"Traffic halogen" 10V 50W bulbs- good for hotwiring?

[Vargen]

Hi, I'm new here and fairly new to modding flashlights. All I've done so far is bore out a SF M3 to take usefully-sized Li-ions, and install LF bulbs.

Now I'm thinking I should try upgrading a Maglite that's mostly gathering dust at the moment. So, I started browsing bulb manufacturers' catalogs in search of suitable bulbs to use.

In the Philips catalog, I came across an entry for a "Traffic Halogen 50W 10V", spec'd to give 820 lumens at 10V and a 9000 hour bulb life. These longlife 10V lamps are apparently made for use in traffic lights (stoplights etc). They have the same base as automotive H3 headlight bulbs, only they're made for 10V rather than 12. Now, with a 9000 hour life these should easily take some modest overdriving; I figure they'll work fine at the voltages you get from 3X Li-ion rechargeables in series.

However, that 50W bulb entry looks a bit incomplete, which makes me wonder if this is a real product. Their listing for an otherwise-identical 35W bulb seems more complete; this one is listed as "Traffic Halogen 6611 35W PKX22s 10V". The 35W version is listed as giving 525 lumens with a 5500 hour life.

Other manufacturers also list 10V bulbs with similar specs; Osram has one with a 950 lumen output, but a rater short bulb life.

The reason I'm looking at these is that it's relatively easy to hack an H3-style lamp base apart so you get a bipin bulb. And something that runs brightly on an easily obtainable voltage from Li-ions sounds good. About the only downside I can see is that the bulbs are relatively large, 11 or 11.5mm in diameter.

Any thoughts or experiences with these bulbs?

 

[LumenHound]

The pins on a hacked H3 bulb can sometimes be only .5mm thick.
If you fit it into a socket that has held something like a 12V 50W G6.35 bulb with it's 1.0mm thick pins previously then the socket may not be able to hold the thinner pins firmly enough.

 

[mdocod]

Hello Vargen,

Welcome to CPF!

My thoughts:

A 10V 9000 hour bulb would require about 17V to wake up to an efficiency level worth running in a portable short-running flashlight. (it would become roughly a 4000 lumen ~120W bulb)

What if: These "10V 9000 hour" bulbs are actually 12V 1000 hour bulbs that have been re-rated for a different application?

When designing a "hotwire" configuration for a bulb and battery combo that will be impressive, we generally shoot to reduce bulb life to the 10-20 hour range. In that range, bulbs wakup to the ~25-35 lumen per watt range and produce a CCT of 3400K or better. This results in an impressive performance. Running long-life halogens at their specified input voltage results in a yellowed light that doesn't generate the kind of lumens for runtime that one would want to see in a portable light to justify the relatively short runtimes one must expect when running high wattage bulbs on a small battery pack.

What size mag are you working with here?

Eric

 

[Vargen]

I was originally thinking I'd use a 3D, but I've also got a 4D I could play with. Note that if I built something like this, I'd want it to be reasonably reliable so it could actually be used; a 4000-lumen light that pops its bulb every few times I turn it on will be of no use to me. Therefore I'd be willing to sacrifice some light output for reliability and a reasonable bulb life.

If I have the mag body tri-bored and run 17500 cells with three cells in parallel and however many is needed in series, wouldn't that be able to deliver the current for this kind of lamp?

 

[mdocod]

Each 17500 size LiCo cell can safely deliver a maximum of about 8 watts of energy. I would shoot for no more than 6-7 watts per cell in a large pack configuration to help introduce a margin of safety. Personally speaking, I would prefer to see a configuration that did not stress the cells beyond 1.5C (~6W per cell)), and ideally, 1.0C (4W per cell). Keeping the drain rate to a reasonable level will help keep the cells wearing more evenly in the pack, and result in better cycle life and overall safety.

