Regulation question - why not *flat* output?

[Darell]

I'm a big fan of regulated flashlights, and hope that more and more mfg's offer this great feature. The whole process is still somewhat of "black box" technology to me but I know there are many folks out there who understand this stuff far better than I.

Why can't a regulated flashlight put out constant light until the battery just can't do it any longer? Even the best of the current crop of regulated lights drop off slowly in output as the battery dies. Shouldn't it be possible to keep consuming the dying battery faster and faster to keep the output curve dead level until the end?

Just curious if it is possible. I'd love my flashlights to have 100% (while taking the regulation losses into account, of course) output for as long as possible, then to have them snap down to moon mode to burn up the remaining voltage.

Thanks in advance...

 

[vcal]

Very interesting subject-keeping in mind that, as Gransee says: 40% of the battery capacity still remains @ .9v

 

[Gransee]

Hey Darell, that's a good question!

Regulators must have an input voltage higher than the target output voltage in order to maintain regulation.

As the battery dies, its voltage drops faster than is remaining capacity.

Here's an example:

-6 volt lithium battery (2, 3v coin cells)
-regulator set at 3.7 volts to run single LED
-minimum input to regulator to maintain set voltage is 4.2 volts (a good regulator).

Several factors to add together:
- Once the battery voltage falls below 4.2 volts, the output can no longer be maintained. The regulator "sags" and allows the output to drop below 3.7 volts.

The regulator and LED put a load on the battery of course so even with fresh cells, the starting voltage may be say, 5.2 volts (beefy coin cells!) but then drop down to below 5 volts in a couple of minutes.

After an hour or two, the battery voltage under load would drop below the required 4.2 volts but battery would still have around 50% capacity remaining. Alkalines are especially good at hiding remaining capacity. Now the light has stop regulating but continue to produce a dimmer light. Depending on the regulator, the transistion from 4.2v to 4.1 volt may produce a dramatic drop off on the LED side.

So now you have a light that regulates fine for a couple of hours, drop off and then slowly dies.

Adding an inverter improves the light further...

A typical inverter can output 3 volts with 0.9 volts going in. Of course the million dollar question is, at how much current?

The problem inverters run into is that in order to maintain a given output they have to draw more and more current from the battery as it's voltage drops. Batteries are willy, remember they drop their voltage faster than their remaining mAH capacity. Soon you run into the minimum input for the inverter to function. A lot of off the shelf inverters don't work below 0.9 or 0.7 volts. So the battery still has 30-40% capacity remaining but the inverter has shut down.

The Arc inverters run with inputs well below 0.7 volts (the actual number is a trade secret). So the battery is pretty dead before the inverter quits.

Now even though the inverter can run down to a pretty low voltage, it's output voltage varies with the input voltage. That is why a regulator is usally attached to the output of an inverter.

So say you have an inverter than can produce 4.5 volts from 1.5 volts in. You regulate it down to 3.7v and all is fine. But the battery voltage drops quickly like they love to do so the inverter output drops quickly too. Soon, the regulator drops out and now the light is in moon mode.

Peter

 

[**DONOTDELETE**]

A good step-up regulator circuit could do just that, if batteries could be called upon to supply any amount of current instantaneously.

Unfortunately, as real-world devices, they cannot.

As battery voltage declines, a regulator demands increasingly higher current from the battery in order to keep output current constant. At some point, the battery will simply not be able to satisfy that demand, and output current will start to wane.

Note that not all DC/DC converter circuits are regulators; true regulators often demand extra components, each of which introduces their own loss, thus lowering efficiency.

 

[Darell]

Thanks for the great, quick responses.

OK, we're on the right track here. For fun, let's just leave efficiency out of it for now. That part can always be improved...

It makes sense that the input voltage needs to be above the desired regulated output voltage - but in the case of single and double cell LED lights, that is typically handled by a step-up converter, right? So we can, in theory, make that input voltage whatever we want already.

So, in Peter's example, let's say we set our regulator at 3.7 V. It will take an amount of time before the batteries are sucked down to or below 4.2 V. So why is it that this theoretical single LED light begins lowering its output from the first moment it is turned on? That's the part I still don't get. Regulated lights dim *slower* than their non-regulated counterparts, but they still begin to dim immediately.

