Like I said, I need to come up with something equivalent to replace that unit. If the OP was willing to spend the extra money a BuckPuck comes with a control input. Their could be 1 for each light. Goes against my grain to spend extra money, but it would definately work.

I'm willing to bet the BuckPuck doesn't come out with continous DC, but high power pulses similar to a joule thief.

 

I was looking at the 700mA PowerPucks. I didn't see a control input for them.

I think your bet that these are switchers is a very good one.

 

Look at this one...

http://www.leddynamics.com/LuxDrive/datasheets/3021-BuckPuck.pdf

 

Hey guys, thanks for the continuing help. I'm following along, struggling to understand, but I think I may have someone to at least help me assemble something at some point...at least someone who knows more than I do.

I would be willing to use 2 BuckPucks if that's what it took...

Tj

 

There are a bunch of LED drivers out there, all you have to do is search for them. Go to some semiconductor websites such as Linear Technology, Texas Instruments, Maxim IC, and Micrel to search.

Austin

 

Here's something I keep in a here document in a shell script for when I need some energy info on batteries. I assembled the data a few years ago from a variety of sources.

Battery capacity in mA*hr

                Alkaline    C-Zn    NiCd        NiMH        Lithium
                --------    ----    --------    ---------   -------
AA               2700       1100    600-1000    1700-2900     3000
AAA              1200        540                800-1000
AAAA              625
C                8000       3800                4500-6000
D               19500       8000                900-11500
9 volt            565        400    120         175           1200
CR123                                                         1500
CR2025   dia=0.787, height=0.098                               160
CR2032   dia=0.787, height=0.126                               225
CR2450   dia=0.965, height=0.197                               610


Silver Oxide   Volts mA*hr     Alkaline  Volts  mA*hr   Dia, in   Height, in
------------   ----- -----     --------  -----  -----   -------   ----------
    SR41        1.55   42        LR41      1.5    32     0.311      0.142
    SR44        1.55  200        LR44      1.5   150     0.457      0.213

 

Bill,
I think that BuckPuck will work OK. I've been kind of under the weather for the last couple of days, and I really don't want to mess this up by steering our OP wrong.

I see that BuckPuck will handle open outputs OK, so that eliminates my earlier concern about both MOSFETs being turned off.

turnertj, don't know how much of a hurry you're in to build this thing, but I really haven't been up to sorting out the requirements yet.

[eta]
OK, I think the 3023-D-N-700 - 5v-32v DC input, wires on output, no dimming or current control - will work.
Input margin is 2v.
We need to know which Luxeon III LEDs you're planning on using.

 

OK, here is how I would do it with two puck bucks. It bothers me, I hate using expensive components when I'm convinced there ought to be a way to use it once, but it will work.

C5 should be directly across the power leads of the 7556, to supress noise from the puck bucks.

Just curious, what are you using it for?

.

 

CR123 batteries are 2.5V each dropping to 2.25V each in one hour at 750mA.
The buckpuck will barely have enough voltage when the batteries are new then will not have enough voltage as they wear out.

 

Nope, you got this one wrong. Try reading the data sheet I posted. While rated voltage is rated 3.3 to 3.6V, nominal voltage is 3.0.

 

Nope, you got this one wrong. Try reading the data sheet I posted. While rated voltage is rated 3.3 to 3.6V, nominal voltage is 3.0.

Sorry, Bill - we both missed that one. Audioguru's closer than you are.

Have a look at this graph, right from the datasheet you posted:

In about 1 hour at 750mA current, the batteries will be down to 2.5v each (5v total).

 

OK, I stand corrected, but it wont' be 750ma. Thing about PuckBucks, the reason people pay the costs, is they're converters. Figure around 400-500ma input, which is over 2 hours. The OP knew about the short life going in, you mentioned it. This PuckBuck also has a spec of 5V, so no problem there.

I was originally going from this table on the same data sheet.

 

Bill,
The Buck Pucks are simply buck current limiters; not transformers. You have to figure that they are not 100% efficient, either. If they're 94% efficient, they'll draw about 750mA for a 700mA output.

[eta]
Just looked in the datasheet; they're 95% efficient with a 5v input. This beats the dickens out of the efficiency you'd get with a linear solution.
700mA/95% = 737mA. Don't forget that the 556 timer circuit will also draw current.

