Fenix PD32UE - some crude, but useful, battery performance data

Tim T

Newly Enlightened
Joined
Jan 27, 2013
Messages
8
Hello everyone:

I've been interested in flashlights for quite some time. However, I have only recently been rather badly bitten by an actual flashlight bug.

If I were to blame one light for causing my current troubles, it would probably have to be the Fenix PD32UE . . . it has a combination of features that I found very intriguing (and had wanted to have for quite some time). I knew that my current stock of 18650's would probably not be up to the demands of this light. They are all getting at least a couple years old, most of them are - admittedly - of the bargain variety, and they spent most of their standby time in pretty much a full charge state.

I initially acquired a couple of 3100 mAhr Panasonics and they certainly do work OK. But, due to the bite from the bug, I have also been looking at some other lights and consequently also wound up looking into lithium ion batteries a little more carefully. Thanks to the valuable information that can be found on this forum, I discovered - surprisingly, to me - that the Sanyo "UR18650FM" might actually provide better performance with a light like the PD32UE, even with its capacity rating of a "mere" 2600 mAhr. It maintains a flatter or more stable voltage curve during discharge, meaning that it would probably be able to better maintain close to full output for a greater portion of a cell's discharge cycle. Having learned this, I ordered a couple of these cells and just recently received them.

Although I do not have a lot of instrumentation available, I do have a simple light meter and it occurred to me last night that I could get an actual measure of relative performance for these two 18650 options. The observations I am reporting here are based on only one run with a Panasonic NCR18650A (3100 mAhr) and a Sanyo UR18650FM (2600 mAhr), so I wouldn't take the actual numbers as too concrete. But even with the data being pretty limited, they do make enough of a point to make the observation worth passing on.

I do have a "hobby charger" and the cells used here have been very recently analyzed from rested full charge to a 3.0 volt cutoff at 1.8 and 2.8 amp rates, both right after receipt and after a few shallow to modest cycles of general use and fooling around. mAhr values were consistent with what one would expect based on those reported in HKJ's battery tests, so I'm pretty confident that the data here, though limited, are not skewed by the unknowing use of a cell with atypically bad or good performance.

I should note, however, that the measurements made with the PD32UE were collected with the cells starting at 4.04 volts. With my acquisition of new (and good quality!) cells, I had decided that they would generally be kept at a bit of an undercharge when in standby. Since the idea of this experiment came to me rather suddenly, I just used them the way they were.

I set the light up to measure relative output by "ceiling bounce". This was done in a spare bedroom that is very cool (North Dakota winter here), so heating was not found to be significant over the course of the run. Since the maximum ("burst") output of the PD32UE cannot be maintained continuously, I decided that I would run a series of 10 minute discharge sequences - each including 30 seconds of burst (reported by Fenix to be 740 lumens), followed by 9.5 minute on turbo (400 lumens). In each case, burst output was recorded at the end of the 30 second burst operation. Although this selected discharge pattern was essentially arbitrary, it is at least somewhat representative of how I might actually use the light in practice.

Like I said, the overall data here are limited, but sufficient to make the important point:

After an hour (that is, six total burst/turbo sequences), the Panasonic NCR18650A was providing a burst output that was about 84% of its initial value (this would actually be as measured at the end of the seventh 30 second run in "burst" output). When I get a chance to collect some more data, I will look into putting things in a graphical form, but I'll just note for the moment that the decline in burst output was pretty much continual, with each burst run being a little dimmer than the previous one. Output on the "turbo" setting stayed constant over the course of the runs with both the Sanyo and Panasonic cells.

With the Sanyo UR18650FM, initial burst output was about 5% higher than with the Panasonic, though this could conceivably be due to reproducibility problems with my setup. Either way, after the same 60 minutes and six full burst/turbo runs, the Sanyo still provided a burst output that was 95% of its initial value. I extended the Sanyo run out by another three cycles (out to 90 minutes, in other words). At that point, burst output was getting down to around 88% of initial. Also, it was now fading by several percent between the beginning and end of the tenth 30-second burst cycle, while in previous cycles it had held pretty constant over the course of each 30 second stint. In other words, the Sanyo was probably in the process of heading over the steeper voltage cliff that it shows at the end of a discharge.

Even though this quick experiment was pretty slapdash, it quite clearly shows that the lower capacity, but flatter voltage Sanyo cell will provide markedly better performance in an application like this. However, this assumes that "performance" is defined as maintaining maximum burst-setting lumen output for as much of the battery cycle as possible. In some situations - anything from a camping trip to an emergency - longer total runtime may be far more important than stable output. In such cases, the higher capacity option may be the better choice. This is especially true since the dropping voltage of a cell like the Panasonic will actually cause the light to use less power (when on the higher settings, anyways) as the discharge continues. All of this assumes, as is the case with the PD32UE, that one is talking about a light that runs on one 18650 and that does not use any "boost" type circuitry.

