Replacement of carbide lamps by LED lamps

[yellow]

build my own test device with 10 switchable places for emitters/optics, so I'm able to test 10 different configurations at one cave trip.

wow,
I bet You plan to make more than just one lamp - quite some efforts ...

 

[5kids]

Have you considered trying an Princeton Tec apex and replace the LED with a warmer emitter? The apex is rated to 1M water depth. They now have an exteme version with a remote battery 8AA pack, or you could use our own battery pack with any of the apex varients. The apex mod is easy and you can pick your tint of emitter to use. I love my modded apex.

BTW, Princeton Tec has no info on the website regarding the apex extreme.

http://www.brightguy.com/products/Princeton_Tec_APEX_Extreme.php

 

[Tobias Bossert]

wow,
I bet You plan to make more than just one lamp - quite some efforts ...

Yes, this is right, we plan to build our own headlamps for some members of our caving club.
***We definitly do not plan to sell those lamps***
Quite the other way round: Any results - failure as well as found optimization - will be posted here. So we hope that some manufacturers will take some ideas from this discussion and develop a new headlamp or optimize an existing model. This would be very helpfull for us and othes too, because everyone can avoid building final lamps by hand!

But up to now, I know only one really good commercial headlamp for caving (www.scurion.ch) - much too expensive! And even this one could be further improved with respect to color temperature and pattern of its bare emitter working light.
There is no option at the moment to choose one model on the market and modify it with actual highly efficient LEDs.

Last but not least: you should calculate neither the time spent to experiment with this things nor the money you spent to build your test-devices and pre-models, it's just our hobby! The efforts you mentioned, that is what enyoys!

 

[uk_caver]

But up to now, I know only one really good commercial headlamp for caving (www.scurion.ch) - much too expensive! And even this one could be further improved with respect to color temperature and pattern of its bare emitter working light.

I guess it depends what you mean by 'improve'.

White LEDs can seem cold, but yellower ones aren't necessarily much better. In the days when I made luxeon-powered lights, and the LEDs I got weren't sold by colour bin, about half were 'proper' white, 30% yellowish, with the rest a little blue (apart from a couple with a purple tint).
Going straight from using carbide, the yellowish ones maybe seemed a bit less cold than the white ones, but didn't seem quite as good for actually lighting things up, or for colour accuracy.
Maybe for a wide beam, a red or amber LED could be added to the mix to give a more carbide-like light, but I'm not sure many people I know would go for that if it meant extra power consumption/loss of efficiency. It'd also be trickier to match in a focussed beam.
Equally, though there can be initial comments when switching to a 'colder' light, that does seem largely to be a matter of adjustment.
Commercially, tint doesn't seem to be much of a selling point, and I'd guess that minor complaints would generally be about something being too off-white in any direction (blue, yellow, red, or green).

Also, 'improving' a wide beam is still a matter of taste. Though I'm happy with what I do regarding narrowing a wide beam somewhat, there are certainly people around here who really want a complete 180 degree spread.
I haven't had any complaints about my approach, but given it was completely free and easily reversible, there wouldn't be much to complain about, and having 2-3x better forwards throw compared to a naked LED gave it a definite advantage.

 

[Bandgap]

Hi Tobias

From my very limited caving experience, I would say arrangement D is the best option as it can be difficult to get your hands to your chest to operate light controls. And the controls would have to stand regular scraping along the ground without adjusting.
You will have fun squeezing the electronics into that small space.

I have been debating battery technology for caves.
Lithium ion seems ideal, except that the contents are inflammable and in amazingly rare circumstances they have been know to burst into flames - from over-charging - not a problem for you - over heating or physical blows, crushing or piercing.

So A nice strong box with not too much thermal insulation if you are running them hard.

Also, a question - what cable are you thinking of using and how would you attach the cable at both ends as these seem to snag on rocks ad get pulled hard.

As for colour rendering - whatever the apparent colour temperature, 'white' leds have very little green and almost no red in their spectrum, so colour rendering is poor compared with a well-driven halogen lamp.
I suspect, being incandescent, a carbide lamp had plenty of red and also most of the spectrum - at least as far as blue.

