Bulb Overdriving question

[FNinjaP90]

Nerd and I, as participants in the Phoenix(formerly Aurora) project, are trying to come up with a bulb. He told me that the rule for overdriving bulbs is:

+10%= current 105.3% of original, lumens 139.8%, color temperature 103.05%, and 70% reduction in life.

This formula works up until the melting point of Tungsten(3695K) right? He said that for safety, we should keep it below 3400K.

So, if I overdrive the Osram 64447 bulb (12V 65W 1700lu 3000K 4000hrs) to 40%, then I will end up with 3366K, which is still acceptable. For those that don't know what the 64447 is, it's the badass axial LOLA bulb originally used for the phoenix with the return wire going down INSIDE the spiral.

If that's right, then I will end up with a 110.3W bulb being overdriven at 16.8V, with a color temp of 3366K, putting out 4406(!) lumens at an ASTOUNDING 39.96lu/W, with a life of 32.4hrs.

We could get brighter with the 64623 bulb, but the filament in it is HUGE and will be very poorly focused. It is also because the 64623 is placed perpendicular to the reflector, while the 64447 axial is placed parallel.

So does that formula always work for every bulb? Is it true that every bulb can be driven to 3400K? Are there any other factors to how high a bulb can be overdriven?

If it's true and there are no other factors, then we have a ~4500lumen bulb that will deliver the whitest light possible from incandescent technology, while retaining a perfect beam and an astounding 40+ lu/W!

 

[Ginseng]

1. The CCT or Correlated Color Temperature of the light is not the same as the actual temperature of the incandescent filament. In fact, it is not uncommon to find bulbs with CCTs of well over 4000K and approaching 5000K.

2. The bulb you're referring to is the 64447IRC not the 64447. The IRC version is the one that is axial and longlife. Same issue with the 64440 and 64440IRC.

3. While the hotspot will be circular as opposed to the oval provided by the 64625HLX and 64623HLX, it will not necessarily focus down tighter or be hotter. I've already explained why this is in a previous thread. Basically, it has to do with the fact that unlike a horizontal filament, if the axial filament is perfectly centered, there is no tungsten at the exact geometric focal point. Additionally, the ends of the filament which are off the focal point result in two separate lobes of light that overlap and make a ferociously bright hotspot. This is why the WA01185 makes a very hot hotspot even with its massive filament. In filaments of this size, it makes very little difference in hotspot intensity until you get up into a 6" reflector and preferably the 9" of the Blitz. I should be able to report on the Blitz bulb in a 3" reflector in a week or so.

3. The rerating formulas become increasingly unreliable the farther you get away from the design operating specifications. 16.8V would be considered extreme. Solux, for example, do not list rerated values for CCT or life above 15V for their 12V, 3,000 hour lamps. BTW, at 15V, their 4700K bulbs have a CCT of 5500K.

4. What is known is that bulbs will get brighter as you overdrive. Bulbs with lives under 100 hours are not reliably overdriven. Bulbs of 3,000 to 4,000 hour life can be overdriven to 16 and even 18 volts, respectively, for lives of between 25 and 50 hours. Efficiency increases as you increase overdrive.

5. The initial overvoltage problem becomes more severe the more cells you stack. For example, to get 16.8V nominal at 50% depletion, you'd need 15-16 of the KAN1800 cells. Your startup voltage, however, would be a sizzling 21V-22.4V or dead-bulb territory. Of course, the way around this is to use a PWM circuit. However, the catch there is that to get 16.8V at a per-cell minimum of 0.9V, you'll need a 19 cell stack. For a low voltage cutoff of 0.8V/cell, you'd need 21 cells. That's a Mag 6D with three columns of 7 cells.

6. I learned a lot, much of it through hard work on the Aurora Project. While I can't participate, I just can't help sharing what I've learned where appropriate. I may ramble on a bit so if you guys want me to clam up, please say so. This is your show and you're entitled to learn as you go.

Wilkey

 

[Deleted member 684]

I greatly appreciate your input Ginseng. We are indeed planning to use a PWM circuit. As for the cells, we're still exploring and experimenting. While we have theoretical hypothesis, we lack the actual data needed that will come when we test the cells.

I think that having a bag of different bulbs to test their overdriving capabilities should prove to be pretty interesting... of cos life span can't be tested... lol!