I found this little experiment to be rather fun as well as informative. Chronic flashaholics would probably warn that this is a bad sign. Either way, I will make some more measurements in a little more of a legitimately "scientific" fashion (and then figure out how to insert a graphical representation of the data into a post). In addition to repeating the same basic measurements, I should also try it starting with fully charged cells, of course. Also (duh!), I should collect a true reference measurement for FULL burst output by measuring it with the light powered by a pair of fresh 123A's. However, it might be a few days before I get to any further efforts. As such, I thought it appropriate to report these initial results because I know that there are at least a few folks who are wondering about appropriate battery choices for the PD32UE and other similar lights.

Thanks much to the folks on CPF who have posted the amazing amounts of useful data and information, both on the operating details of various classes of lights and on the output details of different battery options!

Tim
 

CyberCT

Enlightened
Joined
May 10, 2010
Messages
633
That's quite a long post. Maybe putting a table in there to compare the two instead of a paragraph comparing the two would help a little.
My question is are the cells you are using protected? If so, what brands?

From my protected 18650 testing (AW 2900mah, Keeppower 3400mah), the 3400mah batteries don't last quite as long in turbo mode (very close) but surpass the 2900mah and 2600mah Fenix uses in runtimes for their TK75 in the 2 lower modes. I'm sure the lowest mode it would be an even larger % longer lasting. I did some posts with my findings in the TK75 thread.
 

Tim T

Newly Enlightened
Joined
Jan 27, 2013
Messages
8
I didn't realize how much I yammered when I was doing it! The egghead definitely kicked in on that!

The Panasonic was a 3100 unprotected, while the Sanyo was a protected cell from Fasttech (I'm pretty confident that they are authentic, only warning on them is that they are LONG, also FAT . . . though this part posed no issues in my particular case).

I actually have some unprotected 2600s on the way, which I plan to use in suitable one-cell flashlights (I had originally decided to get protected ones to make them more universally useful - could be used in 2-cell lights also without worries). But, upon finding their length, decided to get some unprotected ones. Of course, this should also make performance a little better by avoiding any voltage drop or loss in the protection circuit.

I figured I would wait until I got some unprotected 2600s for a proper apples-to-apples comparison. Also, to make it worthwhile for putting something in graphical form.

I would suspect that the total runtime in the lower modes may prove to be better with the higher-capacity Panasonic cells. Since lower forward voltages would be needed by the lower settings, the Panasonic should still be above the minimum needed out to a longer total mAhr discharge. The other interesting variable, for comparing runtime data, would be how much actual LED-to-LED variation there is in voltage needed for operation at the various output levels. This could make a very big copy-to-copy difference even when comparing a specific model of light. I've never really looked into the "components" area of the forum to see if anyone knows, or has looked into, how much actual variation there is from light to light. Cree lists some nominal specs, but does Fenix or anyone else do any further testing? I'm totally ignorant on the question, but will snoop around on it. An LED with a rather high Vf would not be as huge of an issue with 2x123A or on a bigger light with 2x18650, but could make quite a difference in how long one could maintain burst output when running something like the PD32UE from one 18650.

I guess the understanding about the different batteries (at least for my needs) clicked for me most when I found HKJ's comparator site: http://lygte-info.dk/review/batteries2012/Common18650comparator.php

If you put in the Panasonic 3400 (he lists both protected and not, and there definitely is a difference) and compare it to his "Sanyo 2600 red" test, there is a quite pronounced difference in voltage maintenance. The unprotected Panasonic holds, as an example, 3.4 volts out to about 1.75 Ahr (at a 3 amp discharge) while the Sanyo (also unprotected) holds this voltage out to ~2.15-2.2 Ahr at the same current. Similar trends are seen at 1 and 2 amps, but if you go down to 0.2 or 0.5 amps, the Panasonic beats the Sanyo, rather badly, at voltages below about 3.5 volts. If a person had a "typical" XM-L, needed Vf would be 2.9 V at 700 mA and 3.1 V at 1500 mA (per the Cree data sheet). Even with some overhead for the driver circuitry, lower brightness settings should be maintained for most of the cycle, even with the more steeply drooping cells. With my particular copy of the PD32UE, at least, "burst" output starts tanking pretty quick with the 3100 mAhr Panasonic.

My main concern - which may be quite different for other folks - is a cell that would allow me to use the light for a fair bit of routine stuff (mostly lower setting use) over the course of the day while still being able to maintain at least a few minutes of respectably decent "burst" if it was needed. People who define performance otherwise (like low setting runtimes) may very well be better served with the super-capacity panasonics.

There is one possible source of confusion in HKJ's comparator to mention (the confusion is due to the supplier labeling): There are two "Sanyo 2600" cells listed in HKJ's testing menu. The one that ends with "bv" shows a rather dramatically lower capacity. It is protected, while the other one is not, but this still seemed a bit severe of a difference. But the details on the one that ends with "bv" (see the actual review) show that the actual cell used here is the UR18650ZT, which is a 2800 mAhr cell, but which is spec'd for a maximum charge voltage of 4.3 V. I suppose it is actually a bit more honest of the supplier to represent this as a "2600 mAhr" cell, if they are selling it to people who are likely using 4.2 volt chargers, but it certainly does cause some confusion if one expects a "2600 mAhr Sanyo cell" to mean the UR18650FM.

Thanks for the reply and comments . . . I will follow up with a bit more "analysis" once I get a chance. MORE graphical and LESS blather-some.
 
Top