And - my preference and I know some people here disagree strongly - I favour adjustable light output - either by continuous control or in small steps - or at least three level including an I-am-only-eating-my-sandwiches-and-don't-want-to-blind-everyone setting.

Happy caving

Steve

 

[uk_caver]

As for colour rendering - whatever the apparent colour temperature, 'white' leds have very little green and almost no red in their spectrum, so colour rendering is poor compared with a well-driven halogen lamp.
I suspect, being incandescent, a carbide lamp had plenty of red and also most of the spectrum - at least as far as blue.

And - my preference and I know some people here disagree strongly - I favour adjustable light output - either by continuous control or in small steps - or at least three level including an I-am-only-eating-my-sandwiches-and-don't-want-to-blind-everyone setting.
Steve

I can't say I've noticed a colour rendition problem with LEDs when caving, whether illuminating rock or other cavers. Even side-by-side with halogen or carbide, the only obvious difference seems to be in terms of colour temperature, not rendition. I guess there just aren't enough things underground where green is important.
The place where there does seem to be a difference is on the surface, illuminating grass, etc.

I like having multiple powers, but after playing around, for practical purposes I reckon power steps of less than a factor of 2 or 3 are probably too close to be useful.

 

[Bandgap]

I can't say I've noticed a colour rendition problem with LEDs when caving, whether illuminating rock or other cavers.....

I like having multiple powers, but after playing around, for practical purposes I reckon power steps of less than a factor of 2 or 3 are probably too close to be useful.

Just thought I would point out the colour thing in case there are purists wanting to enjoy strata to the full. I suppose limestone is pretty dull wherever you find it.

As for the steps, I think steps of 2x or 3x brightness are pretty good- as the eye takes the log (or whatever) of this and makes it apparent steps of 1.4 to 1.7 (ish).
But they do need to go down far enough IMHO.

Apart from the cable, my biggest quastion was how to make a switch or control robust enough while keeping it small.

Steve

 

[TorchBoy]

As for colour rendering - whatever the apparent colour temperature, 'white' leds have very little green and almost no red in their spectrum, so colour rendering is poor compared with a well-driven halogen lamp.

Perhaps you're thinking of blue leds?

 

[uk_caver]

Just thought I would point out the colour thing in case there are purists wanting to enjoy strata to the full. I suppose limestone is pretty dull wherever you find it.

I'm not sure about dull, but most of the stuff I've seen underground tends to be somewhere between white, black and brown.
However, to me, the various kinds of red caving gear still look essentilly the same red under white LED lighting as under daylight, so there's enough red light to allow the brain to construct the impression of a full spectrum.
I don't need any more red than that, though I can understand some mine explorers looking at subtly different minerals might have a different view.

Apart from the cable, my biggest question was how to make a switch or control robust enough while keeping it small.

I guess a lot comes down to the casing. If you're making your own housing and trying to avoid possible water ingress points, it's certainly possible to do something with internal reed switches/external magnets. With a cylindrical housing, making an external rotating ring with a magnet in, (as some people do with torches) is probably not too hardto at least get working, though making one that looks good may be rather harder (assuming there isn't some objet trouve lying around that does the business).

 

[2xTrinity]

I guess it depends what you mean by 'improve'.

White LEDs can seem cold, but yellower ones aren't necessarily much better. In the days when I made luxeon-powered lights, and the LEDs I got weren't sold by colour bin, about half were 'proper' white, 30% yellowish, with the rest a little blue (apart from a couple with a purple tint).
Going straight from using carbide, the yellowish ones maybe seemed a bit less cold than the white ones, but didn't seem quite as good for actually lighting things up, or for colour accuracy.

I think the old luxeons that were essentially cold white LEDs that ended up being "yellowish" by chance were not the same as current-generation warm white Cree LEDs, which appear closer to blackbody and less "greenish" than bin outliers that just happen to be warmer.

