to crap when the input voltage drops near to the output. In those cases, the switch is on almost all the time but you still have switching power losses and bias/driver circuits eating power. I have actually seen buck regs with efficincy LOWER than an LDO when the input voltage is near the output.
 

steveb

Joined Jul 3, 2008
2,436
But more importantly, I was under the impression many buck regulators were well over 90% efficient? Perhaps this isn't true. That said, I've only seen a few linear regulators that handle that kind of current, which usually implies a meaning, hence my suggestion.
Good points.

Yes, some are 92-93 %, although this is usually the maximum number under special conditions. It's not always easy to get these numbers, and one might end up with 85% (or even lower) under the conditions that are most critical, or most typical. This still might end up being the way to go, but I just wanted to point out that it's not a clear decision unless the full constraints are known.

You are probably correct about available high power regulators being not too common. I'm not sure myself because I never tried to buy one. When it comes to higher power requirements, I prefer to design my own solution to have more flexibility and to not get stuck with an unusual "sole-source" device. Doing a good linear solution is not necessarily any easier than a switching approach, in this case. However, depending on the full specifications and constraints it might lead to a better overall solution. (or, it might not!)
 

bountyhunter

Joined Sep 7, 2009
2,512
You are probably correct about available high power regulators being not too common.
It's because die size drives the price out of sight. There used to be a decent selection of 5A LDO regs around and one 7.5A which was REALLY pricey. Not many high current linears around anymore because nobody will pay $10 for a linear reg when you can get discrete FET's for about $.25 and a controller IC for maybe about the same ballpark and build your own reg. Check out the LP2975 from NS.
 
Top