Multi-stage 1AA Review - Part III: Runtimes, beamshots & more!

selfbuilt · 27.04.2011, 12:27

Multi-stage 1AA Review - Part III: Runtimes, beamshots & more!

REVIEWER'S NOTE: This Part III thread is an update of my earlier multi-stage 1AA round-ups, showcasing the latest crop of currently available lights. To see comparisons with older lights, please see my earlier threads:

Part I: Rexlight 2.0; DealExtreme X.1, X.V; Jetbeam MkIIX, C-LE v1.0, C-LE v1.2; Fenix L1D-CE (P4)
Part II: Rexlight 2.1; DealExtreme X.1; Jetbeam C-LE v1.2; Fenix L1DCE (Q2), L1D-R100, L1Tv2.0-R080; MTE 5-Stage Cree, 5-Stage R100; Kai 5-Stage Cree; Lumapower LM-301; Liteflux LF5; Ultrafire C3 5-stage

UPDATE 8/27/2008: This thread needs an overhaul, but in the meantime here are some detailed review of new 1AA lights: Jet-I PRO IBS, NiteCore D10,and LiteFlux LF5XT.

UPDATE 4/25/2008: Added JetBeam Jet-I MK IBS (Q5) to the review.

UPDATE 3/4/2008 : Added Zebralight H50Headlamp (Cree Q5) to the review.

UPDATE 2/5/2008: Added the JetBeam Jet-I PRO (Cree R2) to the review comparisons.

The contenders:

From left to right: Liteflux LF5 (SSC); Rexlight 2.1 Build 002 (Q2); Jetbeam C-LE v2 (Q2), EDGETAC NiteCore DI (Q5); Fenix L1D (Q5); Olight T15 (Q5); Fenix L1D (R100); LumaPower LM31 (SSC). Not shown are the Jet-I PRO, Zebralight H50 (Q5) headlamp and JetBeam Jet-I MK IBS (Q5).

Testing Method: All my output numbers are relative for my home-made light box setup, a la Quickbeam's FR.com method. My relative overall output numbers are typically similar to his, although generally a little lower. You can directly compare all my review graphs - i.e. an output value of "10" in one graph is the same as "10" in another.

Throw values are the square-root of lux measurements taken at 1 meter from the lens, using a light meter.

Beamshots:

All up-close beamshots at ~0.4 meter from a white wall, to show you the different overall patterns. For Jet-I PRO, Zebralight H50 headlamp and JetBeam Jet-I MK IBS (Q5) beamshots, please see my detailed reviews of those lights.

On Max/Turbo/100% with Sanyo Eneloop (2100mAh NIMH)

Runtime Comparison

For all comparisons, “Hi” mode means the maximum output mode of the light (i.e. Max/Turbo/100%, depending on terminology used for that light). I've added an "*" for the Zebralight entries as a reminder that its difficult to compare overall output levels to the other lights in the review (due to its flood-only beam and different positioning in the lightbox).

Hi mode on Duracell 2650mAh NiMH
(except for LM-31 which is on Sanyo 2500mAh since my Duracells wouldn’t fit)

Hi mode on Sanyo Eneloop 2100mAh NiMH LSD (low self-discharge)

Hi/Med/Lo modes on Alkaline Duracells

Hi/Med and Med/Lo modes on an AW Protected 14500 (3.7V, 750mAh)

Output/Throw Summary Chart

Light Summaries:

Liteflux LF5 (SSC)

Uses high freq PWM for lower modes – I can't measure the freq with my setup, and it is obviously quite high as it produces no noticeable flicker in my testing.
Standard output modes are 15% and 50% output, and I've set user-defined to either highest (100%) or lowest (<1%) for testing purposes. Light has an “infinitely variable” output control mechanisms for user defined purposes, but not visually linear like the Nitecore DI (see my NiteCore DI review thread for a graphical comparison of the ramping sequence).
Good build quality and nice smooth beam thanks to the SSC emitter. However, overall spillbeam width is narrower than most the other lights reviewed here.
Very good output and runtime on NiMH and 14500, somewhat similar to the Rex 2.1, although LF5 is one of the best regulated lights I've seen.
Runtime on alkaline is rather disappointing on all levels - consistently lower than the other lights tested here.
Light features lowest low mode of all lights tested here – in user-defined mode, you can get the light down to ~0.3 on my output scale. Continuous runtime is an impressive 15 days on alkaline.
Twisty interface is a bit unusual and takes some practice at first, but does allow you to avoid the strobe/SOS modes if you want. Light has functioned fairly reliably, but you may need to tighten the hidden retaining ring in the head from time to time (if mode switching becomes unreliable, turn the visible disc in the head until the holes line up with the hidden retaining ring holes, and tighten with a pair of fine tweezers or snap-ring pliers).
Light comes as a nice kit with loads of goodies - please see my detailed LF5 review for more info.
Although sometimes hard to find, light retails new for ~$50.

