Custom design of a XHP-70 Flood Light System (Updated Photos 7/11/16)

I didn’t know LiFePO4 could be depleteed of all its capacity. I assumed it behaved like the li-ioms. That means ill have even longer runtimes. My initial plan was to run 3.5 hours but It will hit over 6 hour which is even better.

Oh so I won’t be able to drive off a 12v? Unless I use 6 drives?
Maybe I should bump the battery pack up to 24 volts 60AH?

Still, reading this, man what a monster project to start
Wouldn’t you be better off just doing some basic flashlight conversion projects and get to know different drivers and the XHP70 before going for this 6 XHP70 multiple driver monster?

All or nothing….hahah
I have modded a few lights of mine before. One I did mod a mtg2 to a Xhp70. Also did do some fixing on a Niwalker mm15mb that came with a dead XHP70.
In terms of drivers I’m not as familiar. But doing this project will help me. I can learn from failures. But everyone has been helpful here and it seems to slowly be going forward.
So any tips please let me know so I can improve.
I need to create this light. It’s something that will have use for me and will for others at my local rescue volunteer. Which will be awesome to be using something I made.

Bruno28, I was not meaning you can completely deplete LiFePO4 cells without harming the cells; the absolutely lowest cutoff voltage should be 2V/cell for this chemistry, though these matters are normally handled by the battery management circuitry. For example: 6S 60A LiFePO4 BMS/PCM/PCB battery protection board for 6 Packs 18650 Battery Cell w/ Balance -> Cutoff voltage set at 2'3V.

If you are building/configuring the battery pack then I'd go with a 6S configuration (19'2V nominal) which ensures you'll be able to extract nearly all of the useable battery capacity withouy issues, while at the same time the output/input voltage differentials are kept at a minimum so the driver efficiency maximizes. Please note that the driver efficiency figures claimed are for operation at 32W of output (12 down to 6ish volts), and it is very likely that the efficiency would diminish doing 24 down to 12ish at double the load. Better make sure you heatsink the stuff rather rather well or you may incur into reliability issues.

Cheers ^:)

Ok, there is no point in using all the capacity of the battery if it will harm, i think leaving 20% is fair and will keep its health. So i think leaving 80% for use is doable. Im researching about LiFePO4 batteries now. I don’t know much about these one BU-205: Types of Lithium-ion - Battery University

LiNiCoAlO2 batteries seem to be much higher Wh/kg (260WH/kg). Over twice as much than LiFePO4 (120Wh/kg)….hummm

I will go with 24 volts, it should be better. And the supplier can provide this.

would you better building a light with 2 or 3 x xhp70. primarily for development also if you make 2 or 3 you can aim them in different directions
just thought i’d link this thread by OL his is a triple xhp70

Ive seen that, but thats a flashlight and not a night time woking light for rescue.
Also that has a very bright narrow beam for throw, mine will be a complete wall of flood light.


LiFePO4 is one of the most abuse tolerant chemistries around.

How you do “determine” that 20% left? What's the point?

What it should or not should… Facts?

Seems to me you're after some kind of expensive commercial LiFePO4 truck battery. I mean, if this is true, you seem to have $$$ to spend yet, on the contrary, your lamp design guidelines are “scrooge-class” (only 6 XHP70 emitters “to the brim”, overloaded drivers, etc) and you want that to meet the lofty goal of “20000 lumens, built to endure harsh conditions” gear for rescue work? :FACEPALM:

Sorry, don't get me wrong, I do not mean to be an ass my dear, it just seems to me you are deliberately self-sabotaging yourself. NASA would definitively dissaprove this…

Cheers :-)

Answering your comments in sequence.

1.yes LiFePO4s are tolerable to abuse but fully discharging can kill the cells. LiFePO4 can withstand higher charge cycles and shelf life when compared to lithium cobalt oxide (the ones found in electronic devices)
See this for more einformstion about li-ion battery types and they advantages and disadvantages.

