Zebralight SC52 (XM-L, 1xAA, 1x14500) Review: RUNTIMES, BEAMSHOTS, VIDEO and more!

[MichaelW]

Does this light get hot on H2B? I am wondering if the body gets hot enough to transfer waste heat to the cell, say if you used alkaline in a cold environment, to maintain flat regulation. Maybe if you intentionally insulated the light.
But damn good regulation on M1. Kind of funny that the improved efficiency is making alkalines viable, well to the general populous not CPFers, of course.

 

[Bumble]

looks like this review has finally won me over, i will purchase one soon

 

[pepekraft]

I think I need one of these, dangit. I only have headlamp style Zebras so far, and this looks like the time to cross over to the flashlight form factor.

I'm confused by the 14500 graphs. If I'm reading things correctly, the only reason to put a 14500 in one of these is for the first 30 seconds. After that, in every case, the eneloop outperforms Li-ion.. is that correct -- same output and longer runtimes on the eneloop? Not that I'm bashing that first 30 seconds. That's a lot of light.

 

[selfbuilt]

Nice Review, I like the runtime tests. I'm a little sad to see that this light is not that well suited to the Lithium cells when used on Hi. The SC51 is much better for EDC with Lithium primary in my opinion even if it gives up some brightness.

I've just updated the L91 runtime graph (1xAA) in the review with the H2B results:

The L91 runtime on Hi2B is quite impressive - nearly twice the runtime as Eneloop at this output level, and about 3.5 times the runtime of standard alkaline. oo: I'm surprised that it can outperform the SC51 by such a wide margin. It also demolishes the Nitecore MT1A at this level.

I agree you are probably best sticking with NiMH if you plan to run on H1 ... but the L91s seem like an excellent option for H2A on down.

Does this light get hot on H2B? I am wondering if the body gets hot enough to transfer waste heat to the cell, say if you used alkaline in a cold environment, to maintain flat regulation. Maybe if you intentionally insulated the light.
But damn good regulation on M1. Kind of funny that the improved efficiency is making alkalines viable, well to the general populous not CPFers, of course.

All my tests are done under a cooling fan, and it's definitely possible that alkaline may not have fared as well on H2B without it. Of course, most people don't run lights at this level for an hour continuously, so I think you could expect flat regulation at this level (on down) on alkaline in normal usage.

I agree, it is rather ironic to see a modern light perform so well on the rather archaic alkaline cell. I would certainly never want to store one in there, given their propensity for leaking.

I'm confused by the 14500 graphs. If I'm reading things correctly, the only reason to put a 14500 in one of these is for the first 30 seconds. After that, in every case, the eneloop outperforms Li-ion.. is that correct -- same output and longer runtimes on the eneloop? Not that I'm bashing that first 30 seconds. That's a lot of light.

You are reading everything correctly - and the result is actually not that surprising.

The two typical advantages of Li-ions over standard cells (in multi-power lights) is the ability to drive the emitter harder (i.e., the higher-voltage Li-ion can handle the higher current drain), and the improved ability to maintain flat regulation (i.e., Li-ions can more easily supply power fully regulated). In this case, the SC52 does a remarkable job keeping flat regulation on most cells (at all levels below Hi1), so that advantage is negated somewhat. The SC52 does run brighter on max Hi1 on 14500 - but due to the small heatsinking mass, the circuit won't keep it up for long (i.e., timed step-down after 1 min). So practically, you don't get to benefit much from these two typical 14500 features.

The other question is battery capacity and relative circuit efficiency for different voltage sources. In the capacity sense, the actual work capacity of 14500 is not all that different from NiMH. In simple terms, you can think of this as the work done for the current to flow through a given voltage differential (i.e., the definition of a watt). If you multiply typical amperage capacity (in mAh) by nominal voltage (in V), you basically get work capacity (or Watt-hours, Wh) for a cell. The typical 14500 thus only has a small advantage over the typical NiMH for actual work capacity (how much exactly depends on how much you trust the rated mAh capacities).