Rexlight 2.1 build 002 (Q2):

The original Rex 2.0 used a very noticeable 94 Hz PWM for its low modes, but I am unable to detect PWM freq on the 2.1 versions.
Rex 2.1 versions have 3 sets of light sequence states that you can access ("Daily/Tactic/Advanced"). For a full explanation of the interface, see my Rexlight comparison review here. Switching between them can be a bit annoying (i.e. accidental mode switching when in Lo).
Switch is a good quality reverse clicky, and light has a memory mode of last output level used.
All light modes now work as they are supposed to, on all battery types, with excellent runtimes in all modes now (Rex 2.0 had a number of issues and inconsistencies).
Output on NiMH has been restored to previous Rex 2.0 Hi levels (first Rex 2.1 build had reduced max output on NiMH).
Excellent performance on 14500. Light also features a 3.2-3.0V "low voltage" warning (i.e. light strobes), so you can safely use unprotected cells
Very nice low mode, also available on 14500. Not as low as the NiteCore DI or Liteflux LF5, but quite respectable (especially on 14500).
Well worth the upgrade - only downgrades from 2.0 are in beam "ringiness" on my sample, plus some UI issues with the 3 light states.
For a full comparison to the Rex 2.0 and earlier build Rex 2.1, see my review here.
Light retails new for ~$40.

JetBeam C-LE v2.0 (Q2):

Basically a reverse clicky version of the popular C-LE v1.2. Clicky is good quality (though fairly stiff) and facilitates mode changing (which could previously be unreliable in twisty version when the threads got dirty).
The original C-LE v1 (and DX X.V knock-off) both had a very noticeable 73 Hz PWM. C-LE v1.2/v2.0 both use a more respectable 317 Hz.
C-LE v1.2/v2.0 also have improved threads over previous versions, are a bit thicker in diameter, and features a 2 sec memory mode that retains last output setting used.
Output/runtimes have not changed appreciably from v1.2. Main improvement is a good quality reverse clicky, which makes it easier to reliably switch modes.
On all levels, the C-LE v1.2/v2.0 have increased output compared to earlier versions, with either similar or longer runtimes. Note that this means low mode is not as low as previous versions.
Max output is still somewhat on the low side for a Cree-based light – closer to the SSC lights reviewed here.
See my v2.0 review here.
Note: there has been at least one review showing different output levels in a recent shipping C-LE v2.0. Not sure if this was a defective light, or a Manufacturer’s change in specs – my C-LE v2.0 comes from the first production run.
Although sometimes hard to find, the light retails new for ~$30.

EDGETAC NiteCore Defender Infinity (Q5):

EDGETAC confirms that the light uses PWM, and the frequency is >1 kHz. I'm thinking it's likely considerably above, since I can’t measure it with my setup or detect any hint of it by eye.
The Nitecore allows you to set your low mode (user defined mode) through a continuously variable brightness mechanism similar to Liteflux - but easier to use and more visually linear, as described in my comparison review). This is a very powerful feature, since it allows you to choose your own custom output level.
The light is controlled by a forward clicky switch (i.e. momentary on, followed by a click to lock-on).
User interface is quite elegant – easy to use, simple and straight-forward - I found it very intuitive. Don’t have time to explain it here, see my NiteCore DI comparison review thread. Note that strobe is neatly tucked away so you don’t need to see it if you don’t want to.
No doubt about it - this is the brightest light on standard AA batteries (alkaline or NiMH) that I have ever tested. Runtimes on Hi are very respectable on these battery types.
Output and runtime on 14500 is very good, with output somewhat brighter than regular batteries (although not the brightest I’ve seen – the Fenix, Olight and some JetBeam models are brighter on 14500, but shorter lasting).
Regulated output on all battery types in all output modes is impressive (e.g. unlike Fenix/Olight on 14500, where you loose low modes until light drops to regulated level).
Very good throw for its size, comparable to Fenix/Olight on standard batteries (less on 14500).
For a lot more info, please see my NiteCore DI review here.
Light retails new for ~$80.

Fenix L1D (Q5 and R100):

The Cree Q5 is slightly brighter than the Rebel R100 version, but with less runtime
The Rebel has a more SSC-like beam (i.e. wider spill, smaller hotspot, smoother transition from hotspot to spill). The Cree Q5 has a more traditional Cree beam profile with visible dark rings, despite the OP reflector.
To my eye, color rendition outdoors seems better on the Rebel lights compared to the Cree lights
Fenix remains the output/runtime king on standard batteries - nothing else beats the efficiency of the Fenix's current-controlled low modes. I just wish they could find a way to make lower low modes.
On 14500, you loose all the low modes initially, as light runs in direct drive from max output until it hits the low mode level. At that point, it then switches into regulation – but only briefly, as the battery is typically almost exhausted by then. Note that Turbo mode (i.e. head fully tightened) is difficult to do on 14500, due to extra length of the protected 14500 (I had to use a copper ring spacer for it to make contact and screw down tight).
Cree Q5 version has a bit more overall output than the Rebel R100 on standard batteries, but peak throw (i.e. center of hotspot) is about the same.
See my original Fenix Rebel vs Cree: OUTPUT, RUNTIMES, BEAMSHOTS thread for direct comparison of R100, Q2, and P4 version results
See my Fenix L2D Q5 vs R100, R80, Q2, P4 Comparison Review: RUNTIMES+for an analysis of the premium Cree Q5 and Rebel R100 heads on L2D form.
Although the R100 is hard to find, the lights retail new ~$55.