2.doing the calculation with only using 80% of capacity will provide me with actual use hours. The internal PCB will cut once voltage is low enough which is around the 20% capacity left on the battery pack as per the supplier I’m speaking with.
I’ve done lots of testing with li-ion batteries for my university thesis. I also have done with li-polymer. And I’m a hobbiest for rc helicopters and “drones” which I have to be cautious to never over discharge these batteries. So I know a little about these batteries and how to care for them.
But for more information about discharge see this website. “What constitutes a Discharge cycle”

3.going with 24 volts is better as per what you said before (“With 12’8V of 4S LiFePO4 nominal voltage, you won’t be able to drive 2S 6V XHP70s off that driver; there’s simply not enough input voltage overhead, you’d need at the very least 14V I think.”)
My underdtanding is that by having a higher voltage input will require less stepping up of voltage from the driver to led. This will make the driver have to work less and also will possibly reduce heat generation which is efficiency loss.

4.i prefer investing in quality products than something cheap. Reliability is important for rescue situations. Can’t have something fail mid work. This cost will reflect on the selling cost to buyer. But knowing that it’s going to be a good battery that will last 2000 cycles and around 5-7years (as per manufacture) vs some Chinese crap that will last a few charges is also a strength on the product marketing.

5. I’m using one driver per xhp70 that’s why I bought six. I would like to try and use 3 for 6 led. But reliability is key for me and so there will be 6 drivers and 6 led. Won’t be over driving any drivers.

So the drivers will be feeding XHP70 12V configured emitters, less current, lower losses. Makes sense then.

Still, 6 emitters for 20K lumens is something I'd qualify as “to the brim”/“pedal to the metal”. :-)

Cheers ^:)

I think it might be even more than 20000lm. I haven’t taken I to account the loss in the TIR lens yet. But 20000lm is probably more than it will be used. But the potentiometer I plan on putting will allow users to regulate the right amount of light as a infinite variable output.
If running at 32 Watts as per cree datasheet it should provide 4022lm output per xhp70.
My calculations expectation is that each led will provide only 3500lm at max (pessimist output). And that should be 21000lm which is a LOT already. Most jobs won’t require that much but it’s good to know you have that much light available if it’s a large area that needs lighting.

Let's do a little math. Even for an N4 bin, 1710 lumens is all you get at standard bin drive current. Checking the relative flux/current graph, one can see that 192% is the relative amount of flux at standard maximum drive current and 85°. Thence, maximum flux is 1710 × 1'92 = 3283'2 lumens for a top bin (and likely crap CRI) emitter. So my maximum possible calculated figure is even lower than your pessimistic one. Now, let's say you're not getting top bin, not even close, what happens?

Like flashy numbers? :-) Take a look at this, I've solved your problem already:

Source link: 5000LM CREE Q5 AA/14500 3 Modes ZOOMABLE LED Super Bright Flashlight MINI Torch

Cheers ^:)

Good news. Heatsink arrived. Still don’t have the xph70 so I’ve tested with a not so efficient 100w COB led for 30 mins until now and it’s warm but I can easily touch the fins and not burn myself.
Definetly hotter than I thought but won’t burn any hands with this temperature. Also this was 30 mins with no air flow at all. I don’t have a laser thermometer but I would say it’s around 45-50 degrees Celsius by touch.

Considering that when I use the six xhp70 it will be between 180-190 Watts, I think the heatsink will work well if the leds are run at 100%.

Could you give us a link for this heatsink please? How much does it weight?

It’s an Australian conpany. See the link below for dimensions and weight.
Also see the tab for pricing.

The one I got is the MF30-100

Some updates on build :wink:

Spend a few hours planning on how to wire everything up.

Now for some beam shots comparing a Niwalker MM15MB (6800lm), Light X7 (9000lm) and the Light Bar with wide lens and narrow lens (in theory should be between 20000 and 24000lm)
The photo isn’t showing very well but can clearly see even the trees at 130 meters. But what matters to me is the close wide beam, which is perfect spill with evenly distributed light and a very good 5000k tint which makes it look like sun light.

Long Distance

Close Distance

wauw what a heap of light! Awesome

Yes it is. And it’s a good tint too. Looks like daylight, which is perfect for working light to illuminate a huge area.
I’m happy with the outcome until now. Can’t wait for it all to be put together after the main parts get manufactured. :slight_smile:

As I explained above, unless you have some sort of special bin emitters and/or the heatsink performs wonderfully in keeping the emitters below binning temp, I doubt that breaks 20K lumens, though it should be pretty close.

Theory equals practice when any and all variables are considered.

Cheers ^:)

Update on the first prototype machining. :slight_smile:
Im happy its going somewhere now. First prototype will be complete in less than 2 weeks.