Does Regulation Pay With Rechargeables?

[jayflash]

Given the relatively flat discharge curves of secondary cells, is a regulator worth the expense and added complexity? Certainly it wrings the most from primary cells but their voltage tends to sag more with a heavy load and as they are depleted.

It seems that a simple resistor would allow aproximately 90% of the run time of a regulated light without more components and junctions to fail. Wouldn't the proper value resistor allow an LED to operate at a steady output for most of a rechargeable's capacity?

What are your real world experiences?

 

[evan9162]

Even though you get relatively flat voltage across the life of a rechargable, you still have a high voltage spike with freshly charged batteries.

For example, freshly charged NiMH may be 1.4-1.45V (1.2V nominal), fresh Li-Ion are 4.1-4.2V (3.7V nominal). You still must deal with the initial higher cell voltage. So you either set your resistor to properly drive your LED with the nominal voltage, and accept higher current for the first bit of battery life (possibly overheating and/or damaging your LED), or you pick a resistor that safely drives the LED with the high initial voltage, and suffer lower current later. For alkalines vs NiMHs, you pretty much have to deal with the same voltage range.

In addition, Vf changing due to temperature and life of the LED can also change the set current.

A linear regulator (if that's what you're referring to) satisfies all conditions. You don't have to worry about the initial higher voltage of rechargables, nor do you have to be concerned with Vf changing. When the input voltage is less than about 2V above the Vf of the LED, they can have great efficiencies, as well as providing flat output. They do best when the input voltage is just above the Vf of the LED, there efficiencies can be better than 90%.

Buck regulators go a step further and get you a little more battery life (by drawing less current when the battery voltage is higher) at the expense of higher cost and complexity. However, some buck regulators don't work well when the input voltage is just above the Vf of the LED, since some don't allow the switch to hit 100% duty cycle (just a technical detail there). Mainstream buck circuits are in the high 80's for efficiency, advanced in the high 90s.

So it depends on what you're after. Personally, I like my lights driven by a regulator, even when being driven by rechargables. Both of my Lux V lights will fall out of regulation when even NiMH cells are over 80% discharged, so when they start to dim, the batteries are dead, and it's a sign to change batteries.

 

[RussH]

My primary reason for not using regulators (as opposed to a resistor) is that they require voltage head room. This isn't so good in a flashlight since we need near 3.6v for a lux3 which means that any head room at all requires another battery. I use a lot of LM317 & similar chips in 6v and 12v applications (CV and CC) where I normally have plenty of extra power. I prefer a resistor in most other cases. It's not a perfect solution, but works for most situations IMHO.

 

[evan9162]

LDOs can be easily constructed to have 0.1V of headroom or less. In fact, you have less headroom with an LDO over the life of the batteries than you do with a resistor.

For example: You're driving a Lux3 at 700mA from 3 cells. You choose 0.5 ohms of resistance. The resistor will be dropping more than 0.1V down until the current hits 200mA, which would give a Vf of about 3.2V, so the batteries are at 3.3V.

However, the LDO will drop 0.1V at the point of regulation (Vin=3.7V), while maintaining 700mA. At that point, the dropout voltage gets even smaller, and the light is, for all intents and purposes, direct driven.

With the resistor, when the batteries are at 3.7V, only about 450mA of current flows.

So the regulator has given flat output to that point, whereas the resistored light's output has diminshed from the point the light is switched on.

 

[andrewwynn]

besides the amazing difference in current with a tiny difference in voltage with LEDs... 'flat' discharge is just about a myth on batteries of any type.. the only way to get truly flat discharge is to have waaaaay more capacity than you need... hook up 3 LEDs in series to a 12V car battery and they will not change in brightness for a day or two... but it's a little heavy to carry around.

the headroom thing and efficiency with LDO is unbeatable.. i'm getting over 93% efficiency with my LDO arrangement on my little lights.

re 'headroom'.. so does a resistor. If you didn't already HAVE headroom there is no need for a resistor. I am completely thrilled with the results i've gotten with the LDOs i've made.. the one i made last night i measured less than 0.09V drop on the LDO at 700+mA.. that's the equivalent of a 0.13ohm resistor... or about that of a bit of wire or a dirty switch. In other words, the 'beauty' of LDO is that when it's done 'regulating' it basically becomes a piece of wire... the resistor on the other hand always costs a similar percent of the power.. where the LDO gets more efficient at the battery drains.

-awr