Olight T15 (Q5)

The Olight uses a current-controlled circuit for its various modes, with very similar output and overall characteristics as the Fenix. So close in fact that I suspect it is a variant of the same circuit.
Output/runtimes were generally very consistent with my Fenix L1D Q5, except for Med-Lo modes on alkaline where the Olight runtimes were a bit less for the same output.
Spillbeam width is narrower on the Olight compared to the other Crees, but overall output and max throw are similar to the Fenix L1D.
Build quality is very high. Light has a very substantial feel, bulkier than the Fenix model, but otherwise with very similar overall build features.
Press the reverse clicky tailswitch to activate the light. Soft-pressing the tailswitch will move you into strobe followed by SOS modes. Light intensity is controlled by loosen-tighten switch of the head.
There are 5 distinct output modes controlled by the head switch, although the first two are identical on regular batteries in my sample. Switching sequence is from Max to Min output. Light has a memory mode retaining the last level used.
As you’ll see in the runtimes, output at Level 5 (Lo) is identical to the Fenix Lo, and Level 3 is identical to the Fenix Medium. I haven’t tested the other levels yet.
On 14500, you loose all the low modes initially, as light runs in direct drive from max output until it hits the low mode level, at which point it briefly enters regulation before the battery is exhausted (exactly like the Fenix L1D).
For all intents and purposes, the Olight circuit seems to be exactly the same as the Fenix L1D except with the addition of an extra Lo-Med mode (i.e. Level 2) and a different way of switching between modes.
Light retails new for ~$50

Lumapower LM-31 (SSC)

Formerly known as the LM-301, this is an updated version of the original Luxeon LM-301 that now features a SSC U-bin emitter.
Features a wider spillbeam than most lights (hard to see on the beamshots, since I had to crop it from another comparison – I no longer own the light). SSC emitters provide a less sharply defined hotspot than Crees, with a smoother transition (i.e. no dark rings)
Max output is quite acceptable on standard batteries, though not as high as the more expensive Cree options out there
Light uses a resistor for Low mode, so is not as efficient at other choices - but still acceptable runtime for the output, with a good long "moon mode."
Can take 14500, but Manufacturer does not recommend it on Low mode - circuit not designed for it, produces reduced output and runtime compared to Hi mode
Build quality quite good - very substantial feel and nice anodizing, but longer than most other lights reviewed here and can't tail-stand.
Battery tube not wide enough to take all high capacity NiMH (i.e. none of my 2650mAh would fit, so Hi runtime is based on a 2500mAh battery)
EDIT 2/29/08: Although the light doesn't tailstand with stock orange tailcap installed, I've been informed that the GID tailcap included as an accessory does allow tailstanding. I no longer own this light, so I can't confirm personally.
Light retails new for ~$30

JetBeam Jet-I PRO (R2)

Light uses PWM for its low modes, but frequency is high enough that I can't detect it by instrument (or by eye in everyday use).
Excellent maximum output and throw on all battery types. This is the best throwing 1AA light I have tested to date. if you are looking for 1AA pocket rocket, this is your light.
Relative for the output, runtimes on Hi are excellent on rechargeable battery types (NiMH and Li-ion), comparable to the NiteCore DI.
Output is well regulated on rechargeable batteries (NiMH and Li-ion) in all output modes (e.g. unlike Fenix or earlier JetBeam lights on 14500, where you loose low modes initially)
Unfortunately, efficiency on Med-Lo mode is significantly lower than the competition. Further, there’s an issue with 14500 on Med having both lower output and lower runtime than Hi (JetBeam has informed me that they plan to fix the Med-14500 issue for the second batch of lights produced).
Runtimes on alkaline are also quite poor compared to the competition, especially on Lo. Similarly, there is also lower runtime on Energizer e2 lithium (L91) compared to the competition (see my detailed Jet-I PRO review for more info).
Output levels of Med-Lo are higher than most lights.
The user interface is vastly improved over earlier JetBeam models (i.e. simplified).
Build quality on my sample is very high, best I’ve seen from JetBeam so far.
Light retails for ~$75.