We know that V will sag as current begins to flow, but shouldn't a regulator be able to compensate for that? I just love asking questions on subjects that I know very little about. Makes me feel all-powerful and stupid at the same time

 

[Gransee]

Duggg's explanation was so much simpler and easier to understand than my rambling explanation. That's what I get for talking on the phone and posting to the CPF at the same time. Of course it's better than driving while posting on the CPF at the same time. Or drinking while post...

Ok!

Another good question Darell. You'll notice that the Arc inverters don't actually regulate that well. Its what I call "partial regulation". It is actually regulating so we can say that in our literature. It is just not perfectly regulating. The reason for it is because of a compromise. Once you see why, I think you would agree is a good trade off.

If you see a regulated flashlight with a perfectly flat discharge curve, it's holding back. ie, it not running as bright as it could. For a regulator to work, it must be supplied an excess to work with.

There is only a certain amount of juice in a AAA cell. There is also a sweet spot for the most efficient inverter that will fit in the small housing (we actually were quite fortunate with the Arc-AAA design).

In order to flatten the discharge with real world components, we would have to dial back the regulator. This would mean a dimmer light. This is what we did with the XR series.

But we choose to run the parts flat out to make the light as bright as possible. This means the discharge curve is somewhere between where a normal flashlight and a perfect regulator flashlight. The difference is in the brightness.

Eep! The investers want their reports and I got to get back to work...

Peter Gransee

 

[papasan]

it's harder to consume a dying battery than a fresh one...basically, as the battery discharges it's voltage drops (faster the discharge the more it drops)...this is why an unloaded battery will measure at a higher voltage than a loaded one...so to get the same wattage output as the voltage drops you need to get more current (amps) out of it...the formula is 'volts*amps=watts'...

different batteries handle discharge differently...even within alkaline batteries there's a large difference...the lowest level, or cheapest' alkalines store alot of charge but, because their internal resistance is so high, cannot put out very much amperage and drop their voltage faster while the newer generation of alkalines can put out more current because of lower resistance but hold less total mAH (just as a general rule)...NiMHs, on the other hand, can put out a good deal of current (over an amp) so they hold their voltage for a longer period of time compared to their total mAH...lithium batteries also have a lower internal resistance and can put out more current for longer...

so basically if you have something like the Arc-AAA that draws relatively little power (output is probably about 40mA at 4V which is .16 watts which means that draw off the battery is about .13 amps or so (did i get all my math right?)) with new alkalines and 80% efficiencey...even the lowest end alkalines would be able to supply this much current throughout most of their life...

then you have the 2-cell LS driver circuit...the LS wants to see 350mA at 3.2V (for a white or blue) so drivin off of 2 fresh alkalines they would be putting out almost half an amp at 80% efficiency...at this output the voltage would drop rather fast, especially for older 'hi resistance' alkalines...then as the voltage drops the circuit tries to drain the batteries faster...at 2.4V (2 cells at 1.2V each) they're trying to put out .56A, at 2V .672A...

also, with a step-up regulator the circuit switches 'on' and the batteries charge a coil (or capacitors depending on the circuit type), until the voltage (or amperage depending on the circuit) reaches a certain level then it discharges the coil...all this is trying to achieve the average output required (like 3.2V for a white LS) so that all the peaks and valleys of the charge/discharge cycle average out to your output...so in reality the battery goes through a wave with the top and bottoms at probably like zero to double of the average draw...so, in the last example, the battery is probably seeing a peak draw of almost 1A at 3V and 1.3A or so at 2V...

by the time the batteries get down to 1V a piece it's probably impossible for them to put out 1.3A so your average output falls...

wow...long winded...

anyhow, the output curve of a step-up 'regulator' reflects the output/discharge curve of the batteries...it's much more noticable in high-draw applications than low draw, but even a regulated circuit with an unregulated (ie, bateries) power source,i would imagine, will have some sort of curve in it's output...

 

[papasan]

took me so long to write that gransee replied twice =)...hehe...

 

[JollyRoger]

Interesting topic, guys. I've always asked this to myself, too.