Equally, though there can be initial comments when switching to a 'colder' light, that does seem largely to be a matter of adjustment.

I believe different color temps suit different applications. For example, or trying to do intensive shop/design work, I like 5000k. For "ambient" light, I like 3000k. In most other cases, I prefer a compromise of ~3500k. I generally don't like to stray outside the range of 3000k-5000k.

Also, 'improving' a wide beam is still a matter of taste. Though I'm happy with what I do regarding narrowing a wide beam somewhat, there are certainly people around here who really want a complete 180 degree spread.

If you want a 180% spread, do your buddies a favor and use a diffuser over the LEDs -- this will help with glare immensely. You can buy lenses from Flashlightlens.com with diffusion film and AR-coatings already installed. I'd recommend a shallow reflector, with a diffuser. This will produce a flood light that still has slightly more intensity in the center of the field of vision (similar to what you'd probably get with a "clean" carbide lamp).

As for colour rendering - whatever the apparent colour temperature, 'white' leds have very little green and almost no red in their spectrum, so colour rendering is poor compared with a well-driven halogen lamp.

Most current-generation white LEDs emit the majority of their power in what would be considered the "green" portion of the spectrum -- that is where the high lumens ratings come from (green is where the eyes are most sensitive).The reason older LEDs were much more bluish was that phosphor depositing process was less efficient and would impede light extraction too much, so thinner layers were most efficient. That's improving though -- the fact we have Warm Cree LEDs at Q2 efficiency is evidence of that.

The problem with color rendition is essentially the same as old single-phosphor fluorescents used to have -- the "red" output is actually more like orange, which is enough to counterbalance the blue and make the overall packaging look white, but provides little contrast when distinguishing shades of red/brown. A potential solution to that is to actually use multiple phosphors with the LEDs, just like modern multi-phosphor fluorescents.

Maybe for a wide beam, a red or amber LED could be added to the mix to give a more carbide-like light, but I'm not sure many people I know would go for that if it meant extra power consumption/loss of efficiency. It'd also be trickier to match in a focussed beam.

There isn't much advantage to adding amber -- the only real deficiency is deep red wavelengths. Adding a true red (rather than amber or red-orange) to an array of white LEDs, particuarly ones with a "greenish" off-white tint does produce some very good results. Efficiency loss is neglibile, a surprisingly small amount of red is needed to "fill in the gaps" in the spectrum.

The problem though is with color mixing/packaging. An actual Emitter package that included a red die INSIDE the package would be pretty interesting, but probably not worth the trouble compared to a second phosphor.

 

[2xTrinity]

I guess it depends what you mean by 'improve'.

White LEDs can seem cold, but yellower ones aren't necessarily much better. In the days when I made luxeon-powered lights, and the LEDs I got weren't sold by colour bin, about half were 'proper' white, 30% yellowish, with the rest a little blue (apart from a couple with a purple tint).
Going straight from using carbide, the yellowish ones maybe seemed a bit less cold than the white ones, but didn't seem quite as good for actually lighting things up, or for colour accuracy.

I think the old luxeons that were essentially cold white LEDs that ended up being "yellowish" by chance were not the same as current-generation warm white Cree LEDs, which appear closer to blackbody and less "greenish" than bin outliers that just happen to be warmer.

Equally, though there can be initial comments when switching to a 'colder' light, that does seem largely to be a matter of adjustment.

I believe different color temps suit different applications. For example, or trying to do intensive shop/design work, I like 5000k. For "ambient" light, I like 3000k. In most other cases, I prefer a compromise of ~3500k. I generally don't like to stray outside the range of 3000k-5000k.

Commercially, tint doesn't seem to be much of a selling point, and I'd guess that minor complaints would generally be about something being too off-white in any direction (blue, yellow, red, or green).

I believe a lot of the earlier "warm white" LEDs were essentially off-white toward yellow-green, as they still had too little red. Modern warm emitters, like the Cree Q2 linked earlier, are reputed to be muhc stronger in red -- so that it appears white -- but still not as good as a halogen in that respect.