In an ideal situation, I would expect my 14500s to only have a small runtime advantage over Eneloop NiMH for a given output level. That said, there are also always potential differences in the relative circuit efficiencies of different voltage sources, which may skew things a bit one way or the other (i.e., depending which voltage source the light is optimized for). In the case of the SC52, standard batteries seem to have a greater edge - which gets back to why I consider the NiMH/Alkaline/L91 peformance to be outstanding on the SC52, and the 14500 performance only very good. :wave:

 

[QSL]

Re

I'm confused about the data in the high eneloop chart.

If I'm reading it correctly I'm better off using the brighter of the two sub high modes, it will give me a higher output, with a flatter regulation, and a longer runtime?!

Look like the lower level is unregulated and drops quickly. I must be reading the chart wrong.

Thanks for the great review.
73's Bill

Sent from my Galaxy Nexus using Tapatalk 2

 

[QSL]

Duh,

Nevermind , just ignore the new guy. Apparently he's (me) color blind and can't read a simple chart. Lol. Nothing to see here move along.

Bill

Sent from my Galaxy Nexus using Tapatalk 2

 

[chadvone]

Great Review SB .

Automatic stepping down when battery (AA and 14500) capacity is low

Your run time graphs show any indicator to having this feature ?

 

[reppans]

This is a long post …. my apologies in advance.

....There's been a lot of people waiting for this review - I am glad the lumen arguments can finally be laid to rest....

....Well, I wouldn't say that exactly. Despite how it may sound above, I'm not trying to shut down any discussion of relative or absolute lumen level estimates. I think it's important to critically evaluate ALL claims of purported "lumen" values, to see if they hold up to scrutiny....

I agree w/ Selfbuilt and I'm quite certain that I am the one that started this whole liberal lumen scale thing with ZL and Selfbuilt. There has been only one other CPF member that has publicly back me, and while I don't always agree with his political correctness in voicing opinions, I sometimes really wonder if it is only the three of us (Selfbuilt included) that truly understand the issues here. Selfbuilt is absolutely a stand-up guy, and sounds to be completely open to this discussion, and without taking any personal offense, for which I truly appreciate. I think we both understand each other. I am sorry that so many others do not get the point…. and/or of course, simply do not care.

Reppans,

....it appears like you have identified some lights whose manufacturers may have understated the lumen output.....

.... These types of mistakes are made, both over and under ratings, and the responsibility for these errors rest squarely upon the manufacturer. The onus of correcting their ratings does not fall to Zebra or Selfbuilt.....

......We use ANSI measurements here. There are no Zebra lumens, Fenix Lumens, or collaborations between Selfbuilt and Zebralight. This past month seems to have brought out some unfortunate comments from some decidedly adversarial posters, calling both Selfbuilts and Zebralights integrity into question.....

First off, there are no official "ANSI " lumens here... please read in detail both Selfbuilt's comments here, and his methodology discussion on his website. He is making an estimate based on a sample of manufacturers (best fit line), and if that sample is skewed, so is his scale. I am doing the same and I know my scale is skewed on the conservative side.

I will admit to being the loudest voice on ZLs spec sheet overstatements and I do question the integrity of the company, partially due to the lumen exaggerations, but also for pulling things like tripling the 0.34 moonlight runtimes (vs a QAAX) by actually delivering only 1/3 of the spec'd lumens. Similar with the 2.7 L1 mode which also matches (sorry, "is supposed to match") the QAAX, and I have highlighted this in my earliest reviews of the SC52. These modes happen to be personal favorites of mine, but they are also important indicators of a flashlight's driver efficiency. Is this an honest mistake, grossly poor estimation, or deliberate misrepresentation? Imagine if Ferarri claimed one of its vehicles got 50 MPG.

I don't have any issues with Selfbuilt's integrity and in fact am even more impressed by his willingness to discuss this issue openly.

