To Own One Hotwire: Three LiFePO4 Cells...Which Bulb...?

LEDAdd1ct

Flashlight Enthusiast
Joined
Jul 4, 2007
Messages
3,557
Location
Hudson Valley
The hot wire cannot compete in efficiency with the LED, but there is something special and very nostalgic
about that ancient glow...in that light, I intend to own one hotwire build. I will not own or operate a high current light
with LiCo or LiMn cells stacked in series, so that leaves the much kinder LiFePO4 chemsitry.

Perfect, no, but arguably least cruel and by far least likely to "vent with flame".

While many may argue that with proper care, diligent checking with multimeters before, after, and during use, and constant monitoring of cell behavior you can use, at the least, LiMn cells without too much worry, it is not something I want hanging over my head.

Additionally, I already own three LiFePO4 single cell chargers. There is little I need in that department.

That being established, more than three cells in series is just too much for me to keep track of so we are looking at a three cell build, and therefore, a 9.6 pack voltage.

I've studied Lux Luthor's Destructive Tests, without which I would be hopelessly lost, and narrowed down my search to two potential candidates, the GE 347118 35 watt bulb and the JC 5607 20 watt bulb. In their favor, both of these bulbs, through Lux's tests, will not instaflash at initial powerup, and, both of these bulbs appear to be the brightest 6v nominal bulbs when overdriven.

Of the two, I strongly prefer the GE 347118. According to spec, at a 7.3 amp draw, I should easily net over 2,000 bulb lumens. The JC 5607 does "only" about 1,500 lumens. With a near comparable bulb life, I'd rather have the higher current draw and crank out the extra 500 lumens.

This brings me to my questions:

1) Why can't I find any, or rather, hardly any, information on the GE 347118 bulb? A CPF search yields only a couple pages, and Google only takes me back here.

2) Is there an obscure source of GE 347118 bulbs I don't know about? Feel free to email me if you don't want to publicly expose the secret handshake.

3) Lux Luthor points out here that the Osram 64430 has a higher lux at the same drive current, but fewer lumens. I've nearly always cared more about lumens than lux; I want a more floody light. Am I interpreting his post correctly?

4) Finally, is there another 6v bulb that when taken to 9.6v will outperform any of these three bulbs, GE 347118, JC 5607, and Osram 64430?

I feel like there is something I am missing.
 
Last edited:

JCD

Enlightened
Joined
Apr 12, 2010
Messages
892
Two month bump.

Anybody?

I get the impression that few people are experimenting with the use LiFePO4 cells to power incans. I have just started such experimentation, but with a two cell configuration. I'll be watching this thread with interest.

By the way, what are you using for a host? M@g? Surefire?
 

LEDAdd1ct

Flashlight Enthusiast
Joined
Jul 4, 2007
Messages
3,557
Location
Hudson Valley
I already own three 26650 LiFePO4 cells and three single cell chargers.

I just posted a WTB in "Custom B/S/T" for a 4C Mag that can fit 26mm diameter cells.

So, I will be using a 4C Mag.

I had a 4C Mag for this project, but it went off to vestureofblood for an XM-L U3 mod.

Going for night walks is fun when you light up street signs eight or nine blocks away. :)

Now I want to complete this build, and using those same three cells and chargers,
put a "just right" bulb in a Maglite 4C for some incandescent goodness.

As noted above, the GE 347118 looks awesome.

But the information/availability info on it is scarce.
 
Last edited:

LuxLuthor

Flashaholic
Joined
Nov 5, 2005
Messages
10,654
Location
MS
LED...you don't mind if I just call you LED for short, right? :wave:


I don't read these forums like I used to, so if there is something someone needs in an expeditious manner, they should send me a PM, and/or email me at [email protected] (note the period before the cpf).

Anyway, to the matter at hand...first question about availability of the GE347118. It was apparently a specific local UK design that never had a larger distribution. None of the other companies I ordered from over the years even had it listed in their back catalogs. As I recall, plasmaman (from UK) found a small supply of them, and before buying more wanted me to run tests on the few he had. I wouldn't bother looking for it, as he couldn't even get extra for me later on.