Respectfully

 

[js]

Nerd,

Where did you guys get a 64447-IRC? And how much did it cost? And was it a special order item?

Also, I hope you realize hot HOT a 110 watt setup is going to get. Of course, at that wattage, I'm sure your runtime will be like 5-10 minutes, so perhaps it's not an issue. /ubbthreads/images/graemlins/grin.gif

 

[Deleted member 684]

Heh, js, I haven't got those bulbs, yet. Looking at places to get a bunch of em to test it out, otherwise shipping will kill me, buying from all the different sites.

Actually Phoenix version 2 is s'pose to be running a 250 watt bulb..... /ubbthreads/images/graemlins/sssh.gif And no, don't bother asking about runtime.... /ubbthreads/images/graemlins/jpshakehead.gif

 

[FNinjaP90]

So you are saying we can run the 64623HLX and the 64447IRC at 16 and 18V respectively? What color temperature might we see? If our formula is incorrect, how many lumens might we be able to get?


I personally, want to stay around 5000 lumens max. I'd want a tad less light if I can get more runtime out of it. Of course, I still want an overdriven bulb for whiteness so maybe an overdriven 64447 would suit me best?

Also, you said about the 64440IRC in the original Aurora thread:

"There is no winging or lobing whatsoever. In fact, it throws as tight of a spot with the "stealth" 2" head as the regular horizontal filament bulbs throw in the BigHead 3-incher."

You just contradicted yourself! You said in the first post of this thread that the horizontal filaments had a tighter hotspot (and implied farther throw).

So if the axials DO throw farther, does that mean that a 64447IRC with equal wattage as a 64623HLX's will outthrow the HLX?

 

[Ginseng]

Let's get something straight. I can swap axial and horizontal filament bulbs in my Aurora at will and have done so. I know what these bulbs can do first hand. There is maybe one other person here who can do this or has already done so. What I stated in the original Aurora thread was an apples to oranges comparison meant to illustrate the fact that the spot of an axial bulb is compact and circular due to its filament orientation. In my post above, I stated that the H-bulb spot was "ferociously bright" with no specific reference to the relative tightness.

If there are seeming contradictions in what I report, then it's either my text being misinterpreted or I have been unclear and further elaboration would be helpful. What is perfectly clear are my observations, which I have made the time and effort to document here.

I think I'll just let you guys hammer it out. Let my final comment be that at 4,000 lumens, you really won't give a damn about whether it's 4,000K or 5,000K CCT because it's going to be so much brighter than anything else you're likely to have seen.

Wilkey

 

[js]

Wilkey,

You said "let's get something straight." Then you talk about "first hand" which is either your left or your right, yes? So that's not straight! Hello? And what about "apples" and "oranges"? Everybody knows they're not straight!

Also, doesn't compactness also indicate BOTH apples and oranges?

I'm sorry, but, I'm going to have to lower your CCT rating from 7,500 to 7,499.

This is still 2,478.34 K above the nearest CPF competitor, so you're still the INCAN MASTER but I'll have to stamp your meal ticket with "NO DESSERT" for today.

/ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif

Wilkey, how are you doing? Check your email.

 

[FNinjaP90]

Ok, so with 2 bulbs, one axial and the other horizontal, with the same wattage, the axial bulb will throw farther. Am I right? That's all I wanted to know. As long as we can narrow it down to one bulb, we can do the testing ourselves.

 

[bwaites]

FNinjaP90,

The simple answer is that you cannot make a general assumption like that. Simply because the hot spot is round does not mean that the reflector is optimized for longest throw with that type of filament.

In fact, most, if not all, of the common reflectors are optimized for non-axial filaments, so an axial filament may/may not throw further in any given reflector.

What Ginseng is trying to say is that you cannot make the assumptions you are making without all the other variables being figured in.

P.S. At least thats MY intrepretation!

Bill

 

[FNinjaP90]

Yeah but he has already done it. You don't need to make assumptions if you have already seen it happen.

All I'm asking is that on the axial bulb, does it look like more of the light output is used for the spot than the horizontal bulb?

I guess we'll just have to test for ourselves.

 

[bwaites]

You are asking him to answer a question that has NO answer!