Also, 'improving' a wide beam is still a matter of taste. Though I'm happy with what I do regarding narrowing a wide beam somewhat, there are certainly people around here who really want a complete 180 degree spread.

If you want a 180% spread, do your buddies a favor and use a diffuser over the LEDs -- this will help with glare immensely. You can buy lenses from Flashlightlens.com with diffusion film and AR-coatings already installed. I'd recommend a shallow reflector, with a diffuser. This will produce a flood light that still has slightyly more intensity in the center of the field of vision (similar to what you'd probably get with a "clean" carbide lamp).

As for colour rendering - whatever the apparent colour temperature, 'white' leds have very little green and almost no red in their spectrum, so colour rendering is poor compared with a well-driven halogen lamp.

From LED Museum, Spectrum of Cool White Cree LED:

LEDs have no deficiency in green -- most current-generation actually emit the majority of their radiant output in the yellow/green region of the spectrum. The problem is same as old single-phosphor fluorescent used to have -- the "red" output is actually more like orange, which does make the overall packaging look white, but provides littel contrast when distinguishing say shades of brown. The best engineering solution to that would be to actually use multiple phosphors with the LEDs, just like modern multi-phosphor fluorescents.

Maybe for a wide beam, a red or amber LED could be added to the mix to give a more carbide-like light, but I'm not sure many people I know would go for that if it meant extra power consumption/loss of efficiency. It'd also be trickier to match in a focussed beam.

There isn't much advantage to adding amber -- the only real deficiency is deep red wavelengths. Adding a true red (rather than amber or red-orange) to an array of white LEDs, particuarly ones with a "greenish" off-white tint does produce some very good results. Efficiency loss is neglibile, a surprisingly small amount of red is needed to "fill in the gaps" in the spectrum nicely. I believe some of the warm white LEDs in the past have included multiple phosphors to achieve the same effect, and i expect more to in the future as LEDs are pushed for general lighting.

 

[uk_caver]

I think the old luxeons that were essentially cold white LEDs that ended up being "yellowish" by chance were not the same as current-generation warm white Cree LEDs, which appear closer to blackbody and less "greenish" than bin outliers that just happen to be warmer.

I'm sure the colour was more accidental, though of all the Lux Is/IIIs I used, I don't recall any looking noticeably green.
A friend's commercial LuxIII-powered light does have a definite green cast, and though I guess the eye would adapt to it, when there are cleaner white lamps around, it does look rather nasty.

There isn't much advantage to adding amber -- the only real deficiency is deep red wavelengths. Adding a true red (rather than amber or red-orange) to an array of white LEDs, particuarly ones with a "greenish" off-white tint does produce some very good results. Efficiency loss is neglibile, a surprisingly small amount of red is needed to "fill in the gaps" in the spectrum.

Amber might give the carbide feel some people have fond memories of. I'm not sure how many people with carbide nostalgia are really too concerned about colour rendition.

The problem though is with color mixing/packaging. An actual Emitter package that included a red die INSIDE the package would be pretty interesting, but probably not worth the trouble compared to a second phosphor.

I guess a twin die brings up the problem of sharing power in proportion (or having twin drivers), whereas an extra-phosphor approach does balancing automatically. Maybe when use for indoor lighting takes off, multiphosphor white LEDS would find a ready market.
Still, for a personal experimental light, reddening a wide beam with an extra LED is certainly possible, given enough space in the headset.

 

[Beaker]

Hi Tobias, I'm quite surprised that no one has mentioned the Stenlight S7 in this thread (http://www.stenlight.com/) -- it's a purpose-built caving lamp that is rapidly growing in popularity among USA cavers. I don't own one but many of my caving friends do, and it's really an impressive piece of work in terms of construction, light output, and efficiency. It's not cheap -- a basic setup of light + custom Lithium-Ion battery sells for about US$300 or so -- but all the cavers I know who have them have been extraordinarily pleased with them.

