LEDs mounted near the front of the lens… no reflector, no optics:
.
pic is a link to more info
A mule is a light without any reflector or optic. The light is spread very wide directly from the led, whereas a reflector or optic can focus a beam of varying width depending on equipment. Usually many led’s.
This setup is probably ideal for me as I will be mostly taking macro photo and video of subjects. I already built something similar to a mule with 6 Cree LEDs and had good success.
the floody optics on a 4 LED Emissar D4v2 would probably make an effectively even distributed field of light for photography… without needing to be a Mule
I dislike Mules because they throw diffused light sideways, almost 180 degrees
I would prefer a cone of light, shining forward onto the photo subjects
If you want components deal with Richard Mountain Electronics —— I would think a Triple or a Quad would be sufficient — I’ve built a green Triple Convoy S2+ for hunting
I’m looking into Mountain Electronics and their stuff seems great, but the Cree XTE I need to use isn’t very bright for photography and video, I think I’ll need to use at least 4, preferably 6-9 for best results. But others floated the idea of overdriving the chips which I have not tried.
the Emissar D4v2 uses 4 leds… but it also has Aux lights that you do not need, and the Aux board adds complexity when disassembling to change LEDs
otoh, if you ask, Hank may be able to source and install the LEDs you want.
the DT8 uses 8 leds
both lights are based on a single 18650 battery, which Im confident will give you plenty of power for Macro distance photos of Frog Bellies 
Can you give an idea of how much brighter you need? What current have you been running the single XTE, and how many times faster do you need the exposure to be? I think we can suggest a good light for you, but I need a more specific idea of what you need.
Build a P60 light ( Solar Force etc) with one of those 12 LED mules — Richard has the parts — I’m sure you can find a P60 host some where —Or a light that will accept that 12 Led board
I didn’t have time to look into this more, but I don’t think there is actually anything super special about the XTE’s bandwidth, but rather it’s binning and marketing language. I compared to Luminus SST-20 blue and SFT-20 blue and they are in the same ballpark. XTE by my measurement of the datasheet spectral curve had a fwhm of ~19nm at 50% relative power. The Luminus emitters were rated at 22 and 19nm, respectively. I haven’t check other manufacturers, but I don’t think there is necessarily a reason to limit yourself to the XTE. I could be wrong, but it’s something to consider.
Reflowing any 3535 emitter is super easy. Also, if you make up your mind on what you want them in, I’m sure a member here would offer to do this work for a small fee.
jon_slider, I've been talking to Hank and he is willing to build a light for me. I may take this route, but the more I think about it the more I am leaning toward trying to build a dive light. A non-waterproof light wouldn't be nearly as versatile for me.
EasyB, I used a custom build light with 6xXTEs running at 3V each to take this photo:
I used the following settings: 1/160s, f/3.2, ISO 2500. I think I'll need to increase my speed by 10X+ to get effective video.
JaredM that's extremely interesting. In my research I did not look at Luminous chips, but the other brands I analyzed that had broader bandwidths than the XTE performed poorly (I linked a graph above). From what I am seeing of the Luminus SST-20 they seem like they would also be an effective candidate. I have managed to successfully reflow chips but only by removing the PCB and putting it on a hotplate. When I try using a small heat gun or iron, I burn the chips. Is there another more nuanced method you can recommend?
Everyone: At this point I'd like to make a dive light if possible. (I may also make 3-4 triple-LED lights using mountain electronics for use in surveys with volunteers.) The challenge is this: How do I find a dive light compatible with XTEs/luminous chips? Maybe I can find a dive light and swap out the PCB, is there a way I can get an idea of the PCB diameter in a light? Are they pretty standard sizes?
One of the reasons I want to build a very bright dive light is so I can leave it out with a video rig in the intertidal zone. I read into how some videographers collect tidal footage using a gopro mounted to a cement block. I would drop it off at low tide and then leave it for a few hours until the next low tide and then collect it. This would be great because many organisms hide when exposed at low tide, meaning they can't be easily filmed. But the tide is too dangerous at high tide to swim in.
Sketch of my current idea:
https://drive.google.com/file/d/1HimVlwtQF2w6CNIcvhOr5tbSVGlSYLBF/preview
Basically it would involve a gopro outfitted with a macro lens and an emission filter taking footage of fluorescent sea life.
Your plan would require a lot of power. One LED at 3V is about 750mA or 2.25W. Times 6 is 13.5W. If you want 10x that it is 135W. For 3 hours that is a 405 Wh battery. One 26650 cell is about 20Wh. So this will have to be a custom solution unless you can work with less brightness.
Have you seen the equipment available at Nightsea here . They use the broader band leds and use appropriate filters to eliminate the excess light. This would give you more choices and you could spend your money on filters instead of flashlights.
Nicky Bay has some awesome macro photos of spiders and other things fluorescing. Check out his gear post for some ideas on macro equipment.
Easy B, I may not need to get as bright as I mentioned, but I have no problem building an array of 20+ batteries if that is what I have to do to obtain high-quality footage. I may start with just 12 LEDs to see if that is sufficiently bright for quality video. I think the challenge will be figuring out a waterproof housing that will fit everything while also not insulating and overheating the LEDs. Thanks.
Scientist, I am aware of Nightsea’s products, they are very good. I will look again at their filters, but I don’t remember them being as narrow of a bandwidth. Thanks for the tip.
One more bit of nausea for you to deal with is that wavelength shifts as you drive the emitter harder.
Listed wavelengths is more marketing and binning. Look up the Cree way of listing 4C, 7A, 5D, etc., vs the Nichia way in MacAdam ellipses.
Have a look at the emisar D18. It might be a good starting point to see what will work. You might need to beef up some of the oring seals to make it more waterproof. If you need more battery maybe you can wire in an external battery pack later.
There’s also some submersible options if you search for “12 led diving light”.
I looked at a Luminus datasheet because I was about to declare the same warning. With the Royal Blue, there is surprisingly small amounts of shift. 2-4nm at most. Native red emitters however have a high temp sensitivity.
So the main problem I am foreseeing with using the bandpass filters is the problem of higher bandwidth. I see a lot of filters with a roughly 100nm bandwidth - I tested LED's with similar bandwidth.
Here's an example of what I mean, this is an image of two prototypes I have used in testing:
https://drive.google.com/file/d/1HrbHdutuBNgcYaP3phnc48F7RnXCre8V/preview
The brighter light is an Epistar 10W LED (9v - I had to rewire the 6xAA battery insert into a series to make this ridiculous beast work), and the dimmer light is the Cree XTE (3v).
Despite the Cree being a much dimmer light, it produced much more detectable fluorescence than the Epistar, as you can see in this graph:
https://drive.google.com/file/d/1-ezTv_VfOm8MBEA-twWae87J2lRcMzYn/preview
This is not because the Epistar didn't induce fluorescence, it's because the fluorescence it did produce was surrounded by nonfluorescent tissue illuminated by the non-reactive parts of the bandwidth. So I am concerned I'll order an expensive filter just to encounter the same issue.
But I suppose the only way to know for sure is to just order the damn thing and do a comparison.
Yeh, most of the blues are SiC and similar, whereas reds are GaAs, GaAsP, etc.
Still, if he needs to be picky as far as wavelength, he might have to pick an operating point based on current and stick with it.
There ya go…