Under SOME circumstances the axial filament MIGHT outthrow the other options, but without knowing all the possibel variations from his Aurora, he can't give you an exact answer.

Bill

 

[FNinjaP90]

Oh ok. By the way, I'm asking the throw question based on the bulbs being in a 2102 3" reflector. So there shouldn't really be any other variables.

But it's ok. We'll just do it ourselves. We will get a 1600lm 64611HLX and a 1700lm 64447IRC. Since they both have roughly the same lumen output, we'll see which one throws farther. The result should be the same as a 4000lm 64623 and a 4000lm 64447.

 

[Sway]

FNinjaP90

It all comes down to the angle and size and of the reflector like Wilkey said. I don’t have a 3” MagDaddy to play with (wish I did) and this puts me out of your test parameters but I have tried a 100W axial filament in Oto’s PMR with less than impressive results the filament is to long vertically for the reflector and angle.

Yes it’s a super bright light but the corona is just about as bright as the hot spot /ubbthreads/images/graemlins/icon3.gif My SWAG is the horizontal filament is going to give the most bang for the buck with this style reflector with nothing to chase down here till you get into a deeper reflector designed to make the best of the vertical filament with the proper angle to get the most out of it then you will see a difference in the throw.

Hope this helps /ubbthreads/images/graemlins/smile.gif

Later
Sway

 

[js]

[ QUOTE ]
FNinjaP90 said:
Oh ok. By the way, I'm asking the throw question based on the bulbs being in a 2102 3" reflector. So there shouldn't really be any other variables.

But it's ok. We'll just do it ourselves. We will get a 1600lm 64611HLX and a 1700lm 64447IRC. Since they both have roughly the same lumen output, we'll see which one throws farther. The result should be the same as a 4000lm 64623 and a 4000lm 64447.

[/ QUOTE ]

Not only will you have to do it yourself, you SHOULD do it yourself! What's going on here? Hello !?! You're modding. Roll up your sleeves and start getting your hands dirty. There's simply no substitute for direct experience. I'm not saying not to try to get good advice from experts such as Ginseng; I'm just saying that it seems to me that you can't afford NOT to try out both axial and horizontal OSRAM lamps for yourselves. I can't even believe that you were going to eliminate one or the other without having tried it. Even if you're sure the axial is the better choice (and it IS pretty cool /ubbthreads/images/graemlins/grin.gif ) you'd want to be able to say how and why it was better than the horizontal filament. Can't do that unless you've seen both.

Besides, it's just plain FUN to try out the different options. Go ahead. Have some fun. And be nice to Wilkey. He generously and unstintingly provides a great deal of incan expertise and wisdom to all of us here at CPF.

 

[Ginseng]

Good responses all. Let me see if I can wrap up a few loose ends before I go offline. Sorry, it is a bit long but I wanted to be as complete and as descriptive as possible.

js,
I wonder about you sometimes, buddy. /ubbthreads/images/graemlins/wink.gif

bwaites,
You have nudged up against one of the hidden aspects of reflector-filament interaction and the resulting effect on collimation and spot formation. There is an interesting thread in the "Machining" forum about parabolic reflectors. In essence, there is nothing mysterious about parabolic reflectors. Nothing at all. Parabolic reflectors have a simple mathematical description and their function is equally simple: collect light from a point source and, through reflection, redirect that light into a beam of parallel rays of diameter equal to the diameter of the reflector. That's it. That's all. Nothing more, nothing less. Why then do we see such a vast assortment of behavior from the myriad of reflectors in the myriad of lights out there?

The answer, as you may have intuitively sensed and empirically observed is not in the theoretical structure of the reflector itself, but rather, in the deviation from theoretical. That is to say, what happens when a real non-point source of a given geometry is placed in a real reflector of a given geometry. The most basic aspect of the reflector that affects performance is the depth to diameter ratio. This is related to the focal length of the reflector. On the one extreme are wide and shallow reflectors like the 6" unit found on the Vector 2MCP handheld spotlight. Let's call this the type "S" reflector. Characteristic of the "S" type is a shallow dish and a long focal length (FL). The FL of the V2MCP reflector is about 25mm. There's nothing magical about this number. It is simply a consequence of the particular parabolic formula used to generate this dish. Another more extreme "S" type is the Edmund Optical 12" parabolic. It is very shallow and has a FL of 75mm. I do not know the FL of the Blitz 9" unit.