There are discussions about it both on this board and at cavechat.org (the US NSS discussion board), eg

http://www.forums.caves.org/viewtopic.php?p=296&sid=85eec0fe766dd5347106387cb3516203

If you are looking for a caving-specific LED headlamp, I'd highly recommend you check out the Stenlight if possible. (I'm assuming it's available in Europe...)

 

[uk_caver]

Doesn't the Stenlight run both LEDs in series?
If so, it's not great for people looking for a smooth (non-spot) beam unless they lose the optics on both LEDs and then get another light for spot use.

 

[Tobias Bossert]

If you are looking for a caving-specific LED headlamp, I'd highly recommend you check out the Stenlight if possible. (I'm assuming it's available in Europe...)

Yes, I read the thread upon Stenlight headlamp and I looked at the page of the manufacturer. It seams to be a good and robust headlamp indeed. But it still doesn't comply with our requirements as I posted in the beginning of my thread.
There are two LEDs in series, always used together. Each using its own optics. Consequently the only possibility to change the spacial pattern under use would be to attach and remove an additional diffuser, not very handsome in caves...
The beam is a compromice between throw and flood: It neither meets one of both requrements. We need flood working light all over the day and for short term usage only super narrow beam (without spill).
On the other hand it provides 70 lumens output for 3 to 5 hours only (consuming 4 to 5W, not quite good nowadays). So you will need 2 or 3 battery packs (up to 6 pieces of 18650 cells) per day...

There is indeed a much more suitable headlamp on the market, www.scurion.ch, but I would like to surpass it technically and nevertheless avoid its high price (about $ 500).

 

[uk_caver]

Leaving aside the underground practicalities, how easy would it be to fully diffuse a spot beam without losing a deal of light in the process?
Wouldn't a diffuser tend to cause backscatter?

 

[Tobias Bossert]

After some month of experimenting I come back to this treat. We tested all kinds of optics – now I can demonstrate some reasonable solutions.

1) Normal light for walking and climbing

Neither TIR with narrow, medium or wide beam nor plain LED satisfied our requirements. On the one hand all TIR are not wide enough for walking (you tent to stumble) but on the other hand, plain LED without any optics is to fade (you can't see where to go).
We tested many combinations of two simultaneous LED and find one combination satisfying us in real cave tests: Combination of a medium beam with FWHM of about 16° and a very wide beam with FWHM of about 80°.
I realized this with XR-E Q5 using Carclo ripple medium (10210 & 10205) and with SSC P4 U using Carclo wide angle reflector (10170 & 10363). The axis of the medium optics is horizontal (when you stand upright) and the wide reflector is declined downwards 20 to 30°, as shown in Figure 1.

This free mounting without case was for photo documentation only.

It is not easy to demonstrate the feeling of the combined beam in a picture. I put the LEDs 50cm apart from a vertical white wall while the photo was taken from about 150cm from the wall – otherwise I would need a fisheye photo lens losing the orientation. The medium beam is horizontal and the pattern therefore is circular. The wide beam is declined downwards and therefore projects elliptical; it just reaches the area immediately in front of your feet. In reality you will not realize the sharp boundary of the beam, this is due to the photo taken from thee times the distance from the wall.

Both LED were driven by different current levels. The XR-E Q5 for the medium beam needs about 30% of the current of SSC P4 U in the wide angle reflector to achieve the impression of the figure above. We plan to provide two levels: 40m + 130mA and 90 + 300mA, so we can hold this level for about 24 or 10h with 6 alkaline AA.

2) Pause light

We plan to use the above mentioned medium beam with a current level of about 20mA.

