Chinese  English

Reviews Appreciation

Thread: Fenix TK35UE (MT-G2, 2x18650) Review: BEAMSHOTS, RUNTIME, VIDEO+(by Selfbuilt, CPF)

Publish Time:2014-04-08

Warning: even more pic heavy than usual. 

This is the first Fenix light to use the popular Cree MT-G2 emitter. Interestingly, Fenix has chosen to update an existing build – the relatively compact 2x18650 TK35

Although this light was officialy announced a little over a week ago, Fenix is expecting to actually launch the model sometime around mid-month.

Let's see how it compares to the original TK35, and to other MT-G2-based lights that I have tested. 

Manufacturer Reported Specifications: 
(note: as always, these are simply what the manufacturer provides – scroll down to see my actual testing results).

  • LED: Cree MT-G2 emitter
  • Uses two 18650 rechargeable batteries (Li-ion) or four 3V CR123A batteries (Lithium)
  • Low: 25 lm/150hr; Mid: 250 lm/13hr 30min; High: 750 lm/4 hr; Turbo: 1800 lm/1hr 30min; Strobe: 1800 lm; SOS: 250 lm
  • Max beam distance: 242 meters
  • Peak beam intensity: 14,640cd
  • Digitally regulated output - maintains constant brightness
  • Reverse polarity protection to protect from improper battery installation
  • Low-voltage warning function to remind users of low power
  • Dual button switch in the tail, convenient operation
  • Made of durable aircraft-grade aluminum
  • Premium Type III hard-anodized anti-abrasive finish
  • Toughened ultra-clear glass lens with anti-reflective coating
  • Intelligent memory circuit
  • Over-heat protection
  • Low-voltage warning function
  • Impact resistant: 1m.
  • Waterproof to IPX-8 standard, underwater 2m.
  • Dimensions: 165mm (Length) × 44mm (Diameter) × 52mm (Head)
  • Weight: 265-gram weight (excluding batteries)
  • Include accessories: Lanyard, holster, o-ring
  • MSRP: $129.95 (confirmed by Fenix)

Packaging is unknown, as I was just sent the bare light for review. I would expect the usual Fenix cardboard box (with specs and details print the outside), with the standard accessories described above (i.e., extra o-rings, wrist lanyard and belt holster) and manual and warranty card.

From left to right: AW Protected 18650 2200mAh; Fenix TK35UE, TK35; Eagletac SX25L3; Sunwayman T45C; Thrunite TN32.

All dimensions directly measured, and given with no batteries installed (unless indicated):

Fenix TK35UE: Weight 270g, Length 164mm, Width (bezel): 48.6mm
Fenix TK35: Weight 256.1g, Length 162mm, Width (bezel): 48.6mm
Fenix TK51: Weight: 476.9g (627g with 3x18650), Length: 187mm, Width (bezel narrowest): 58.2mm, (bezel widest): 70.8mm
Crelant 7G10: Weight 643.4g (827g with 4x18650), Length: 198mm, Width (bezel): 79.0mm
Eagletac SX25L3: Weight: 315.9g, Length: 150.2mm, Weight (bezel): 47.0mm
Niwalker BK-FA02: Weight: 687.6g (870g with 4x18650), Length: 209mm, Width (bezel): 80.0mm, Width (tailcap): 50.3mm
MiniMax Nova MM15 Prototype #2: Weight (with handle): 332.7g (516g with 4x18650), Length: 115.3mm, Weight (bezel): 60.7mm
Rofis TR51: Weight: 242.2g, Length: 146.7mm, Width (bezel) 45.9mm, Width (widest part): 48.1mm
Sunwayman T45C: Weight: 216.6g, Length: 136.0mm, Width (bezel) 40.4mm, Width (widest part): 40.8mm
Thrunite TN35 (MT-G2): Weight: 571.4g (723g with 3x18650), Length: 201mm, Width (bezel): 78.9mm

The overall body design of the TK35UE has not changed greatly from the earlier versions. As before, the light remains very streamlined, and quite comfortable to hold and handle. 