Zebralight H50 headlamp (Q5)

Much more than just a headlamp, the Zerbralight is a very versatile little flood light.
If it uses PWM for the low modes, then the frequency is high enough that I can’t detect it with my setup, or notice it by eye.
The Zebralight has 3 modes, accessed by a twist of the tailcap (i.e. twisty interface). Light comes on Low when first screwing the tailcap tightly closed. A rapid twist off-on and you move to Medium. Another twist and you are on Hi. There is no memory mode.
Good selection of output levels on standard batteries, with Lo and Medium being considerably lower than most other lights. Runtimes are truly excellent for the apparent output levels.
Light lacks true low modes on 14500, and max output is insanely bright (so much so that I won’t do runtimes at this level – I doubt the heatsinking could handle it).
Interestingly, on Med/Lo the light has a long “moon mode” on protected 14500 (i.e. doesn’t just rapidly drop to zero output when the protection voltage is reached). I stopped the run at ~4.5 hours, and it was still producing a low level of light on my AW protected 14500.
Build quality is top-notch, good machining and anodizing (type III – HA). This light has one of the thinnest layers of aluminum that I’ve seen, but still feels solid. Weight of the bare light is a negligible 18g.
Can tailstand (or headstand for that matter! )
In headlamp mode, the silicone bracket holds the light fairly firmly, but has enough play so you can still rotate the light easily. This means that the light can also rotate on you inadvertently while you are moving around.
Light retails for ~$50.

JetBeam Jet-I MK IBS (Q5)

The IBS versions are JetBeam's new continuously variable circuits, similar to the EDGETAC NDI. See my full review for more details.
All JetBeam lights use PWM, to my knowledge. On this light, the frequency is high enough that I can’t detect it by eye or instrument, even at the lowest output settings.
Highest output on 14500 of any 1AA I’ve seen, by far. Good news for all you output freaks out there! … but I must caution about the risks of excessive heat if run this way in a prolonged fashion.
Runtimes on the highest and lowest output levels are very good on rechargeable battery types, but lower than the competition on standard batteries in the Med-Hi range.
Lowest output mode (5%) is as low as I’ve seen on a multi-stage 1AA light – with output and runtime as good as my EDGETAC NDI. Well done!
The feature set and user interface on this IBS circuit are more advanced than most lights, and is a considerable improvement over earlier JetBeam models. Although straightforward, I found the triple-flash to enter variable brightness and other modes a bit cumbersome compared to the simpler NDI interface.
Light lacks a memory mode, but doesn't need one since you have the flexibility to set 3 programmable states yourself.
Build quality on my sample is very high, consistent with other JetBeam lights. Standard issues like anodizing mismatch, lettering imperfections, and stiff reverse clicky switch with plastic switch retaining ring persist, but overall quality is still quite high.
Light retails for ~$70.

Final Word:

On regular batteries, the NiteCore DI, Jet-I PRO and JetBeam Jet-I MK IBS (Q5) are currently the brightest 1AA lights out there.
The NiteCore DI and JetBeam Jet-I MK IBS (Q5) both have very low Lo modes, “infinitely variable” output control system, ability to retain all levels on 14500, excellent build quality and UI, etc. These lights are very sophisticated - yet easy to use.
The Jet-I PRO is clearly the best thrower of the bunch - a veritable 1AA pocket rocket! Very well built, there are unfortunately some efficiency issues to be resolved on the original circuit, as well as poor overall alkaline performance. However, an IBS circuit upgrade should be available shortly for this light.
The Fenix L1D Q5/R100 lights remain the output/runtime efficiency leaders for their current-controlled low modes (although the Olight and NDI are close). Lower low modes would be nice, as well as the retention of regulated output levels on 14500. Build quality is high with a straightforward interface - it’s hard to go wrong with either model for a general purpose light.
Olight T15 is basically a Fenix L1D Q5 with an extra Med-Lo mode and different switching mechanism. Build quality is fraknly even higher than the Fenix, although the light is a bit bulky in comparison. I just wish the output sequence went from min to max, instead of the other way around – but a minor quibble since light comes with a memory mode.
If you like a light with the lowest Lo mode and long runtime, the LiteFlux LF5 is your best pick (~0.3 on my output scale, with two weeks worth of continuous runtime at this level on alkaline). The LF5 is a powerful light with a lot of features, but somewhat complex to program with its twisty interface. Overall output is less than the Cree lights, and runtime performance on alkaline is unimpressive. However, it does retain all levels on 14500 (including nice low modes).
The Rex 2.1-002 is a great jack of all trades – very good Hi, Med and Lo levels on all batteries, retains output levels on 14500, excellent runtime performance overall. User interface is a bit complex with its 3 sets of output states – easy to accidentally move between them if you like to use the light on low (switching is controlled by a time delay on the lowest setting). But a very strong performer for a Q2 light, and my previous EDC before the NiteCore DI came out.
JetBeam C-LE v2.0 and Lumapower LM-31 are both excellent choices in the low-mid price range of quality lights (~$30). Both lack many of the features of the higher end lights (e.g. 14500 not supported on C-LE, and not recommended on LM-31), but both are good solid choices for general use. I would recommend the C-LE v2.0 for those wanting runtime efficiency, extra modes, or small size, and the LM-31 for those wanting a simple interface with good low mode (e.g. reading at night).
Budget DX/Kai models have been omitted from this Part III round-up since I’ve found their quality quite inconsistent. You never know what you are going to get from one batch to the next - output and build quality is incredibly variable over time.
The Zebralight H50 Q5 is a very unique and versatile little light – more than just a headlamp, this is a true multi-purpose flood lamp with outstanding runtimes.

Whew - there you have it!

