Need (lots of) help on DIY light

[bradsab]

Got a new point-n-shoot camera & housing for Christmas with HD video. I just shoot for my own pleasure and to share with family, and I was always happy with the stills I took with my old camera and internal flash. I never shot much video because the quality was lousy, but now I hope to take a lot more.

I want to build a combo dive/video can light. Compact size is a high priority; if it gets too bulky, I’m not going to do it. I want to be able to have one hand free even on night dives because I don’t like the light/camera juggle. The light head will mount to a tray screwed to the bottom of the camera housing, and will have 2 or 3 LED emitters. One will be my primary dive light, the other one or two will be video flood(s).

I’ve spent many hours searching SB and CPF trying to decide what to put in this light. The more I read, the more confused I get. Add to that is a bit of urgency: I have a dive trip planned in six weeks, and my current work load is going to make it difficult to get it done. To speed things up, I’m begging for some input. I need help choosing LEDs & optics/reflectors, drivers, batteries, and canister design.

For starters, I’d like to focus on the LEDs and driver. If I can decide on the emitters & optics I can at least get started machining the head and building the arm & tray. I’m leaning to the Cree MC-E with Ledil optics. While the Seoul P7 and the Luminus SST90 look similar, I can’t find as many optics options. The Cree XP-G is more efficient, but then I’d need two video lights to equal the output of the MC-E, wouldn’t I?

When it comes to drivers, I’m completely lost. The last light I built was direct drive (linked in my signature), but it drops off noticeably during use, and even with fresh batteries, it doesn’t seem as bright as when it was new. From what I understand, a driver regulates higher voltages and keeps them constant to the emitter? Higher voltage = more current? More current thru the diode creates more lumens, more heat, less efficiency?

There appears to be a wealth of knowledge on this forum, but I’m just not putting it all together enough to get it. Any help would be greatly appreciated.

 

[bradsab]

Well, I just spent the whole day reading threads here and at CPF. Still confused, but if I can find the right optics, I'd like to try the SST-90. It's frustrating because Ledil has dozens of optics & reflectors dedicated to the MC-E, complete with specs and beam shots, but almost nothing for the SST-90.

 

[meesier42]

try PM member Packhorse

he seems to know what to do.

 

[Packhorse]

SST-90 is a bugger because there is very little in the way of drivers that will work with it.
I have built a SST-90 video light but the driver I used is not exactly easy to implement.
I would suggest 2 mag lite based light heads each with a MC-E LED driven at 3.5 amps behind a 44mm aspheric lens from DX.
Check out my Aspheric mag and W300 posts on CPF for details. It would be a mix of the two.

 

[bradsab]

SST-90 is a bugger because there is very little in the way of drivers that will work with it.
I have built a SST-90 video light but the driver I used is not exactly easy to implement.
I would suggest 2 mag lite based light heads each with a MC-E LED driven at 3.5 amps behind a 44mm aspheric lens from DX.
Check out my Aspheric mag and W300 posts on CPF for details. It would be a mix of the two.

Well, I might have rushed into this a little, but I'm worried about getting it done before my trip. I've already ordered three SST-90's, and they've been shipped. I've also ordered several different reflectors & optics hoping I'll find one that works acceptably for spot, and two others for flood.

My cannister will house nine CTA D 12,000mAh Ni-MH batteries arranged 3X3 to give me 3.6Vf. I want to build this around these batteries because I already have them and a Maha charger.

My current plan is to direct drive the SST-90s, switching between the spot and floods via an on-off-on toggle rated at 15A. The cable will be 16ga., but inside the head I may drop to 18 or 20 gauge because of the tight space. The two floods would be wired in parallel so they both get the 3.6Vf.

I would like to have a low power option. Can I do that with just a resistor? I seems to me a simple resistor would work differently with the single spot SST-90 than the twin flood LEDs. I'm in over my head on this, and I don't want to make a mistake and fry $150 worth of LEDs. At the same time, I want to keep it as simple as possible.

 

[DIWdiver]

I'm not a big fan of direct drive as it offers no limitation to the max current (which can fry your LEDs if you have too much chutzpa in your batteries) and because the output changes as batteries discharge. But I can offer some help to get you out of a jam.

Yes, you can drop the power with a resistor. Yes it will be different for the single and the double (two in parallel) LEDs. And you will sacrifice some efficiency vs. a properly designed driver.

If you use a double-pole double-throw switch for your high/low switch, it's not a problem using different resistors for the single and double LEDs. If you really want to use a single pole switch and one resistor, that's fine too. You will just see more reduction in the double than the single LED.

Direct driving, you'll get something like 10A per LED with full charge, quite a bit less nearing the end of charge. Most folks call 1.0V/cell the end of charge, which would give you 3.0 Vf. The data sheet shows about 2A at 3.0V That's a pretty big difference! Now decide how much you want at the 'low' setting. A factor of two in power is noticable, but not large, so I'd shoot for a 3-4 times reduction.

On the single LED, if we shoot for 3A at full charge, the data sheet shows about 3.2V on the LED, and the batteries are 3.6V. Thus we need to drop 3.6-3.2 = 0.4V across a resistor at 3A. Since V=I*R, or R=V/I, R=0.4/3.0 = 0.133 ohms.