I have one of the original TK35s, and can see some small changes over time (although I'm not sure when these changes were introduced). As before, anodizing remains a matte black (hard anodized), and without blemishes on my sample. Lettering is clear, but less bright against the dark background now.

The light still lacks knurling as such, but the ridge detail along both sides of the battery handle has been improved (i.e., less fine now, with sharper edges). I would describe overall grip as ok.

As before, screw threads are square-cut and anodized. 

The tailcap area features boot covers for a physical clicky off/on switch and an electronic mode-changing switch, as before. However, I find both switch covers have been extended further than before (i.e., both protrude more now). Although the original TK35 looked like should be able to tailstand, it was always somewhat unstable due to the bulging on/off switch. The TK35UE takes this even further, with an even easier to access bulging switch. That said, the light can still tailstand at something of an angle (just a bit wobbly). 

Let's look at the carrier:

The carrier is largely unchanged from before. This was always one of the more unique aspects to the TK35, as you don't often see carriers in this class. Made entirely of plastic, it slides fully into the aluminum handle when correctly oriented for the tailcap buttons. The carrier chambers hold all my protected 18650 cells snugly, even the newer high capacity ones. There is a raised button at the positive terminal in each chamber, so flat-top cells work fine.

As before, the carrier contains two different types of switches – a forward clicky for on/off activation, and a secondary electronic switch for mode changing. At the base of the body handle, you will see two buttons. The smaller of these is basically just a small plastic rod with a rubber top that serves as the mode changing button. When assembled, it makes contact with the actual electronic switch on the battery tube. The larger button is connected to the traditional on/off clicky switch on the carrier. Switch feel is good for both buttons, with perhaps a slightly higher traverse than typical. See User Interface discussion below for more info.

Note that there can be a bit of a rattle effect, even when fully assembled. This is likely due to the mode-changing plastic rod. 

The Cree MT-G2 is certainly a different beast from the standard XM-L2 emitters most people are used to. The MT-G2 is a remarkably large emitter, with a dome diameter of almost 8.9mm (vs. 5mm on the XM-L2). Of course, what really matters is the surface area of the die underneath, which is only 2x2mm on the XM-L2. I am not sure of the actual die dimensions on the MT-G2, but there appears to be a grid of 72 distinct segments on it.

Note that the MT-G2 only comes in neutral-warm tint bins. All the MT-G2 samples I've seen have certainly been in the typical "Neutral White" range, although I would peg this example at closer to the cooler ~5000K end of that range.

Due to the large die, don’t expect great throw from a MT-G2 light – especially with the relative small reflector that comes with the TK35UE. The reflector is heavily textured here as well, to help smooth out the beam. This should produce the most "floody" single-emitter MT-G2 I've tested to date. Scroll down for beamshot comparisons.

User Interface

User interface is unchanged from the original TK35. Turn on/off by the raised large forward clicky switch (press-on for momentary, click for locked on). 

Mode changing is controlled by the smaller secondary electronic switch. Click and release the smaller button to advance through output modes, in sequence from Lo > Med > Hi > Turbo, in repeating sequence. Note the manual incorrectly states the reverse level order. The light has mode memory, and retains the last level set when you turn the light off and back on. 

Note that you cannot set the output level while the light is off. The electronic switch only works when the light is powered on by the clicky switch first. As such, there is no standby current on the TK35UE.

The “hidden” strobe and SOS modes are accessed by clicking and holding the mode-changing electronic switch for more than 1 sec. Press and hold for ~1 sec, and the light will first enter a rapid Strobe mode. Press and hold for ~3 secs and the light will enter into SOS mode. To return to the constant out modes, a quick press of the electronic mode switch will suffice (or turn the light off/on at the main switch). 

The TK35 always had an automatic step-down period when run on Turbo (originally, after 25 mins). On the new TK35UE, the light automatically steps down after 5 mins of continuous runtime. Turning off/on restores the initial max Turbo mode.