I will continue to update this thread as new 1AA lights come out.

27.04.2011, 12:27	1
selfbuilt Увлеченный Регистрация: 26.04.2011 Последняя активность: 10.11.2015 10:40 Адрес: Canada Сообщений: 412 Сказал(а) спасибо: 0 Поблагодарили: 349 раз(а) в 129 сообщениях	Multi-stage 1AA Review - Part III: Runtimes, beamshots & more! Multi-stage 1AA Review - Part III: Runtimes, beamshots & more! REVIEWER'S NOTE: This Part III thread is an update of my earlier multi-stage 1AA round-ups, showcasing the latest crop of currently available lights. To see comparisons with older lights, please see my earlier threads: Part I: Rexlight 2.0; DealExtreme X.1, X.V; Jetbeam MkIIX, C-LE v1.0, C-LE v1.2; Fenix L1D-CE (P4) Part II: Rexlight 2.1; DealExtreme X.1; Jetbeam C-LE v1.2; Fenix L1DCE (Q2), L1D-R100, L1Tv2.0-R080; MTE 5-Stage Cree, 5-Stage R100; Kai 5-Stage Cree; Lumapower LM-301; Liteflux LF5; Ultrafire C3 5-stage UPDATE 8/27/2008: This thread needs an overhaul, but in the meantime here are some detailed review of new 1AA lights: Jet-I PRO IBS, NiteCore D10,and LiteFlux LF5XT. UPDATE 4/25/2008: Added JetBeam Jet-I MK IBS (Q5) to the review. UPDATE 3/4/2008 : Added Zebralight H50Headlamp (Cree Q5) to the review. UPDATE 2/5/2008: Added the JetBeam Jet-I PRO (Cree R2) to the review comparisons. The contenders: From left to right: Liteflux LF5 (SSC); Rexlight 2.1 Build 002 (Q2); Jetbeam C-LE v2 (Q2), EDGETAC NiteCore DI (Q5); Fenix L1D (Q5); Olight T15 (Q5); Fenix L1D (R100); LumaPower LM31 (SSC). Not shown are the Jet-I PRO, Zebralight H50 (Q5) headlamp and JetBeam Jet-I MK IBS (Q5). Testing Method: All my output numbers are relative for my home-made light box setup, a la Quickbeam's FR.com method. My relative overall output numbers are typically similar to his, although generally a little lower. You can directly compare all my review graphs - i.e. an output value of "10" in one graph is the same as "10" in another. Throw values are the square-root of lux measurements taken at 1 meter from the lens, using a light meter. Beamshots: All up-close beamshots at ~0.4 meter from a white wall, to show you the different overall patterns. For Jet-I PRO, Zebralight H50 headlamp and JetBeam Jet-I MK IBS (Q5) beamshots, please see my detailed reviews of those lights. On Max/Turbo/100% with Sanyo Eneloop (2100mAh NIMH) Runtime Comparison For all comparisons, “Hi” mode means the maximum output mode of the light (i.e. Max/Turbo/100%, depending on terminology used for that light). I've added an "" for the Zebralight entries as a reminder that its difficult to compare overall output levels to the other lights in the review (due to its flood-only beam and different positioning in the lightbox). Hi mode on Duracell 2650mAh NiMH (except for LM-31 which is on Sanyo 2500mAh since my Duracells wouldn’t fit) Hi mode on Sanyo Eneloop 2100mAh NiMH LSD (low self-discharge) Hi/Med/Lo modes on Alkaline Duracells Hi/Med and Med/Lo modes on an AW Protected 14500 (3.7V, 750mAh) Output/Throw Summary Chart* Light Summaries: *Liteflux LF5 (SSC)* Uses high freq PWM for lower modes – I can't measure the freq with my setup, and it is obviously quite high as it produces no noticeable flicker in my testing. Standard output modes are 15% and 50% output, and I've set user-defined to either highest (100%) or lowest (<1%) for testing purposes. Light has an “infinitely variable” output control mechanisms for user defined purposes, but not visually linear like the Nitecore DI (see my NiteCore DI review thread for a graphical comparison of the ramping sequence). Good build quality and nice smooth beam thanks to the SSC emitter. However, overall spillbeam width is narrower than most the other lights reviewed here. Very good output and runtime on NiMH and 14500, somewhat similar to the Rex 2.1, although LF5 is one of the best regulated lights I've seen. Runtime on alkaline is rather disappointing on all levels - consistently lower than the other lights tested here. Light features lowest low mode of all lights tested here – in user-defined mode, you can get the light down to ~0.3 on my output scale. Continuous runtime is an impressive 15 days on alkaline. Twisty interface is a bit unusual and takes some practice at first, but does allow you to avoid the strobe/SOS modes if you want. Light has functioned fairly reliably, but you may need to tighten the hidden retaining ring in the head from time to time (if mode switching becomes unreliable, turn the visible disc in the head until the holes line up with the hidden retaining ring holes, and tighten with a pair of fine tweezers or snap-ring pliers). Light comes as a nice kit with loads of goodies - please see my detailed LF5 review for more info. Although sometimes hard to find, light retails new for ~$50. *Rexlight 2.1 build 002 (Q2):* The original Rex 2.0 used a very noticeable 94 Hz PWM for its low modes, but I am unable to detect PWM freq on the 2.1 versions. Rex 2.1 versions have 3 sets of light sequence states that you can access ("Daily/Tactic/Advanced"). For a full explanation of the interface, see my Rexlight comparison review here. Switching between them can be a bit annoying (i.e. accidental mode switching when in Lo). Switch is a good quality reverse clicky, and light has a memory mode of last output level used. All light modes now work as they are supposed to, on all battery types, with excellent runtimes in all modes now (Rex 2.0 had a number of