The power dissipated in this resistor is given by P=I^2*R, or 3*3*0.4=3.6 Watts. That's hefty, but not unmanagable.

For the double LED, all the currents are double. We want 6A at low power, R=0.4/6 = 0.67 ohms. P=6*6*0.4=7.2 Watts. Another way to think of it is that for two LEDs in parallel, you need two resistors in parallel, so calculate for one LED, then just use two of them.

A number of different 0.13 ohm 5W resistors are available from electronics distributors. I picked 3 from DigiKey Corp. | Electronic Components Distributor | United States Home Page, which are progressively smaller and more expensive:
TWW5JR15E
ALSR-5-.13-1%
AC05000001507JAC00

Note that if you want an even bigger reduction at the 'low' setting, say to 1A, you'd have to do some experimenting, as the data sheet doesn't show currents that low. Also note that the power in your resistors is much less (making them smaller), and that the drop in output due to discharging batteries is less significant.

Hope this helps.

D

 

[bradsab]

D-
That helps a lot, thanks! Like I said, it's a bit over my head, but if I keep re-reading it, I think I can get it. Part of the fun of building my own stuff is what I learn along the way. I just don't want my education to cost me some expensive LEDs.
From what I've been reading, the SST-90 will take a lot more current that I can feed it with the battery pack I'm building. Do I have to worry about frying the LEDs with what I'm doing?

 

[Packhorse]

I very much doubt you will get anywhere near 10 amps to the SST-90 with that battery set up. EDIT Dependant on Vf of SST-90

I ran a direct drive setup using 7 x 2500ma 18650 Li Ions. It started at 6 amps and quickly fell to 3 or 4 amp.

I wouldnt worry about a low mode if you decide on the DD method.

Personally what I would do ( with what you have) is run a 2P4S pack so you get 4.8volt then make up a 7 amp driver using 5 of these drivers. For low mode just use one or 2 boards.

 

[DIWdiver]

I find myself in the unusual position of disagreeing with Packhorse. I think that if you use short, heavy guage wire, you WILL get 10A or more from your pack to the SST-90, at least for a little while. That's assuming your cells are in good shape.

However, in the end I have to agree with his recommendation. His setup will give you nice regulated light output that's constant for most of the pack's life. And despite having only 2/3 the amp-hour rating in the pack, you'd get full output longer than with the 3p3s pack. You will get less total time, and a more precipitous drop at the end, but do you need more than 2 hrs on a charge?

I would recommend only 6 boards though, as 7 will give you 9.8A, which is well above the absolute max current for the SST-90. If your heatsinking is really good, and you plan to use it only in the water, you can probably get away with 7.

However, to run your double LED, you'll need twice as many boards (12 or 14 total). Fortunately, you still only need one SPDT and one DPDT switch, one of which needs to be on-off-on.

D

 

[bradsab]

Packhorse & DIWdiver-

Thanks so much for your input. If I'm going to expand beyond the simplicity of direct drive, I need to note some things about this build.

First, the power supply. I hope to start building the canister tomorrow, but I haven't done anything more than gather the supplies for it, so at this point I can make changes. Whatever I do will be done in parallels of 3, so to jump to 4.8v, I would simply make a longer can and have 3p4s. I can also allow room for drivers in the canister, but the on-off-on switch for spot/flood will be in the light head.

I'm trying to keep the head as small as possible so the whole camera/housing/light package is still compact enough to operate with one hand. The design I'm working on doesn't leave room for much more that the LEDs & switches. I'm probably limited to resistors up there. I could make room for a small driver behind each star, but then I'd need mount the stars to a separate heat sink, and it would only transfer heat to the head around its edges. I would rather attach the stars directly to the back of the head and get better cooling and smaller size.

My ignorance about drivers & regulating current prevents me from understanding the how’s & why’s of what you’re recommending. Is there a thread somewhere here or CPF where I can educate myself a little? I’ll be glad to just do what you suggest, but I’d really like to understand it, too.

The heaviest cable I can get thru my glands is 16ga. Since I don't have room in my head for 14 drivers, the glands will also fit an 18ga. three-wire cable, and i could supply two separate regulated power circuits up to the head, using only one for the dive light, and switching both into the two flood LEDs. But then the ground wire would only be a single 18ga. for two LED's pulling 7A each. Is that OK?

If I went with the 3-wire cable, the Hi/Lo switch could be in the canister, and I could switch out some of the drivers to give me the low power setting. I would rather have the switch in the head, but having it in the canister is an acceptable trade-off to get better circuitry. Is it a dumb idea to send regulated current to the head, and then dim the LEDs with simple resistors up there? If I did that, would there be a single driver that would work for each SST-90 without having to use 14 of the ones from DX?

I want to build this light so it's safe to use out of the water without worrying about it frying. In reality it's going to be mostly in the water, and the longest surface times would be prepping for night shore dives, most likely at low setting. But I don't want someone who doesn't know better cranking the thing on high, and then leaving me in the dark. My luck, it'd be the first dive of a trip. Thanks for your help!

-Brad