Coincidence would have it, that a 3S3P configuration of 17500 cells (this would fit in a 3D tri-bore) driving the bulb you are considering would result in an operating power of around 60W, or about 6.7W per 17500 cell, not bad. Assuming that the manufacture listed specifications are accurate and the bulb is not vaporware: Bulb life would be approximately 2500 hours, output approximately 1100 bulb lumen (~700 torch lumen). 18 lumens per watt. 30-35 minutes runtime. Moving up to the 4D host and using 17670 cells in the same configuration would extend run-time to 45-50 minutes.

I did a little searching around on the net and came up with some strange results for 10V 50W traffic signal bulbs. Some are listed, as the one you found is, at 820 lumen, but I also found some rated a measly 220 lumen for the same power consumption and even less bulb life. Weird eh? The other weird thing, is that, most halogen bulbs rated in the 800-1000 lumen range with 50 watts of power consumption are 2000-4000 hour bulbs.

Anyways, moving up to 4S3P 17500s in the 4D, you could drive the same bulb if you wanted to. Things would start to look a little more exciting though: (again, assuming the bulb specs are real): ~2500 bulb lumen (~1600 torch lumen), ~90W, ~50-100 hours life (better than most tactical incandecent flashlights), 28lm/W, 20-30 minutes runtime. This configuration would push the cells right up approaching the 8W per cell discharge rate and the runtime suffers a bit more.

Running the numbers for that 35W version you found:
on 3S3P 17500: 42W, 700L (450TL), 17L/W, 1500HR, ~50-55 minutes run
on 3S3P 17670: 42W, 700L (450TL), 17L/W, 1500HR, ~70-75 minutes run
on 4S3P 17500: 65W, 1600L (1000TL), 25L/W, 25-50HR, 45 minutes runtime.

-------------

If I were looking to do mods in similar classes of output, I would use cheaper, easier to find bulbs... Use the 4S3P configuration of 17500s to drive 12V long-life bulbs. The overdrive will still leave you in the 150-300 hour of life range (which would be considered by almost anyone here to be a dramatically conservative level of drive with near-zero chance of a premature failure from aging).

Typical 12V20W2000HR bi-pin halogen will give you ~600 bulb lumen for just 28 watts. (as apposed to 700 lumen costing 42 W in the 35W bulb above). That's ~21.5L/W

Typical 12V35W2000HR bi-pin halogen will give you 1100 lumens for about 50W, 22L/W or better.

How about a 50W IRC bulb. The 64440. That would give you 2200 lumens for under 70W. That's 31+L/W and still leaves 300 hours of bulb life on the table in this 4S3P configuration.

Hope that helps

Eric

 

[Vargen]

Thanks, that was very informative.

I guess what I really want is a dependable light with at least 1000 lumens out the front, more output would be nice but is less important than reliability. Runtime should be at least 20-30 minutes, again more would be nice and ideally something in the 45-50 minute range. I can live with a 20-minute runtime though, I've got other lights for when long runtime is needed.

Size isn't critical, but I'd like it not to be much larger than a 4D. Ideally it should look like a reasonably stock maglite, so my colleagues and clients aren't distracted by a piece of glaringly nonstandard kit dangling from my belt (that's when the light isn't on, of course ).

I'm handy with a soldering iron and I have some machining capabilities, so modifying switches etc is no obstacle.

Finally, I would like to use a bulb that's in production and seems likely to be easily available (for a given value of "easy") for years. I'm in Europe, if that matters. I'm pretty sure I can order anything that's actually in production out of the major brand-name catalogues (Philips, Osram etc). More obscure stuff might be trickier.

My reason for choosing 17mm diameter Li-ion cells is simply that I'm already using 17500s so have the charger and some cells, a DMM for testing voltages, and some understanding of the risks involved with that battery chemistry. Switching to a different type of cell would involve more research, and probably a new charger.

The reason I was looking at these signal bulbs is that they seemed to be made for about the right voltages. I'd done some back-of-the-envelope calculations using formulas found here, but hadn't thought to figure out the load per cell properly. Also I hadn't really thought about the efficiency or lack thereof, I was looking at bulb lumens and amps but hadn't thought to think in terms of lumens per watt cost.

I did think the manufacturer's specs for that 50W bulb seemed a bit weird, especially as that bulb lacked a proper model number in the Philips catalogue (unlike the 35W version).