issues and inconsistencies). Output on NiMH has been restored to previous Rex 2.0 Hi levels (first Rex 2.1 build had reduced max output on NiMH). Excellent performance on 14500. Light also features a 3.2-3.0V "low voltage" warning (i.e. light strobes), so you can safely use unprotected cells Very nice low mode, also available on 14500. Not as low as the NiteCore DI or Liteflux LF5, but quite respectable (especially on 14500). Well worth the upgrade - only downgrades from 2.0 are in beam "ringiness" on my sample, plus some UI issues with the 3 light states. For a full comparison to the Rex 2.0 and earlier build Rex 2.1, see my review here. Light retails new for ~$40. *JetBeam C-LE v2.0 (Q2):* Basically a reverse clicky version of the popular C-LE v1.2. Clicky is good quality (though fairly stiff) and facilitates mode changing (which could previously be unreliable in twisty version when the threads got dirty). The original C-LE v1 (and DX X.V knock-off) both had a very noticeable 73 Hz PWM. C-LE v1.2/v2.0 both use a more respectable 317 Hz. C-LE v1.2/v2.0 also have improved threads over previous versions, are a bit thicker in diameter, and features a 2 sec memory mode that retains last output setting used. Output/runtimes have not changed appreciably from v1.2. Main improvement is a good quality reverse clicky, which makes it easier to reliably switch modes. On all levels, the C-LE v1.2/v2.0 have increased output compared to earlier versions, with either similar or longer runtimes. Note that this means low mode is not as low as previous versions. Max output is still somewhat on the low side for a Cree-based light – closer to the SSC lights reviewed here. See my v2.0 review here. Note: there has been at least one review showing different output levels in a recent shipping C-LE v2.0. Not sure if this was a defective light, or a Manufacturer’s change in specs – my C-LE v2.0 comes from the first production run. Although sometimes hard to find, the light retails new for ~$30. *EDGETAC NiteCore Defender Infinity (Q5):* EDGETAC confirms that the light uses PWM, and the frequency is >1 kHz. I'm thinking it's likely considerably above, since I can’t measure it with my setup or detect any hint of it by eye. The Nitecore allows you to set your low mode (user defined mode) through a continuously variable brightness mechanism similar to Liteflux - but easier to use and more visually linear, as described in my comparison review). This is a very powerful feature, since it allows you to choose your own custom output level. The light is controlled by a forward clicky switch (i.e. momentary on, followed by a click to lock-on). User interface is quite elegant – easy to use, simple and straight-forward - I found it very intuitive. Don’t have time to explain it here, see my NiteCore DI comparison review thread. Note that strobe is neatly tucked away so you don’t need to see it if you don’t want to. No doubt about it - this is the brightest light on standard AA batteries (alkaline or NiMH) that I have ever tested. Runtimes on Hi are very respectable on these battery types. Output and runtime on 14500 is very good, with output somewhat brighter than regular batteries (although not the brightest I’ve seen – the Fenix, Olight and some JetBeam models are brighter on 14500, but shorter lasting). Regulated output on all battery types in all output modes is impressive (e.g. unlike Fenix/Olight on 14500, where you loose low modes until light drops to regulated level). Very good throw for its size, comparable to Fenix/Olight on standard batteries (less on 14500). For a lot more info, please see my NiteCore DI review here. Light retails new for ~$80. *Fenix L1D (Q5 and R100):* The Cree Q5 is slightly brighter than the Rebel R100 version, but with less runtime The Rebel has a more SSC-like beam (i.e. wider spill, smaller hotspot, smoother transition from hotspot to spill). The Cree Q5 has a more traditional Cree beam profile with visible dark rings, despite the OP reflector. To my eye, color rendition outdoors seems better on the Rebel lights compared to the Cree lights Fenix remains the output/runtime king on standard batteries - nothing else beats the efficiency of the Fenix's current-controlled low modes. I just wish they could find a way to make lower low modes. On 14500, you loose all the low modes initially, as light runs in direct drive from max output until it hits the low mode level. At that point, it then switches into regulation – but only briefly, as the battery is typically almost exhausted by then. Note that Turbo mode (i.e. head fully tightened) is difficult to do on 14500, due to extra length of the protected 14500 (I had to use a copper ring spacer for it to make contact and screw down tight). Cree Q5 version has a bit more overall output than the Rebel R100 on standard batteries, but peak throw (i.e. center of hotspot) is about the same. See my original Fenix Rebel vs Cree: OUTPUT, RUNTIMES, BEAMSHOTS thread for direct comparison of R100, Q2, and P4 version results See my Fenix L2D Q5 vs R100, R80, Q2, P4 Comparison Review: RUNTIMES+for an analysis of the premium Cree Q5 and Rebel R100 heads on L2D form. Although the R100 is hard to find, the lights retail new ~$55. *Olight T15 (Q5)* The Olight uses a current-controlled circuit for its various modes, with very similar output and overall characteristics as the Fenix. So close in fact that I suspect it is a