On the other extreme are the deep dish or "D" type reflectors. The Mag and Carley reflectors are of this design. The Carley RF1940, a popular replacement for the identical Mag reflector, is a 2" diameter unit with a 5.08mm FL. The RF2102 parabolic has a 3" diameter and a 6.1mm FL. I'll leave it to you to compare the diameter to FL ratios of these representative units.

Ok, so if you'll accept that among commercial reflectors, there are two possible extremes in terms of the depth to diameter relationship (and the resulting leverage on focal length) what then is the significance of all this? There are two results of signfiicance:
1. Deeper "D" reflectors collect more of the light for collimation than shallower "S" reflectors. This is strictly a geometric consideration. Deeper reflectors encompass a larger total solid angle of the filament radiation.

and the more significant result...

2. "S" type reflectors are more insensitive to whether the filament is orientated axially or horizontally. This is because the shallower (or relatively wider) reflector will have a higher tolerance for off-focal point light sources (especially in the horizontal plane parallel to the front face of the reflector's largest diameter).

So what? Well, one practical fallout of this is that handheld spots almost without exception use "S" type reflectors because the diameters are large and a deep 6" reflector would make the light unacceptably bulky. The unintended benefit of this is that in lights of 6" or larger "S" type, the resultant spot will be pretty darned compact. Also, since handheld spots typically employ high output bulbs in the 1,200-3,000 lumen range, the slight hit in collection efficiency will hardly be missed. Here's an example of what I'm talking about.

As for smaller lights like the MiniMag and Mag C/D type, they use pretty dim stock bulbs so collection efficiency is a significant design consideration. Also, a 2" deep 2" wide reflector encompasses maybe 4 cubic inches whereas a 6" deep 6" wide reflector would require over 110 cubic inches. The unfortunate result of the "D" type geometry in small torches is the dreaded "batwing" or bilobe. You'll see this to some degree with nearly any bulb you wish to put in but to a severe degree with the popular high output bulbs from WelchAllyn and Carley.

So, getting to Sway's point, the reflector must have the proper tolerance in the dimension most critically related to the orientation of the filament. That's right, there is no one "best" solution. And as for throw, believe it or not, whether you go axial or horizontal in a small reflector like the RF2102 in the Aurora doesn't make a hell of a lot of difference (I used the 64440non-IRC and 64440IRC as tests). The H-bulbs throw a hotter spot just as far as the V-bulbs. The V-bulbs throw a circular spot that is more diffuse than the axial. This all changes when you go to the 6" unit in the V2MCP with the two spots starting to look indistinguishable at a distance. And I'm talking at a range of 250-400ft. Realistically, I don't have white walls at this distance so that's interpreting what I see against a stand of trees.

That's it. That's all I've got. If you want more, you'll have to try it for yourself to find the specific information you need.

Wilkey

BTW, a 6" true "D" type reflector doesn't exist...not right at this moment. /ubbthreads/images/graemlins/wink.gif

 

[FNinjaP90]

Thank you for that information guys. We should be able to get started soon.

 

[Deleted member 684]

I wonder how much it would cost for Carley to machine a 4" wide, 4" deep reflector............. /ubbthreads/images/graemlins/grin.gif 5 inches? /ubbthreads/images/graemlins/wink.gif

 

[PeLu]

Maybe it is uneccessary to tell here, but when overdrivng these bulbs it is, of course, important that they get no additional overvoltage and that they get a soft start. Best device to manage that is the Willie Hunt LVR.

I always thought that a soft start is not necessary but Willie proved the opposite.

 

[NewBie]

In re-doing my Thor, I found that Soft Start allowed me to over drive alot further, and that utilizing a regulated voltage so there is no overshoot whatsoever, goes a long way in preventing a person from going through bulbs like a flashbulb. I also like the fact that the voltage regulator holds the output rock solid for output level. It even holds the CCT. The fact it was a boost allows me to pull more juice out of the battery at a useful brightness. One of these days I want to build a big brother of the 240W Soft Start boost voltage regulator to get into the 500W-1000W range, but I'll have to find batteries that can handle the current draw or use higher voltage battery stacks.