The TK35UE steps down in output as the batteries drain, and has an additional low-voltage warning feature. As the batteries drain further in the Lo mode, eventually the light will blink 3 times per second, every 5 mins. There is no automatic battery shut-down feature in the circuit.


For more information on the overall build and user interface, please see my video overview:

Video was recorded in 720p, but YouTube typically defaults to 360p. Once the video is running, you can click on the configuration settings icon and select the higher 480p to 720p options. You can also run full-screen. 

As with all my videos, I recommend you have annotations turned on. I commonly update the commentary with additional information or clarifications before publicly releasing the video.


There is no sign of PWM on any level – I believe the light is current-controlled, like other Fenix TK-series lights.  

TK35UE Strobe:

As with other Fenix TK-series lights, strobe is an oscillating frequency strobe, switching between 6.6Hz and 15.3Hz on my sample. Each frequency lasts for about ~2 secs. Here is a blow-up of each strobe frequency individually:

Due to the automatic white balance, don't infer any significant difference between the various Cree MT-G2s shown above. All MT-G2s are somewhere in the traditional "Neutral white" range.

The most obvious observation is that the TK35UE has lower max output than the 3x/4x18650 MT-G2 lights (as you would expect, given the smaller size and lower 2x18650 battery source). The TK35UE is also relatively less throwy – again as expected, given the smaller reflector used here. It is certainly a different beam pattern from the original XM-L-equipped TK35 shown above.

Spring has finally arrived in my part of Canada, but there isn't much of interest for outdoor pics at the moment (i.e., still a pretty bleak landscape). I will be heading out within the next few weeks for some outdoor pics, and will update this review when ready.

In the meantime, you will have to make do with some indoor shots in my basement. For your reference, the back of the couch is about 7 feet away (~2.3m) from the opening of the light, and the far wall is about 18 feet away (~5.9m). Below I am showing a couple of exposures, to allow you to better compare hotspot and spill. For these beamshots, the camera is on auto white balance again, which will distort tints somewhat.

For these comparisons, I'm comparing the Fenix TK35UE to the Eagletac SX25L3 (MT-G2) and the original TK35 (XM-L).

Again, don't make a big deal of tint – the MT-G2 lights are actually warmer in real life (i.e., this is just an artifact of the auto white balance).

As you would expect, the SX25L3 has more overall output than the TK35UE, and is slightly "throwier" as well. A very different beam pattern from the original XM-L-equipped TK35. Scroll down for my detailed testing results

Testing Method: 

All my output numbers are relative for my home-made light box setup, as described on my flashlightreviews.cawebsite. You can directly compare all my relative output values from different reviews - i.e. an output value of "10" in one graph is the same as "10" in another. All runtimes are done under a cooling fan, except for any extended run Lo/Min modes (i.e. >12 hours) which are done without cooling.

I have devised a method for converting my lightbox relative output values (ROV) to estimated Lumens. See myHow to convert Selfbuilt's Lightbox values to Lumens thread for more info. 

Throw/Output Summary Chart:

My summary tables are reported in a manner consistent with the ANSI FL-1 standard for flashlight testing. Please see for a discussion, and a description of all the terms used in these tables. Effective July 2012, I have updated all my Peak Intensity/Beam Distance measures with a NIST-certified Extech EA31 lightmeter (orange highlights).

As expected, the TK35UE has relatively low center beam "throw" for the MT-G2 class (although my sample was slightly higher than the specs would indicate). I suggest you refer back to the beamshots, as that is probably the best way to compare beam profiles between the MT-G2 lights.

It terms of output, you will see my max and min measures look pretty close to the Fenix specs. Here is a breakdown of the estimated lumen values for each level, on my sample:

I generally get very good concordance between my estimated output measures and Fenix specs. Of course, my calibration standard is based only on the lights I have tested, so I don't insist on the absolute value accuracy of relative output values. In this case, I would note that my throw measures for the TK35UE on Turbo were similarly slightly higher than spec. I trust my throw measures to be accurate in absolute terms, as they are based on a NIST-certified and calibrated lightmeter, with the lights tested using standard ANSI FL-1 testing conditions. 