variant of the same circuit. Output/runtimes were generally very consistent with my Fenix L1D Q5, except for Med-Lo modes on alkaline where the Olight runtimes were a bit less for the same output. Spillbeam width is narrower on the Olight compared to the other Crees, but overall output and max throw are similar to the Fenix L1D. Build quality is very high. Light has a very substantial feel, bulkier than the Fenix model, but otherwise with very similar overall build features. Press the reverse clicky tailswitch to activate the light. Soft-pressing the tailswitch will move you into strobe followed by SOS modes. Light intensity is controlled by loosen-tighten switch of the head. There are 5 distinct output modes controlled by the head switch, although the first two are identical on regular batteries in my sample. Switching sequence is from Max to Min output. Light has a memory mode retaining the last level used. As you’ll see in the runtimes, output at Level 5 (Lo) is identical to the Fenix Lo, and Level 3 is identical to the Fenix Medium. I haven’t tested the other levels yet. On 14500, you loose all the low modes initially, as light runs in direct drive from max output until it hits the low mode level, at which point it briefly enters regulation before the battery is exhausted (exactly like the Fenix L1D). For all intents and purposes, the Olight circuit seems to be exactly the same as the Fenix L1D except with the addition of an extra Lo-Med mode (i.e. Level 2) and a different way of switching between modes. Light retails new for ~$50 *Lumapower LM-31 (SSC)* Formerly known as the LM-301, this is an updated version of the original Luxeon LM-301 that now features a SSC U-bin emitter. Features a wider spillbeam than most lights (hard to see on the beamshots, since I had to crop it from another comparison – I no longer own the light). SSC emitters provide a less sharply defined hotspot than Crees, with a smoother transition (i.e. no dark rings) Max output is quite acceptable on standard batteries, though not as high as the more expensive Cree options out there Light uses a resistor for Low mode, so is not as efficient at other choices - but still acceptable runtime for the output, with a good long "moon mode." Can take 14500, but Manufacturer does not recommend it on Low mode - circuit not designed for it, produces reduced output and runtime compared to Hi mode Build quality quite good - very substantial feel and nice anodizing, but longer than most other lights reviewed here and can't tail-stand. Battery tube not wide enough to take all high capacity NiMH (i.e. none of my 2650mAh would fit, so Hi runtime is based on a 2500mAh battery) EDIT 2/29/08: Although the light doesn't tailstand with stock orange tailcap installed, I've been informed that the GID tailcap included as an accessory does allow tailstanding. I no longer own this light, so I can't confirm personally. Light retails new for ~$30 *JetBeam Jet-I PRO (R2)* Light uses PWM for its low modes, but frequency is high enough that I can't detect it by instrument (or by eye in everyday use). Excellent maximum output and throw on all battery types. This is the best throwing 1AA light I have tested to date. if you are looking for 1AA pocket rocket, this is your light. Relative for the output, runtimes on Hi are excellent on rechargeable battery types (NiMH and Li-ion), comparable to the NiteCore DI. Output is well regulated on rechargeable batteries (NiMH and Li-ion) in all output modes (e.g. unlike Fenix or earlier JetBeam lights on 14500, where you loose low modes initially) Unfortunately, efficiency on Med-Lo mode is significantly lower than the competition. Further, there’s an issue with 14500 on Med having both lower output and lower runtime than Hi (JetBeam has informed me that they plan to fix the Med-14500 issue for the second batch of lights produced). Runtimes on alkaline are also quite poor compared to the competition, especially on Lo. Similarly, there is also lower runtime on Energizer e2 lithium (L91) compared to the competition (see my detailed Jet-I PRO review for more info). Output levels of Med-Lo are higher than most lights. The user interface is vastly improved over earlier JetBeam models (i.e. simplified). Build quality on my sample is very high, best I’ve seen from JetBeam so far. Light retails for ~$75. *Zebralight H50 headlamp (Q5)* Much more than just a headlamp, the Zerbralight is a very versatile little flood light. If it uses PWM for the low modes, then the frequency is high enough that I can’t detect it with my setup, or notice it by eye. The Zebralight has 3 modes, accessed by a twist of the tailcap (i.e. twisty interface). Light comes on Low when first screwing the tailcap tightly closed. A rapid twist off-on and you move to Medium. Another twist and you are on Hi. There is no memory mode. Good selection of output levels on standard batteries, with Lo and Medium being considerably lower than most other lights. Runtimes are truly excellent for the apparent output levels. Light lacks true low modes on 14500, and max output is insanely bright (so much so that I won’t do runtimes at this level – I doubt the heatsinking could handle it). Interestingly, on Med/Lo the light has a long “moon mode” on protected 14500 (i.e. doesn’t just rapidly drop to zero output when the protection voltage is reached). I stopped the run at ~4.5 hours, and it was still producing a low level