While I have more than voiced my opinion about Selfbuilt's liberal lumen calibration, I have also repeatedly said that his data is consistent and relative, and I frequently use his numbers as the basis from which to recalibrate the conservative manufacturers "up" for apples-to-apples comparisons in my posts. The big problem I have with liberal interpretations, especially from someone as respected as Selfbuilt, is that people automatically take it as a de facto ANSI certification. I personally believe, in the longer run, it will reward manufacturer bad behavior and eventually lead to less transparency for all us flashaholics. Having said all that, and I've repeatedly said this too, if I were in Selfbuilt's shoes, I would do the same exact thing. Selfbuit has modestly stated here…..

...Thanks. Yes, at the end of the day, I'll I can say is my relative scale is consistent over time. Even if it does seem slightly inflated, not sure there's much point in trying to come up with a new calibration - it would likely just create confusion between the "new" numbers and the "old". The consistency and backward-compatibility is more valuable in my mind than the absolute correlation.

…. that it doesn't make sense to recalibrate for practical reasons. However, I personally feel there is NO WAY HE CAN RECALIBATE, if only for POLITICAL reasons. Imagine if Selfbuilt were to adopt a more conservative lumen scale and then start reporting that ZL, and a few other liberal manufacturers, are not meeting their specifications. I suspect that Selfbuilt's reviews will diminish in quantity as his supply of flashlights dry up. A man in his position, just like the automobile journalists, must walk a politically fine line with the manufacturers. The only resolution to this, in my mind, is to free Selfbuilt up, by turning him into a "Consumer Reports" of sorts, and have CPF members fund his flashlight purchases. If there's any interest in this, I will happily offer to fund 10% of the pot, when it hits USD 2,000. Selfbuilt, I hope you do not take any offense to my suppositions here, once again, if I were you, I wouldn't do a single thing differently.

Now back the original question of whether ZL is overstating, and Selfbuilt's scale is too liberal. I attach a couple of Selfbuilt's recent Output/Runtime graphs for the 3V, 1xCR123 and 2xAA classes. I took the liberty of including manufacture max lumen specs as best as I could determine, and attempted to superimpose the SC52/L91 graph onto CR123 graph… I hope I've been reasonable fair and accurate with the data, please correct me where I am wrong. Here are the saiient points based on Selfbuilt's review population above:

1) 280 lumens appear to be the highest manufacturer specification (using a non Li-ion battery), and most are significantly lower than that, in the 3V class. The SC52 is spec'd right at that high water mark and with a 1.5V.
2) Despite claiming THE highest lumen spec, the SC52 appears to be the bottom dweller on both charts, and probably, the furthest away from it's claimed spec. (assuming the others are remotely correct)
3) Looking at Selfbuilt's CR123 and 2xAA max lumen tables above, it appears that every light, save ZL and Jetbeam, have been "underrated" by their manufacturers.
4) So, perhaps it is not so much that I've found the few manufacturers that understate their lights, rather, it appears that the current bar is set low enough to accommodate the lowest common denominators.

When we are talking about really statistics and the use of "best fit lines" I strongly believe in bell curves, and for every light that has been understated by its manufacturer, there should be an equal number of lights that have been overstated by its manufacturer. So, where are all these overstated lights?

Selfbuilt, apologies for referring to you in the third person, I feel I am still trying to explain my position to others. In the end, I suspect you may be one of the only people that understands, or cares, what I am talking about. Please do not take any of this personally, I hope it is merely construed as constructive criticism..... you're the best.:twothumbs

 

[tam17]

Great review, Selfbuilt! This petite ZL is bound to be a real killer...

Cheers

 

[kosPap]

putting it up against the 2AA lights. what a bold thing to do!

Since the light is 14500 compatible, it would be interesting to see how it does with the 14505 3V batteries

then you can make a graph with the various battery chemistries on high mode.