BTW, Osram lists an otherwise-identical 10V 50W bulb (SIG 64014) as giving 950 lumens with a 2000 hour bulb life. Doing the math it seems driving that off 4S3P would be very close to instaflash on freshly charged cells, unless voltage loss to resistance in the flashlight (or a regulated driver, obviously) gets it down to barely-manageable levels.

Osram also has the SIG 64015, a 10V 50W bulb with 750 lm and 8000hr life. These specs seem not too horribly different from the Philips ones.

I'll have a look at the various 12V bulbs, in particular the availability of 64440s. Thanks!

 

[mdocod]

Also check out the 64432 lamp. It'll give you ~1750 bulb lumens on the 4S3P pack for about 53W. 50+ minutes runtime there. 300 hr bulb life. That will have just enough oomf to meet your "over 1000 out the front" requirement. ~33L/W.

Please use protected cells in any setup you choose.

 

[Vargen]

Thank you, that one sounds like it fits very nicely. 50 minutes runtime would be great, it's more than I usually need but there have been a few situations where it would have been useful.

And yes, my main reason for choosing a series/parallel setup is so I can get enough amps without having to use unprotected cells. Those things are scary, more so in a pack of 12 cells!

 

[LuxLuthor]

Thanks, that was very informative.

I guess what I really want is a dependable light with at least 1000 lumens out the front, more output would be nice but is less important than reliability. Runtime should be at least 20-30 minutes, again more would be nice and ideally something in the 45-50 minute range. I can live with a 20-minute runtime though, I've got other lights for when long runtime is needed.

Size isn't critical, but I'd like it not to be much larger than a 4D. Ideally it should look like a reasonably stock maglite, so my colleagues and clients aren't distracted by a piece of glaringly nonstandard kit dangling from my belt (that's when the light isn't on, of course ).

I'm handy with a soldering iron and I have some machining capabilities, so modifying switches etc is no obstacle.

Finally, I would like to use a bulb that's in production and seems likely to be easily available (for a given value of "easy") for years. I'm in Europe, if that matters. I'm pretty sure I can order anything that's actually in production out of the major brand-name catalogues (Philips, Osram etc). More obscure stuff might be trickier.

My reason for choosing 17mm diameter Li-ion cells is simply that I'm already using 17500s so have the charger and some cells, a DMM for testing voltages, and some understanding of the risks involved with that battery chemistry. Switching to a different type of cell would involve more research, and probably a new charger.

The reason I was looking at these signal bulbs is that they seemed to be made for about the right voltages. I'd done some back-of-the-envelope calculations using formulas found here, but hadn't thought to figure out the load per cell properly. Also I hadn't really thought about the efficiency or lack thereof, I was looking at bulb lumens and amps but hadn't thought to think in terms of lumens per watt cost.

I did think the manufacturer's specs for that 50W bulb seemed a bit weird, especially as that bulb lacked a proper model number in the Philips catalogue (unlike the 35W version).

BTW, Osram lists an otherwise-identical 10V 50W bulb (SIG 64014) as giving 950 lumens with a 2000 hour bulb life. Doing the math it seems driving that off 4S3P would be very close to instaflash on freshly charged cells, unless voltage loss to resistance in the flashlight (or a regulated driver, obviously) gets it down to barely-manageable levels.

Osram also has the SIG 64015, a 10V 50W bulb with 750 lm and 8000hr life. These specs seem not too horribly different from the Philips ones.

I'll have a look at the various 12V bulbs, in particular the availability of 64440s. Thanks!

My first "real" incan light was one that FiveMega made in several runs, using a Welch Allyn 1185 bulb in a "tri-bored" 3D mag. I have two of these, and put some Pachmayr "Pac-Skin" (5x20 size) which is a perfect circumference fit, and gives great insulation in cold weather. With the 3x3 cell setup, it gives "1200 lumens" for about 50 mins. It's another option.

There are smaller ways to get the job done, but 45-50 mins narrows the +1000 lumen candidates.