of light on my AW protected 14500. Build quality is top-notch, good machining and anodizing (type III – HA). This light has one of the thinnest layers of aluminum that I’ve seen, but still feels solid. Weight of the bare light is a negligible 18g. Can tailstand (or headstand for that matter! ) In headlamp mode, the silicone bracket holds the light fairly firmly, but has enough play so you can still rotate the light easily. This means that the light can also rotate on you inadvertently while you are moving around. Light retails for ~$50. *JetBeam Jet-I MK IBS (Q5)* The IBS versions are JetBeam's new continuously variable circuits, similar to the EDGETAC NDI. See my full review for more details. All JetBeam lights use PWM, to my knowledge. On this light, the frequency is high enough that I can’t detect it by eye or instrument, even at the lowest output settings. Highest output on 14500 of any 1AA I’ve seen, by far. Good news for all you output freaks out there! … but I must caution about the risks of excessive heat if run this way in a prolonged fashion. Runtimes on the highest and lowest output levels are very good on rechargeable battery types, but lower than the competition on standard batteries in the Med-Hi range. Lowest output mode (5%) is as low as I’ve seen on a multi-stage 1AA light – with output and runtime as good as my EDGETAC NDI. Well done! The feature set and user interface on this IBS circuit are more advanced than most lights, and is a considerable improvement over earlier JetBeam models. Although straightforward, I found the triple-flash to enter variable brightness and other modes a bit cumbersome compared to the simpler NDI interface. Light lacks a memory mode, but doesn't need one since you have the flexibility to set 3 programmable states yourself. Build quality on my sample is very high, consistent with other JetBeam lights. Standard issues like anodizing mismatch, lettering imperfections, and stiff reverse clicky switch with plastic switch retaining ring persist, but overall quality is still quite high. Light retails for ~$70. Final Word: On regular batteries, the NiteCore DI, Jet-I PRO and JetBeam Jet-I MK IBS (Q5) are currently the brightest 1AA lights out there. The NiteCore DI and JetBeam Jet-I MK IBS (Q5) both have very low Lo modes, “infinitely variable” output control system, ability to retain all levels on 14500, excellent build quality and UI, etc. These lights are very sophisticated - yet easy to use. The Jet-I PRO is clearly the best thrower of the bunch - a veritable 1AA pocket rocket! Very well built, there are unfortunately some efficiency issues to be resolved on the original circuit, as well as poor overall alkaline performance. However, an IBS circuit upgrade should be available shortly for this light. The Fenix L1D Q5/R100 lights remain the output/runtime efficiency leaders for their current-controlled low modes (although the Olight and NDI are close). Lower low modes would be nice, as well as the retention of regulated output levels on 14500. Build quality is high with a straightforward interface - it’s hard to go wrong with either model for a general purpose light. Olight T15 is basically a Fenix L1D Q5 with an extra Med-Lo mode and different switching mechanism. Build quality is fraknly even higher than the Fenix, although the light is a bit bulky in comparison. I just wish the output sequence went from min to max, instead of the other way around – but a minor quibble since light comes with a memory mode. If you like a light with the lowest Lo mode and long runtime, the LiteFlux LF5 is your best pick (~0.3 on my output scale, with two weeks worth of continuous runtime at this level on alkaline). The LF5 is a powerful light with a lot of features, but somewhat complex to program with its twisty interface. Overall output is less than the Cree lights, and runtime performance on alkaline is unimpressive. However, it does retain all levels on 14500 (including nice low modes). The Rex 2.1-002 is a great jack of all trades – very good Hi, Med and Lo levels on all batteries, retains output levels on 14500, excellent runtime performance overall. User interface is a bit complex with its 3 sets of output states – easy to accidentally move between them if you like to use the light on low (switching is controlled by a time delay on the lowest setting). But a very strong performer for a Q2 light, and my previous EDC before the NiteCore DI came out. JetBeam C-LE v2.0 and Lumapower LM-31 are both excellent choices in the low-mid price range of quality lights (~$30). Both lack many of the features of the higher end lights (e.g. 14500 not supported on C-LE, and not recommended on LM-31), but both are good solid choices for general use. I would recommend the C-LE v2.0 for those wanting runtime efficiency, extra modes, or small size, and the LM-31 for those wanting a simple interface with good low mode (e.g. reading at night). Budget DX/Kai models have been omitted from this Part III round-up since I’ve found their quality quite inconsistent. You never know what you are going to get from one batch to the next - output and build quality is incredibly variable over time. The Zebralight H50 Q5 is a very unique and versatile little light – more than just a headlamp, this is a true multi-purpose flood lamp with outstanding runtimes. Whew - there you have it! I will continue to update this thread as new 1AA lights come out.