 

[WilsonCQB1911]

I don't know if anyone has noticed, since we've been so entrenched in the weeds on lumen issues... but the SC52 is actually a really nice flashlight. I'm really/happy excited with mine.

 

[reppans]

putting it up against the 2AA lights. what a bold thing to do!

Since the light is 14500 compatible, it would be interesting to see how it does with the 14505 3V batteries

then you can make a graph with the various battery chemistries on high mode.

I personally asked Selfbuilt to include onin the 2xAA graphs (which he did), and also the CR123 graphs (which he did not), since this light attempts to compete with the top of the class in those batteries categories.... ummm, on paper at least.

Good point about the 14505s, this was discussed in one of the SC52 threads. They don't work properly in this light due to the idiot-proof automatic low voltage step-down protection, and probably the same with IFRs too.

 

[WilsonCQB1911]

I personally asked Selfbuilt to include onin the 2xAA graphs (which he did), and also the CR123 graphs (which he did not), since this light attempts to compete with the top of the class in those batteries categories.... ummm, on paper at least.

Good point about the 14505s, this was discussed in one of the SC52 threads. They don't work properly in this light due to the idiot-proof automatic low voltage step-down protection, and probably the same with IFRs too.

Let it rest man. It almost seems like you are getting paid by one of ZL's competitors.

 

[stp]

reppans there are more people who understand the issue, for example me. I will just point few things, keep in mind that it's just my point of view and I don't have time/will and this is no place to argue about it anyway.

-Introducing few lumen scales named after manufacturer was counter-productive. Just think what a mess it would create if other people would catch to it and it would spread over the forums. There is only one scale: THE ANSI. Others (manufacturers, reviewers) should try to stick to it as close as possible.

-..as close as POSSIBLE. And here is the problem because it's not easy. You would need few sources of light calibrated and certified by ANSI to calibrate your measures. And it still wouldn't be enough. You would have to control many aspects: temperature, power source, the way you measure every light and so on.

-It's still much better now than it was before. We had OTF lumens, led lumens, calculated lumens, lumens after 10s, lumens after 30s, lumens from thin air...introducing ANSI FL-1 by manufactures is a great progress. The FL-1's runtime until 10% is stupid but still it is big progress.

-There is difference between the same bin leds, batteries, drivers. Manufacturers introduce silent updates. Is there a high chance that the lumens measured by ZL or selfbuilt are higher than in some ANSI laboratory or other manufacturer? Of course. But I and you can't be really 100% sure. And if yes than by how much? You can't use the bell curve because the sample pool is too small and there are too many factors.

-Wouldn't it be great if selfbuilt could say that his measures stay within for example 1% of ANSI? It would of course. And I believe that he could do it but the problem is with money. We should be happy with what he is doing now because without him most of us would be unable to compare output of the lights. We don't know the exact lumens but we know that light A emits 120% more then light B thanks to him.

-For me lumens are like Mpg(km/liter) in cars. To many factors to trust manufacturer even with heavy standardized test and hard to compare with real life usage. It's much better if somebody compares cars doing the same route. They get different Mpg than manufacturer but can compare the two cars much better and get real life data. This is what selfbuilt is doing with lights.

-You and especially Shelm made big deal about it. FOR ME from the beginning it did look like you have some beef against ZL and doubt the integrity of selfbuilt. You did way better later on. SC51 was the most efficient AA light on the market. So it looks strange when somebody tries to downplay its successor and potentially the new king of efficiency. Now we know that SC52 is the most efficient AA light on the market whatever the lumens are thanks to selfbuilt.

-It's much better to do it in positive way: You could ask manufacturers about certificate from certificated by ANSI labs in Europe or US and make the big deal about the ones that do it. I think that nobody is doing it currently but some could see the PR value in that if enough people will start asking them.



selfbuilt:

Thanks a lot for your work.

I have one question. Did you think about adding the area under the graph up to 50% and 10% of output as two additional parameters describing the light and its mode? It would help a lot in comparing efficiency of the lights and modes. We would know precisely which mode gives us the most bang for the buck for example. I think it should be not hard to do from spreadsheet.