Basically, the 750mA curve from the datsheet is within 2% of what this projects' load will be.

 

Lets go through the math.

Conversion ratio is: 3.6V / 5V = 72% (worst case, their will be more voltage DC wise).

Add the inefficiency: 72% / 0.95 = 76%

700ma X 76% is 532ma.

Wattage is the constant, not current.

 

Thanks guys...I'm going to try building it in the next day or so here...once I find all the components.

Where is the best place to find the 7556 chips?

Bill, by looking at the schematic you sent, are the numbers around the two 7556 chips the leads for the chip? So I'm only using different leads on both chips? Please forgive my electronics ignorance...I plead the mechanical engineer affliction!

 

There is only one chip. It is a dual 555 timer, it has two separate 555s inside one package. This can be a really small package, the PuckBuck is around the size of a quarter.

The numbers around the parts are the pin numbers of the chip.

So I assume white LEDs (it makes a difference), whatcha using for?

 

OK, you can either use two 555 timers, or one 556 timer. They (or it) should be CMOS. BJT (transistorized) 555 timers will not work correctly, as they will not raise the gates of the MOSFETs high enough.

Digikey.com is a good place for making small orders; they're hobbyist friendly and they will ship via USPS 1st Class Mail, which will save a good bit on shipping costs. Handling costs are built-in to the price of the parts.

I still don't know which LEDs turnertj is using, and what their typical Vf @ 700mA is. Until that is known, it can't be determined if this thing will work.

 

I have several lights that are high power LED.

They all run on CR123 batteries, with a life of 1-4 hours.

Peak LED Solutions uses one cell and gives me good light (100+ lumens) for the longest run time on a CR123, but starts flashing around 20Hz when the battery gets low. The switching supply in it is not visible to the user without bending aluminum.

I have a Surefire E1L which is a bit brighter, bu goes through batteries much quicker, I'm guessing their regulator isn't as efficient as the Peak LED one.

Point of all this is that CR123 batteries carry a pretty good punch. 3V and about 2000mAH for decent brands today, in a much smaller physical package than 2 AA's to get the same Watt Hours.

I'll be interested in how this project turns out!

 

Lets go through the math.

Conversion ratio is: 3.6V / 5V = 72% (worst case, their will be more voltage DC wise).

OK, we know we're going to take a "hit" due to losses in the batteries; that's unavoidable in batteries that small.

Add the inefficiency: 72% / 0.95 = 76%

Ahh, whoops! 72% times 95% = 68.4%

700ma X 76% is 532ma.

Bill, sadly that's not how it works. If you could figure out how to actually make it work like that, I would be more than happy if you would give me just 1% of the profits you'd make from it, as you would be creating energy from thin air.

[eta]
If the BuckPuck contained a step-down transformer with a chopper driver, then part of what you're thinking might be possible; however it's just a switching current regulator that has a simple inductor to keep the current flow reasonably constant, along with associated driver circuitry. It cannot possibly output more current than it receives, as it isn't a transformer; it's just a buck current regulator.

The BuckPuck will attempt to keep very close to 700mA current through the LED presently selected.
As it is 95% efficient with an input of 5v, it'll take 737mA current to do so.

Back at the ranch (er, batteries...) the CR123's measure - let's say - 3.3v with no load on them; for a total of 6.6v. So, 1.6v is being dropped across the battery internal resistance. With a 'ballpark' current of 737mA and a 1.6v drop, roughly 1.18 Watts of power will be dissipated in the batteries alone.

Until we know what the Vf of the LEDs are @ 700mA current, it'll be tough to calculate the wattage dissipated in the BuckPuck itself - or even if it's going to work as we think it will.

In order to get an idea of the LED specs, we need to see a datasheet.

 

Wow...

I didn't realize the LED color might make a difference, or maybe better put, I didn't realize that the Vf would make that big a difference.

My first thought is to use 1 red and 1 green Luxeon III LED.
Green: http://ledsupply.com/lxhl-lm3c.php (Vf 3.9)
Red: http://ledsupply.com/lxhl-ld3c.php (Vf 3.2)

But then I was pointed toward these:
Green: http://ledsupply.com/creexre-g.php (Vf 3.6 @ 700mA)
Red: http://ledsupply.com/creexre-r.php (Vf 3.2 @700mA)