step down convertor for 5W LS?

[r2]

As I understand it, step-down convertors are more efficient than step-up convertors. Also, the 5W LS emitters generate enough heat that 2x123 size lights tend to overheat (the heat the batteries generate at that load doesn't help either).

With that in mind, it seems like a 4x123 housing with a step-down convertor could make an excellent 5W LS light. 4x123 cells end-to-end is still pretty short (a little longer than 2xAA?) and would be a nice comfortable size to hold.

The batteries would start out supplying 12V w/out load and the individual cells could drop all the way down to around 1.5V before underdriving the emitter.

Questions: would this give better battery life than running 2 sets of 2x123 cells one after the other? Would the batteries generate less heat? What about the convertor? How hard would it be to make such a circuit?

This struck me as a good configuration that would also give a little extra body size to help dissipate the heat from the emitter. Anyone know of a good housing that would be about the right size?

- Russ

 

[carbonsparky]

with the right down converter, yes 4 in series would be as good as 2 sets of two. It would even be better since batteries must be will matched to work in parallel.

As for step down converters, I to would be interested if anybody has one they like to use.

I have been working on one on my bread board using a cmos 555 timer driving a mosfet. This one is adjusted for desired ideal output and left. It does not compensate for battery depletion. I have ran standard LS off of 12 volts with only a 10 ohm resistor in series. It could have been even smaller resistance that just happens to be what I used. I plan to get it on a scope and do some tweaking sometime soon.

 

[dat2zip]

I've two in the works.

One for high input voltages like 12-18 volts and the other would be more for under 10V situation. The second one that would run at 10V or less would run right down to the Luxeon voltage. This is the most efficient use of the converter to run all the way down.

I'm leaning back towards the Maxim family of step downs for the second one since some of them go into linear regulation mode.

I've only found one step down that has a high switching frequency which means the capacitors and inductors will be very small. The part is the Linear Technology LT1616. I've got a LT spice on this for the 1W design.

I don't have any details or info on the 5W yet.

There are others from National and Maxim and other MFGs but, most switch in the 150KHz range and the inductors are incredibly large.

I've ordered parts for the first proto LT1616 1W design and should be getting those latter this week.

The LT doesn't run down very low, but, maybe for the 5W this would be OK since I know it won't go down to ??4V. At 6.8V or so the 5W might be useable with the LT1616.

Four 123 batteries should provide 4 times the run time as would two batteries. The Watt/Hr density in four batteries is twice that of two batteries and with the proper DC/DC converter should draw the same power. In the 12V configuration, the converter should be drawing 1/2 the current it would be in the two cell configuration.

 

[Albany Tom]

[No message]

 

[Albany Tom]

[No message]

 

[Albany Tom]

Ok - crazy question:

Why couldn't you use a switched capacitor system, rather than an inductor based system? This is just off the top of my head, but charge a small cap from the battery, by firing an scr or something, then dump that cap into a lower voltage, higher capacity cap that runs the light.

If the second cap could power the light for a few hundred miliseconds, and the frequency of the thing was in the low kilohertz, you wouldn't need fancy PWM. You could just use a "bang-bang" type of thing to turn on and off the oscilator between the high/low set points.

1. I hate winding coils.

2. This is going to keep me up at night until either I figure out why it won't work, or someone else points it out to me...

 

[Ron Schroeder]

Hi r2,

Step down is not much more efficient than step up. Maybe only a % or 2. The big problem with step down is you need to have enough batteries to be above your output voltage at end of life. With 123 batteries still having energy in them at 1.5 volts, you would need at least 5 or 6 cells for a LS 5W. Step down is more practical with rechargables that have less of a differance between full charge and end of charge. 8 nicad or NMh cells would be good.

BTW, capacitor charge pumps are very inefficient except at very narrow specific divide or multiply ratios, not over the whole full charge-discharge range.

 

[Lucien]

Thinking about it, its a great idea. Imagine, a 4xCR123 sized light that puts out ~120 lumens

I figure runtime would be around 2 hrs too, not too bad. (Unless I made a big mistake somewhere)

The worry would be as Ron said, 4 might not be enough, since they would run down to 1.5v@, giving 6v output, below what we need.

Another reason a step-down might be a good idea: lower current draw from batteries compared to a step up. I'm given to understand large current draws aren't good for batteries, or aren't they

dat2zip,
I think the LT1767 might be useful, has a switch current of 1.5A. Can the LT1616 supply enough current for the 5W LS? I'm under the impression it can't go much higher than 400-600mA

 

[r2]

I may be flaunting my ignorance here--please correct me if I'm wrong.

As I understand it, the voltage drops significantly under load with all batteries. One of the reasons I thought a 4x123 configuration would be nice is that it would put less load on the individual cells. Would they still drop to 1.5V too quickly to be useful? The point of adding more batteries would be to use them more efficiently and get _more_ light out of them. Lithiums are already pretty good at delivering high current and have a relatively flat discharge curve. Let's see... at 12V if we assume 6W used (5W + 1W lost in the circuit) that's about 500 mA drawn from each battery. Something like the SF E2e pulls about twice that and still holds up for reasonable battery life.

I don't know how load would affect all this. I'd love to see 2 hours of regulated output running at the full 5W. This was the best configuration I could think of to achieve this in a compact size. A body with sufficient surface area (read: fins) to dissipate the head would also be necessary.

Any other thoughts on an ideal configuration for a compact 5W LS light?

- Russ

 

[ElektroLumens]

I have a circuit I put together, using a LM2595 chip. As dat2zip says, it's huge. But, it does work. I come in with 12 volts from 8 AA's, and it regulates for either the 1 watt or 5 watt just fine.

This is an older 150khz chip, and it is big. But it is very simple to design, etch the board, solder the parts, etc. Takes inexpensive electrolytic caps, etc.

Wayne J.
www.elektrolumens.com