So how about using a combination of a step down and step up regulators with a starting high input voltage...say 4AA's...to run a single luxeon?

Would this be impractical as the efficiency of the thing would be below 70%???

 

[**DONOTDELETE**]

One thing important to note is that regulators provide either constant voltage or constant current, but not both.

On the other hand, most flashlight users really expect constant power, which is the product of voltage times current.

This is not a problem as long as the battery can supply that power. But, as battery voltage declines, so does the battery's ability to provide the additional current required to keep power constant.

Suppose, for example, at 3 volts, a battery can supply 1 ampere. At 2 volts, it needs to supply 1.5 amps---certainly possible, but for how long? At 1 volt, it needs to supply 3 amps. Can it still do so? How about at 0.1 volt---30 amperes??! You can easily see why regulation cannot just go on forever.

Now, you mentioned that you observe a decrease in power shortly after activation. Here are two possible reasons:

1. The circuit is not really a regulator. Output power is proportional to input voltage, as in the Zetex circuit.

2. The circuit requires so much power that the battery starts to heat up, thus increasing its internal resistance. Then as its internal resistance increases, its ability to provide constant power decreases. A larger battery will address this problem.

 

[Darell]

Wow you guys. As usual, the CPF members are out-doing themselves to help answer my question.

Though I'm certainly learning from all the posts, I think Duggg wins for conciseness (is that a word?).

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Duggg:
One thing important to note is that regulators provide either constant voltage or constant current, but not both.
<HR></BLOCKQUOTE>

Ah, see.... this is an excellent point that I hadn't thought of.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>
Suppose, for example, at 3 volts, a battery can supply 1 ampere. At 2 volts, it needs to supply 1.5 amps---certainly possible, but for how long? At 1 volt, it needs to supply 3 amps. Can it still do so? How about at 0.1 volt---30 amperes??! You can easily see why regulation cannot just go on forever.
<HR></BLOCKQUOTE>

Excellent example. No, I don't expect the regulation to go on forever - just until the battery is dead. But it sounds like some things may melt first. I guess what I'm looking for is possible - but is it practical (as far as hardware and run-time are concerned)? Yeah, 0.1V at 30A is a bit extreme. You've opened my eyes, certainly.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>
1. The circuit is not really a regulator. Output power is proportional to input voltage, as in the Zetex circuit.
<HR></BLOCKQUOTE>

Well, from what Peter just posted, I guess this is the common issue. These lights are "semi" regulated.

Learning, learning, learning.

 

[**DONOTDELETE**]

Semi-regulation is actually a good thing.

<UL TYPE=SQUARE><LI>The circuit is simple, inexpensive, and very efficient.
<LI>It provides reasonable output for a relatively long time.
<LI>More suitable for rechargeable batteries which don't like to be overdischarged.
[/list]

A more strictly regulated circuit would not have the same run time, because

<UL TYPE=SQUARE><LI>It would likely be less efficient
<LI>The battery would be consumed faster
[/list]

That said, we all would like better regulation

We just have to remind ourselves from time to time that it doesn't come for free.

 

[Darell]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Gransee:
Or drinking while posting...
<HR></BLOCKQUOTE>

Hey! I resemble that!

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>
You'll notice that the Arc inverters don't actually regulate that well. Its what I call "partial regulation". It is actually regulating so we can say that in our literature. It is just not perfectly regulating. The reason for it is because of a compromise. Once you see why, I think you would agree is a good trade off.
<HR></BLOCKQUOTE>

Thanks for the included great explanation Peter. I sound like a broken record here - but I will say it again before launching into my blurb: I REALLY like the Arc AAA lights, and can't imagine being without (at least) one at all times.

OK, now that that's out of the way - let me give you the impression I had of the lights before I owned one: From everybody's description (both here and on the Arc site) but without consulting some of the output/time charts that I've since discovered, it seemed as if the Arc would supply that elusive constant brightness "sun mode" right up until it snapped into "moon mode."

