It is most certainly a design issue. I think it takes quite some electronics and software expertise and probably extra components on the driver to get the modes right for both AA and 14500 when not all modes are made by PWM (the low mode isn’t). They chose to get the AA modes right and left the 14500 modes for what it is and that is understandable.
Is the Li-ion mode spacing problem because of the current? So there would be the same (or near the same) problem with NiMh eneloop?
Cheaper lights I have offer better aa/14500 versatility. This one kinda disappointed me, it should be advertised as AA only
NiMH works fine in it, it’s the Voltage difference that causes problems when using Li-ion.
When designing a driver that can use both AA/NiMH and Li-ion, compromises must be made. Expense considered…
I don’t have the Jaxman E3 but from the current values I think the 350 lm may be true off the emitter for NiMH for the Cree version:
Fully charged NiMH: 2.3A x 1.3V = 2.99W. — 2.99 x 0.8 = 2.4
Partially charged NiMH: 2.3A x 1.2V = 2.76W — 2.76 x 0.8 = 2.2
Assuming 80% efficiency that means between 2.4W and 2.2W at the emitter.
From CREE PCT, a XP-G2 S4 should output (at 25ºC):
340.3lm at 700mA 3.04V = 2.13W
380.7lm at 800mA 3.07V = 2.46W
419.6lm at 900mA 3.10V = 2.79W
So, 350 lm at 2.4W is doable if the led doesn’t get too hot. That value ignores losses at the reflector and glass.
From Texas_Ace test a Nichia 219C D240 should output about 280lm at 2.1W and 350lm at 2.9W. Assuming 80% efficiency I would expect about 300lm with a NiMH AA for the Nichia version.
2.7A with a li-ion is almost like a 8x7135 linear driver, it should be able to output at least 800 lumens off the emitter.
To get an estimated of the actual output you can measure the lux (aiming the flashlight at the ceiling and measuring the reflection) both for a flashlight with an known accurate lumen rating and for the tested flashlight at the same position. The former will get you a lm/lux ratio for the lux meter position. You may need to repeat the meaurements with the lux meter at 2 or 3 different positions because how floody the light is may affect the results a bit.
See this thread for more details: How accurate is a ceiling bounce test to measure difference in output?
Thanks for explaining.
When I get my E3 I’ll use NiMh for it, then eventually change the driver to use Li-ion and maybe gain a little power boost too.
Jaxman replied me on AliExpress and said the listing on other sites was updated to warn about this, but the listing on AE ended up being left behind and not updated to reflect this. So it’s by spec that it behaves this way. A pity as my intentions involved the dual chemistry capability, but it’s a nice AA flashlight nonetheless and build quality is above average for products I have in this size
Try a LFP cell, 3.2V vs 4.2V.
The Jaxman E3s are wonderful lights.
I have the two-mode version from a few months ago. It works well and the tint and beam are just beautiful. I have more expensive AA lights from EagTac and ZebraLight and the E3 is strong competition for all of them. At the price, it is a no-brainer.
I love the ergos — the light just fits my hand so well. It is light enough and compact enough to easily EDC.
The E3 works well on both Eneloop and 14500. (One of my AW’s is too thick to fit, apparently because of the label; my other AWs fit and work just fine.)
I just ordered two of the three-mode version — a 4000k and a 5700k.
Jaxman, I hope you keep the E3s in production — they will make fine gifts for non-flashaholic friends.
How are the modes with 14500 on the previous driver?
As Jaxman says, there’s just one mode with 14500s.
I have mostly used my two-mode E3 4000k with Eneloops. High produces enough lumens to walk the dogs around the block in a suburban setting. There’s also little heat on high.
Eneloop use also makes the E3 perfect to stash in a drawer, tool chest or car.
I do sometimes switch to 14500s. But that’s mostly out of curiosity to see how much light this little guy can produce. I have 18650s for when I know I want lots of lumens.
“one mode for the 14500” — actually it’s not one mode, it also has 3 modes, however, there is very little visible difference in light output from Low-Med-High. I measured 0.9A, 1.3A and 2.3A using a (nearly-but-not-full-charged) 14500. However, the actual brightness of Low and Med is almost negligible, and even the High mode is not that visibly much brighter than Low and Med (although when High mode, the flashlight will get warmer more quickly).
I suppose that’s what Jaxman means by “one-mode”…
Correction — upon further investigation, there are TWO levels of output using the 14500 on the “two mode” Jaxman E3.
The difference is clearly visible although it is not huge, at least on my wall. And “hi” does put out more heat.
I am looking forward to getting my three-mode E3s — I love this little light.
My only gripe is it being presented as a double chemistry flashlight, but an AA flashlight, yes, it’s impeccable.
For those interested, I posted ceiling bounce measurements of my AA flashlights. I didn’t convert them to lumens because I don’t have benchmark lights that I can claim to be super accurate to obtain a multiplier. However, since I measured so many, you can compare their brightness relative to each other and get a good idea how each perform relative to other AA flashlights.
I just picked one of these up on Aliexpress, and I am really impressed by the quality. I didn’t use it much in stock form (not the flashlight’s fault…I never intended to), but I can tell you that it is an excellent AA host. It is MUCH smaller and thinner than I anticipated. With some filing, I squeezed a triple XPL HI V2 5D Noctigon soldered to an MTN electronics copper S2+ spacer under a Carclo 10507, and dropped in one of MTN’s 17mm FET/DD drivers. It’s a tight fit, but it is INSANELY bright for a AA and handles heat very well. I can also confirm that the deep carry clips from both the Astrolux S3/SC and the Olight S1a fit it perfectly. Fantastic little EDC torch!
That’s an insane mod XD
Apart from the XPL-HI, that is the kind of mod I’ve been wanting to do with that light…when I get one 
Did the Carclo optic fit without being “shaved” or filled? That is my only “if” on that building as I already know what I will prefer for the driver and emitters
If you can, show us the mod on the thread “What did you mod today”
Wow thats a really cool mod. High output and beautiful 5D tint. Do you have any lumen estimates or comparisons with stock emitter?
Unfortunately you do have to shave the optic, but it still had quite a bit of real estate to spare. The inside diameter of the steel bezel is about 18mm, and you lose another 1-1.5mm to the lip. It still had quite a nice beam profile imo, though, for what it’s worth.
I managed to hose up the driver trying to install a bleeder resistor for a lighted tailcap because, you know, I never learned to leave “well enough” alone, but I will post pictures as soon as I have it replaced!