3) Spot to explore deep pits or high shafts

At the beginning I was sure we should use Gaggione Mobdar or Gaggione LL3 because so many users mentioned them positive. But these TIR are relatively big and provide a lot of side spill too. There exist completely different TIR ('catadioptric reflector optics' using total reflection on the front plain and being metalized on the backside, e.g. Carclo 10144) with extremely narrow beam, but these beams are too narrow for us (at the ground of a deep pit you will see one big stone only instead of the floor of the pit completely). We tested almost all TIR and reflectors from Kaidomain and Dealextreme and others and found nothing satisfactory.
Cree XR-E has its own internal optics narrowing the beam a little bit, so it is possible to gather almost all of its flux with a thick aspheric collimator lens. We also tested almost all acrylic lenses we found. One very cheep lens with 23mm diameter was suitable and performs quite good. This lens is sold with different holders (DX 4614, DX 4628) which do not fit for XR-E. Therefore we cut the bottom of the holder away and inserted a Carclo holder for Cree (10205). With this combination you can adjust the lens. Both holders form something like a telescope.

This device is much smaller than Mobdar or LL3 as you can see in the following picture.

The combination of the two holders allows adjustin the beam. The beam is narrowest when the lens forms an image of the die on the wall:

The rectangular shape of the die is about 40x40cm at a distance of 5m. As you approach the LED to the lens, the beam gets broader, round but weaker too; the left beam is about 150cm in diameter at 5m.

I also compared the single lens with Mobdar. The beam of the lens is set to about 70cm at 5m (rounded but not jet round). The peak intensity of illumination in the center of the beam of Mobdar is about 70% of this, but it seams that the overall flux of the Mobdar is higher and the beam is much more "beautiful". On the other hand the beam of the lens is "without" spill, that's what we need.

We plan to use the single aspheric lens with XR-E Q5 at about 1A. This is for short term use only. It works fine with 6 fresh alkaline AA, but when they got discharged they will limit the current rather than the regulator. With NiMH AA cells this is no matter.

4) As much light as possible for big halls

We tested the SSC P7 for this. We found no compact TIR (< 30mm). Since P7 has no internal lens the original beam is too wide to gather all flux with an aspheric collimator lens. Smooth reflectors throw donuts or beams with a dark cross in the center. So you must use orange peel reflectors. But – as I know now – orange is not orange. The roughness should be high enough with respect to the focal length. Most OP reflectors I found at Kaidomain and Dealextreme were not rough enough. Only two of the smaller types can be used with P7 avoiding severe donut effects:
DX 5951 (similar to KD 3317) has diameter of 18mm only. The hole for the emitter must be opened to fit over the dome of P7 (9mm). This gives a relatively wide beam with only a week donut.
DX 3257 has diameter of 26,5mm. You must cut the holder side so that surface on the backside of the hole for the emitter is plain. Additionally you must open the hole to fit over the dome of P7. This gives a medium beam with a week donut.
If you don't like this week donut, use a diffuser from L2Optics or Polymere. But this will cost at least 15% of the flux! If we could find a stronger orange peeled reflector, the donut should be avoidable without diffuser. I just ordered KD 5405 (23mm), KD 3318 (24mm) and KD 5151 (25mm) to try these too. For the moment I prefer the DX 3257. You can see the beam of modified DX 3257 with P7 in the picture below.

We plan to drive this P7 with about 3A. This is for short term use only. It works fine with 6 NiMH AA cells, but with alkaline AA the current will be limited to their strength naturally.

5) Put all together

At the moment we plan to build a headlamp with 4 LEDs and optics as described above. This should get not to big for a caving helmet, as you can see on the following picture:

But what is about the housing?
No idea at the moment – I still hope that some manufacturer takes my ideas and starts its own development on that base.
But at the moment, there is no such manufacturer. Therefore I have to continue...
For this I assembled the circuitry for such a headlamp using cheep drivers from far east:

Next I will construct a prototype and lend out to my caving friends to optimize it. This will consume some time...

P.S.: Moderator please help!
The links to my pictures work well, but I can't insert images. After saving changes, they were gone. Why? Please, can anyone tell me what I'm doing wrong:
I push the button 'insert image' than the window for the URL appears with default 'http://'. When I overwrite it by the link and accept, than the text window shows an icon at the place of insertion. After 'save changes' no pictures appear but all icons are gone too...
P.S.: Problem solved now:
I used Firefox 2.0.0.16 with add-in Control de Scripts 1.03 and add-in AVG Save Search 8.0, which do not insert images correctly. With older versions of Firefox or with MS IE there is no such problem.