In reality, of course, that just doesn't happen. I've now burned through many batteries in my lights (on purpose, you understand - had to try real hard) and I still don't have a clue about when "mood mode" begins. There never seems to be a steep drop off at any point. They just s-l-o-w-l-y get dimmer and dimmer and dimmer. If you let the battery rest for a bit, it does come on noticeably brighter for a short while, but not what I'd ever call "sun mode" bright. One time I showed an Arc AAA to a friend and explained that the battery I had in the thing had powered it for 4-5 hours already. And though I had trouble convincing myself that it was still as bright as before, I proudly exclaimed "Still as bright as when the battery was new - its regulated." Then I put a new battery in it to prove my point. Oops.

It still astonishes me that you managed to build a waterproof step-up regulator into the head of this little thing, and still make it so much smaller than a Solitaire. I am curious what the trade-offs are of achieving better regulation. Complexity, efficiency, cost and size, I'm betting (what else is there, really?)

But before you respond, please keep the investors happy.

 

[Darell]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Duggg:
We just have to remind ourselves from time to time that it doesn't come for free.<HR></BLOCKQUOTE>

Yeah. It's that "no free lunch" deal that always comes back to bite us in the butt, huh?

Let's say the current Arc AAA runs for 5 hours at a decent brightness. What I'd be thrilled with is 3 hours of light equivalent to current new-battery brightness, followed by however long of 20% "moon mode" brightness. So, I'm willing to pay 2/5 of my battery time for better regulation. But I'm not sure if it's for sale that cheap.

 

[JollyRoger]

Darell, I second your request. I think this would sell well....but of course there comes a point where you have to have a fine balance of runtime and "constant brightness"

I don't really experience the dimming, since I use rechargables and recharge them regularly.

About moon mode: I usually use a nimh or nicad AAA in my ARC LE. The drop off is *noticeable* with these kinds of batteries. So I think I've witness the "turning into moon mode" phase...

of course I charge my batteries immediately.

 

[bikeNomad]

I've been playing around with real step-up regulators (current mode) to drive Luxeons in conversions, from 2 AA cells. These regulators will deliver a constant current for as long as they can, and hence a constant power. When they can't output the full 350mA, they go to a very low level (30mA or so). Suddenly.

As others have noted, constant power isn't really what batteries like to do, since what it means is that the more discharged the batteries get, the higher the current you're asking from them.

Which they're increasingly unable to do, since their internal resistance is also increasing.

So, even though my circuit does a great job of regulating the current, it doesn't treat the batteries as nicely as it could.

With alkalines, it discharges them to maybe 0.96V (for 2 cells) before it goes to the 30mA mode. So right before it switches over, it's drawing about 1.4A from the AA cells. Which alkalines don't like to do at all, especially dead ones.

It works much better for NiMH cells than alkalines, since their discharge curve is flatter.

 

[**DONOTDELETE**]

Hi guys,

Why not use a MAX711 IC.

It is a step up/down power supply IC.

MAXIM-IC link

Looks like this could be what you are after...

One luxeon star seems to be pushing the limit of this device, but is within the absolute maximum ratings....

Not bad for my first post here.

 

[**DONOTDELETE**]

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by darell:
I still don't have a clue about when "mood mode" begins. There never seems to be a steep drop off at any point. They just s-l-o-w-l-y get dimmer and dimmer and dimmer.<HR></BLOCKQUOTE>

As Peter mentioned, a step-down regulator only works when Vin is higher than the pre-set output voltage, Vout.

What happens when Vin falls to Vout? Two possibilities. If the circuit opens completely, the light suddenly goes out.

But if the circuit closes completely, Vout starts to track Vin. By this point, most of the battery energy is spent, so it's not capable of much power, so you get... moon mode!

 

[**DONOTDELETE**]

The MAX part is good up to 11 volts input. It seems it will operate at 3.7 volts, 350ma output at 3 volts input. As the voltage drops below 3, it looks like it will have trouble, but this seems the most promising for developing a "one size fits all" regulator/boost power supply for LEDS.

This could allow your 6 cell flashlight to have an extremely long run time.

 

[**DONOTDELETE**]

The max part can output 5v@250ma with only 1.8v input. It can keep that voltage regulated up all the way to 11v input.

According to the specs, it can output 5v@500ma at 3.6v input.

the MAX 711 part's output voltage is independent of the input voltage, there is no correlation, only if it drops too low for the part to boost it back up.