 

[MichaelW]

So what is Zebra's timeframe for the xm-l2 ?

 

[Polki]

Thank you for the very nice Review

I hope the SC52 will delivered soon

Polki

 

[selfbuilt]

Automatic stepping down when battery (AA and 14500) capacity is low

There is some clear evidence of it in some of the 14500 runtimes - check out the H1 and M1, where there is at least one clear step before the protection circuit kicked in. In the other 14500 cases, I suspect the battery protection circuit kicked in before the step could occur.

On standard batteries, I don't see much evidence of (certainly not on eneloops, but perhpas I didn't let them run long enough). On some of the alkaline traces, you can see what looks like defined drops near the end of the runs (e.g. H2B alka), but it seems like the circuit is unable to draw regulated low level power from the nearly depleted cells.

I agree w/ Selfbuilt and I'm quite certain that I am the one that started this whole liberal lumen scale thing with ZL and Selfbuilt. ... Selfbuilt is absolutely a stand-up guy, and sounds to be completely open to this discussion, and without taking any personal offense, for which I truly appreciate. I think we both understand each other. ... Selfbuilt, apologies for referring to you in the third person, I feel I am still trying to explain my position to others.

To be honest, I only briefly scanned through the other threads, so I may have missed some of the points raised. But I certainly don't feel any sort of offence to any of the comments I've seen - and I think it is important to raise and discuss these issues. Not sure of the best place for it, but I'm fine with leaving it in this thread (along as everyone stays respectful of everyone else).

Like most here, I have my own personal views of which manufacturer lumen values seem to be the most consistently believable. I try to keep that speculation out of my reviews, and just comment on how each given light perfroms relative to my scale, which is all I can say. :shrug: But one of the huge confounds here is the nature of sampling variability, as reppans alluded to - there are a lot of variables that can affect a given light's output (including specific output bin, where in the output bin range it falls, how the circuit responds to different Vfs - and Vf variability relative to output bin, variation in circuit components, centering and focusing of reflectors, etc, etc.). This is why my output/runtimes are just a guide - one sample, at one point in time - and may not be representational of the "average" (which, in turn, is a moving average that changes over time from batch to batch).

Something else to keep in mind is that some manufacturers may choose to stick with minimum lumen specs (at time of launch), whereas others go for typical output (and some perhpas higher still ). Over time, as output bin and other variables change, manufacturers may not update those specs, so later lights could over-perform. I'm thinking of Foursevens and Surefire as potential examples of this - they seem to go for a minimum spec at time of launch, and don't typically change them (even when improved outbin bins are used). Case in point: the original Foursevens Quark Turbo AA-2 (XP-G R5), the re-labelled Quark QB2A (XP-G, no output bin given), and the new QB2A XP-G2 (no output bin given) all share the same official "205 lumen" spec. But from my testing, my QB2A XP-G2 appear to one output bin higher than my late-model QB2A, which in turn appears to be at least one output bin higher than my old-stock Quark Turbo AA-2. This is as you would expect, as currently available XP-G2 output bins are about two steps up from the old common XP-G R5 (see my QB2A XP-G2 review for more of a discussion).

Incidentally, Foursevens was one of the manufacturers I used (along with Fenix and Novatac) for the comparison data for my calibration. But both Foursevens and Fenix were using peak "out-the-front" lumens in those days, which I found correlated very well with my intial activation data. I suspect they were closer to "average" lumen values for their lights too ... in Foursevens case, it appears to me that they have adopted a more conservative "minimum" spec rating around the time they moved to official ANSI FL-1 lumens. But again, that's just speculation - I don't have any direct knowledge of their lumen measuring processes (or any other manufacturer, for that matter).