 

[Yucca Patrol]

Tobias,

I applaud your effort and thoroughness. As a fellow caver, I am always searching for the ultimate caving headlamp and will look forward to seeing what you end up with. :twothumbs

 

[uk_caver]

A four-LED light? That does seem a fairly serious project.

Might it not be easier as a general 2-led caving light, and a more 'specialist' room-lighting and shaft-spotting light?

I imagine there might be some kind of market for a specialist light among the large numbers of people who already have some other kind of caving light that they're happy with (Duo LED, etc)
Possibly a specialist light would be better as a handheld unit, since for shaft-examining (at least when looking upwards), a helmet mounted spot beam can literally be a pain in the neck.
When I was playing with some P7s, the main use I found for a very bright flood beam was for photography, which is also a use where a lamp is more useful in the hand than on the helmet.

Also for the P7 room-light, have you tried using a conical reflector? If you don't want any kind of beam, just a rather more forward-biased light than you get from a naked LED, it's definitely worth trying.

With your FWHM of 80° using reflectors - looking at the datasheet, a naked Cree has a FWHM of only about 90°.
How different is the reflectored Seoul to a naked Cree?
Is it just a case of the Seoul+reflector eliminating 'wasted' extreme-wide light?

 

[Tobias Bossert]

A four-LED light? ...
Might it not be easier as a general 2-led caving light, and a more 'specialist' room-lighting and shaft-spotting light? ...
Possibly a specialist light would be better as a handheld unit, since for shaft-examining (at least when looking upwards), a helmet mounted spot beam can literally be a pain in the neck.

Yes, uk_caver, I know this discussions. There are pros and cons. About 20 years ago I constructed my first handheld halogene thrower (10W 6V in the housing of flashlight for 'flat battery' 4,5 V with 6 NiCD AA). Works very well at that time. Good when standing at the floor of a high shaft in the ceiling, because there is no problem taking it out of your baggage ant it is good for your neck. But bad in case you are descending a rope into a pit. On my personal icon you can see me with carbide lamp plus halogene thrower on my helmet (constructed 1995).
Since the additional space and weight of the LED optics are very small, I decided to integrate all three lamps (working light with 2 LED, thrower with 1 LED and flooder with 1 LED) into the headlamp.

Also for the P7 room-light, have you tried using a conical reflector? If you don't want any kind of beam, just a rather more forward-biased light than you get from a naked LED, it's definitely worth trying.

Yes, I tried P7 with conical smooth reflector Carclo 10170. The beam is not homogeneous. I will look for some orange peeled alumine to try it for my own. But at the moment I have no source for alumine sheet with high orange peeling.

With your FWHM of 80° using reflectors - looking at the datasheet, a naked Cree has a FWHM of only about 90°.
How different is the reflectored Seoul to a naked Cree? Is it just a case of the Seoul+reflector eliminating 'wasted' extreme-wide light?

There are some differences.
1) XR-E has its maximum on axis and decreases the intensity continuously apart from it; at about 45° off axis intensity is half maximum. If you combine this with a narrow beam 30° off axis, than you get two maxima in the vertical. SSC P4 with Carclo 10170/10363 gives an flat intensity over almost 80° and than drops fast. If you combine this with a narrow beam 30° off axis, than you get one maximum only in the vertical.
2) XR-E has reasonable spill outside of its 90° beam. This may dazzle your friends when talking with them. With SSC P4 and reflector you just have to decline your head about 10 to 15° and the eyes of your friends are out of beam.

But the problem of dazzling is still unsolved. I experimented wit diffusors of Cree (L2-Optics) and Polymere, but are not happy with it. Both work well to soften and broaden a narrow beam (reducing the overall flux about 15%). But used with a extremely wide beam, the losses ar much higher and increase with off axis angle. This could be solved only by spherical or domed diffusors, but I could not find such devices on the market.

Regards Tobias