Next, you have to see what is the actual supplied voltage of those cells under various current loads before matching them up with a particular bulb. Your 3s 9.6V default may drop to 8.3 V under a 7-8 amp load with a bulb like the 35W GE, but only to 9.0 V under a 4 amp load of the 20W Hikari 5067. That alone can determine which will deliver higher voltage (& lumens) with a given light resistance setup & specific battery source.

There are issues of bulb quality, glass & filament size & shape, reflector quality/shape/coating/focal length/bulb opening, etc. that can collectively affect the best bulb choice matchup. Some define a good performance based upon optimal throw/center hot spot. Others want a large, diffuse, smooth, flood type output.

Once you know the actual 3s Vbat under 5, 7, 10 amp loads (including the resistance in the flashlight body/switch/springs), the next thing is to look at Ictorana's graphs of my destructive bulb test data in this pinned thread. Finally, after you narrow things down to several choices you have to look at availability, price, rated life, physical bulb size relative to reflector opening and focal length.
 

LEDAdd1ct

Flashlight Enthusiast
Joined
Jul 4, 2007
Messages
3,557
Location
Hudson Valley
Thank you, Lux.

1) Your explanation regarding the lack of availability of the GE 347118 really helps. Thank you.

2) A PDF of my 3.2v cells under load is here. It appears a three to four amp draw is ideal for a combo of runtime/brightness in terms of the voltage curve.

3) The 20W Hikari 5607, your test linked here, is increasingly seeming like the ideal candidate for four key reasons:

a) Availability
b) Cost
c) Brightness
d) Runtime off desired cells

4) A good incandescent bulb to me provides a good mixture of large hotspot with peripheral illumination.

I love the Surefire P90 beam profile.

Right now, it is the only incandescent I own.

5) I am wondering how the Hikari 5607 in a MOP would do.
 

LuxLuthor

Flashaholic
Joined
Nov 5, 2005
Messages
10,654
Location
MS
Those cells hold up close to 3.0V even at 10A, and look pretty awesome! Where did you get them?

I think the 5607 in a MOP is a good choice, but I don't know how it would compare to the P90 (never had one & it is a whole other reflector & small bulb setup).
 

LEDAdd1ct

Flashlight Enthusiast
Joined
Jul 4, 2007
Messages
3,557
Location
Hudson Valley
1) I enjoy smaller reflector setups because I rarely need throw. I like to be able to see the path in front of me.

2) K2 26650 3.2v LiFePO4 cells are easily obtained online, from a dealer and not a shady e-Bay seller.

Unfortunately, they are not cheap.

However, they feature:

a) U.S. sourcing with no wait from China
b) A published discharge curve (good luck finding a published discharge curve for most of the brandless stuff out there)
c) Much safer attributes overall than the li-co chemistry cells insomuch as they are much more tolerant of abuse,
overcharge, and over discharge

When searching, use this term on Google:

"LFP26650EV"

1) The first link will allow you to purchase them
2) The second link is the same discharge curve PDF linked above

Again, they are not cheap. $13.00 each plus expensive shipping.

However, with high capacity 18650 li-co cells going for $20.00+ each, I suppose $13.00 per cell really isn't all that bad.

If you are feeling lucky, you can get ten of them for $100.00 delivered on e-Bay.

Can't speak for the quality from that seller, though.

To charge them, google this query exactly or just copy and paste what is contained between the quotes:

"voltphreaks Single cell LiFePO4 charger (1 to 3)"

It is the first hit.

I've always enjoyed reading your posts over the years, and the LiFePO4 chemistry is the one way I feel comfortable putting into play so much of the data you (and others) have accumulated on hotwires.

Hope they provide some fun for you! :thumbsup:
 
Last edited:

LuxLuthor

Flashaholic
Joined
Nov 5, 2005
Messages
10,654
Location
MS
Thank you sir! That's not a bad cell price at all.

Did a quick comparison to what used to be the Gold Standard A123 "M1" cells (pdf) that we (first found by the RC Forum guys) couldn't get enough of those "safe chemistry" cells back in 2008. Cannibalizing Dewalt power tool packs was the main source.