Output/Runtime Graphs:

As always, note that all my runtimes are done under a cooling fan, and I use AW protected 2200mAh cells for all my 18650 tests.

Fo start, I am limiting comparisons to 2x18650 sources (to keep the comparisons fair).

As you can see above, the TK35UE has flat regulation at all levels, with defined step-downs as the batteries drain (with an initial timed step-down of 5 mins on Turbo). Note that you can re-activate max output by an off/on click of the switch. 

Output/runtime efficiency is excellent for the class. In absolute terms, runtime is at least as good as the latest current-controlled XM-L2 lights driven to equivalent output levels.  And of course, the MT-G2 is brighter initially on Turbo than any of my XM-L2 lights.

Now let's see how it does against all the 3x/4x-18650 MT-G2 lights in my collection:

Again, all these comparisons are taking into account different numbers of 18650 cells. But the 2x18650 Fenix TK35UE actually manages to do pretty well for the class, nearing the runtime performance of some of the larger lights.

Potential Issues

The TK35UE uses a plastic battery carrier, as before. Normally, I have some concerns when using all-plastic carriers in lights than can generate a fair amount of heat. However, the TK35 carrier has been around for awhile, and seems to be holding in well.

The TK35UE is not as "throwy" as most lights in this size or class (i.e., has a relatively "floody" beam).

Light is not as "grippy" as some, but is improved from the original TK35. Tailstanding is possible (at an angle), but is not likely to be very stable.

Quality of bundled extras in unknown as full retail packaging wasn't provided (i.e., I haven't seen the holster and lanyard). On my original TK35, quality of these items acceptable (but fairly basic).

Preliminary Observations

In my initial review of the TK35 (XM-L version), I described it as something of a Goldilocks model - as it had a pretty good balance of output and throw, with a reasonable range of levels, while still sitting comfortably in the hand. I would say the TK35UE manages that same sweet spot pretty well for higher output – although with a more evenly "floody" beam now.

It has been interesting to watch the MT-G2 emitter make large in-roads into the flashlight world. Designed for directional lighting applications (e.g., as a halogen spot bulb replacement), this relatively large and "floody" emitter has a different beam pattern than typical flashlight emitters. But it does have a pleasing neutral white tint, and a lot of people do seem to prefer a more even flood beam over throw. The main limiting factor right now seems to be a relative scarcity of MT-G2s – I understand it's hard for manufacturers to secure sufficient inventory.

With the exception of the pure flood Niwalker Nova MM15, most makers have tried to compensate for the large MT-G2 die size by coupling it with a large reflector. Fenix has gone a different route, choosing instead the more compact TK35 build. Keep in mind that this will mean a relatively floody beam, with little directed throw. That said, the beam is very nice and even on my sample. 

One area where you can always expect Fenix to excel is in overall output/runtime efficiency and regulation. TK35UE typically shows slightly longer runtime (for equivalent output) compared to recent current-controlled XM-L2 lights I've tested.  The flat step-down pattern regulation is also appreciated here. Note the TK35UE steps down on Turbo after 5 mins (previously 25 mins on the TK35), although you can restore max output by an off/on restart. Keep in mind the TK35UE may get hot quickly if you attempt to maintain Turbo output in this way.

In my original TK35 review, I was concerned about the durability of the all-plastic battery carrier/switches, and the overall "grippiness" of the light. Grip has been improved somewhat, and there have been a few minor tweaks on the switches as well - but generally the build is largely unchanged. The TK35 has been quite a popular model for Fenix, and I am not aware of any significant build issues. This makes me think the design has held up well to continued use.

At the end of the day, the TK35UE builds on a successful model, incorporating a popular new style of emitter. It is nice to see a MT-G2 emitter meeting a mainstream model – I'm curious to hear what the general community thinks of the novel beam pattern. 

Original text link:

Product Registration | Supply Chains Support | Partners | Resources Download | Legal Declaration | Site Map

Copy Right 2013, All Right Reserved By Fenix 粤ICP备08112176号