The issue of the low levels on Zebralight is worth considering. My calibration for the lower <20 lumens levels is heavily based on the Novactac and Foursevens (sidebar: the older Novatac lights were well regarded in terms of their lumen values, it is only the later models when confidence here seems to break down). I generally continue to find that Foursevens "moonlight" mode values fit within a reasonable range of my calibration estimates, even on newer lights. But by that calibration, my SC52 is producing unusually low readings (compared to ZL specs - see the lumen table in my review). :thinking:

In any case, runtime and lumen estimates for really low levels are tricky. There seems to be a LOT more variability here than higher levels (for both outut and runtime, from sample to sample). This is especially true for current-controlled lights. When I've queried this with some makers, a few have told me that they have difficult ensuring consistent lumen levels (especially <1 lumen) between samples/batches. Others have also speculated that Vf can greatly influence runtime at these low levels (apparently for circuit funcitioning reasons, but that's getting beyond my level of knowledge or expertise).

Since the light is 14500 compatible, it would be interesting to see how it does with the 14505 3V batteries

Good point about the 14505s, this was discussed in one of the SC52 threads. They don't work properly in this light due to the idiot-proof automatic low voltage step-down protection, and probably the same with IFRs too.

Yes, I can confirm that. I just tried a primary 14505 (which is a little over 3.2V fully charged), the light came on in H1, then immediately stepped down to a medium-level mode. Same thing if I try to switch back to max - it goes there for a second, then immediately steps down.

The circuit is thus interpreting the primary 14505 as a nearly depleted Li-ion, and refusing to stay in Hi modes (i.e., triggers automatic step-down). Actually makes perfect sense, when you think about it ... that's what you get for the Li-ion low-voltage step-down feature.

I don't know if anyone has noticed, since we've been so entrenched in the weeds on lumen issues... but the SC52 is actually a really nice flashlight. I'm really/happy excited with mine.

 

[selfbuilt]

I have one question. Did you think about adding the area under the graph up to 50% and 10% of output as two additional parameters describing the light and its mode? It would help a lot in comparing efficiency of the lights and modes. We would know precisely which mode gives us the most bang for the buck for example. I think it should be not hard to do from spreadsheet.

I've thought about it, but I prefer that people look at the graphs themselves rather than just a couple of arbitrary time-under-the-curve points. There's more information to be gleamed about the regulation patterns that way.

The other problem is that I rarely let lights run down to 10%. For high-output runs, they usually hit that point fairly quickly once they fall out of regulation. But on the Med and Lo levels, it can sometimes take quite a bit of time to make it all the way down to 10% (i.e., they spend a lot of time just slightly above it). Parathentically, this is why I'm not too enamored with the ANSI FL-1 standard runtime to 10% ... it preferentially benefits lights run at lower levels on alkaline cells (which have a quick drop-off at the high-end, and a slow drop-off at the low end).

In practical terms, I also don't like tying up the lightbox for extended runs to 10%, when I have a pretty good answer by ~20% (and there's always plenty more lights waiting to be tested). It's also not good for NiMH and Li-ions to spend too long at nearly depleted levels. Again, that's not a issue for Hi mode runs, but can start to be one at the lower outputs (i.e. where 10% of initial low output means an almost dead battery, for a long period of time).

 

[stp]

I've thought about it, but I prefer that people look at the graphs themselves rather than just a couple of arbitrary time-under-the-curve points. There's more information to be gleamed about the regulation patterns that way.

The other problem is that I rarely let lights run down to 10%. ...

You could always place it in the legend of the graphs . There is more information in the graphs but some of it is harder accessible. For example it would be easier to compare (and conclude) the efficiency of Hi1, Hi2a, Hi2b, on ni-mh's and alkaline if the graphs would be accompanied by that numbers.

The 10% was just example. Even just 50% would be great.

Anyway just food for thought. Thanks again for the reviews.

 

[GordoJones88]

I think it is important to raise and discuss these issues.

Ditto. I've learned a lot of good stuff about lights from some of these discussions. People should always question things.