The bankrupted MIT grad student startup company, A123 Systems was one of many failed US DOE taxpayer grants, and now bought at bankruptcy sale for a song by Chinese company. In any case, while your K2 cells have less amp/watt load rating, they have a whole additional 1000 mAh capacity. Some of us used those A123 M1 cells to make our own lightweight battery packs for the Peakbeam Maxabeam, but as their can was aluminum, we couldn't use our magnet charging leads, which led to various custom charging cradles/clamps that would interface with our balance lead hobby chargers. These companies are all going crazy trying to get chosen for various electric car projects--so this size cell will be around for a good while--and with ongoing improvements in capacity, lifespan, current output profiles..and the thing that sold the RC guys--really fast recharge capability so they could get their planes back up in the air.
 

LEDAdd1ct

Flashlight Enthusiast
Joined
Jul 4, 2007
Messages
3,557
Location
Hudson Valley
With the information contained in this thread, I can use my three 26650 LiFePO4 3.2v cells to power a JC/Hikari 5607.

At 9.0v under load and a measured 4.30 amp draw, and assuming an actual 2600mAh capacity, we get:

2600mAh / 4.30 amp draw = 36 to 40 minutes runtime.

Not bad to play around with.

Now, let's say I want something to run for a longer time.

Colonel Sanders was kind enough to offer to review the Feilong 32650 LiFePO4 3.2v cells I recently purchased.

Assuming they can provide close to their claimed 5000mAh capacity...

I have a copper 2D Maglite, new, sealed in the pack. Been waiting to do something cool with it.

Shouldn't be hard to get another G4 socket and glass lens.

For a long runtime light with two 32650 LiFePO4 3.2v cells in series, what do you think of the MagCharger bulb?

At 6.4 volts, your graph states 1.70 amp draw and 273 lumens. Plenty to hike with, even with some ambient light.

Let's say Colonel Sanders tests show that in reality, they only have 4000mAh capacity at a two amp load.

4000mAh supply / 1700mAh draw = 2.35 hours runtime, or, 2 hours and 21 minutes.

1) Are my math/bulb/lumen calculations correct?

2) What do you think of a 2D copper Maglite with MagCharger bulb for 273 incandescent lumens?

Like the JC/Hikari 5607, MagCharger bulbs are both cheap and widely available. :D
 
Last edited:

bouncer

Enlightened
Joined
Jan 5, 2005
Messages
249
I'm thinking a 3854-L ROP would be fairly bright with a decent runtime with 3 Lifepo4 32650 cells. It comes close to flashing anybody else think it will work or is it too close for comfort? Might try one with my refund check. I'm planning on building an MR-16 with 6 Feilong 32650 Lifepo4 cells anyway what's a few more batteries?
 
Last edited:

Colonel Sanders

Flashlight Enthusiast
Joined
Aug 17, 2010
Messages
1,022
Location
ROLL TIDE!
"Assuming they can provide close to their claimed 5000mAh capacity..."

They do! :cool: Very close in fact (4800-4900) and at any current 5a or under (I didn't test higher as you didn't request that but I would be glad to go as high as you want if you're curious.) I can also tell you that at a 5a load they stay nice and cool and that's a good sign.

So far they seem to be very nice cells.

I am pretty much finished testing now. Just gotta run a few tests on the second cell to make sure it matches well with the cell I've been testing. Then I'll get 'em back in the mail to you.
 

Colonel Sanders

Flashlight Enthusiast
Joined
Aug 17, 2010
Messages
1,022
Location
ROLL TIDE!
There is no question in my mind:

3.65v full charge * 3 = 10.95v

According to this, 3854-L ROP bulb flashes at 9.7v...

I would avoid.

"I was under the impression LIfepo4 cells were 3.2v."

They are 3.2v nominal. I don't have much experience with LiFePO4 cells but what I found in testing these Feilong cells is that they quickly (like, pretty much instantly) drop back to no more than 3.36v resting after being charged to 3.60v. Under load they drop to about 3.10-3.18v and stay there until shortly before exhaustion. Then they quickly drop. Their discharge curve is extremely FLAT. I like that.
 
Last edited:

Colonel Sanders

Flashlight Enthusiast
Joined
Aug 17, 2010
Messages
1,022
Location
ROLL TIDE!
"3.65v full charge * 3 = 10.95v

According to this, 3854-L ROP bulb flashes at 9.7v...

I would avoid."


So, I think the ROP-L would be fine with 3 x LiFePo4 since you're actually looking at no more than about 10v and that's before any resistance losses.
 
Top