# lumens vs. candlepower

#### BugLightGeek

##### Flashlight Enthusiast
So, I'm doing some shopping and trying to determine which light is brighter. One manufacturer rates everything in lumens and one rates in candlepower.

How can a newbie figure out which will be brighter?

#### Darell

##### Flashaholic
Unfortunately, it is almost impossible to compare.

Your best bet is to buy them both, then tell everybody here what you've found. "But them both" is somewhat of our unofficial motto around here...

#### BugLightGeek

##### Flashlight Enthusiast
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by darell:
Unfortunately, it is almost impossible to compare.

Your best bet is to buy them both, then tell everybody here what you've found. "But them both" is somewhat of our unofficial motto around here...<HR></BLOCKQUOTE>

Sorry, I wasn't specific. OF COURSE I'll probably end up buying them both, but for now, I'd like to have a sense of comparison between candlepower and lumens...

#### rlhess

##### Enlightened
You can use a (walk-in) integrating sphere, or you can integrate the light output over area and do the math.

I haven't done it yet, but I might try it on the plane tomorrow--or maybe I'll sleep on the plane.

Cheers,

Richard

#### BugLightGeek

##### Flashlight Enthusiast
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by rlhess:
You can use a (walk-in) integrating sphere, or you can integrate the light output over area and do the math.

I haven't done it yet, but I might try it on the plane tomorrow--or maybe I'll sleep on the plane.

Cheers,

Richard<HR></BLOCKQUOTE>

Whoa...you lost me on that one. But, if you could provide me with some sort of conversion, that would be appreciated.

#### Chris M.

##### Flashlight Enthusiast
Greetments herr Doppelganger....

There is no easy or direct comparison between Lumen and Candlepower ratings.

Lumens are a measure of the total amount of light emitted in all directions, and while not entirely appropriate when applied to directional light sources like flashlights, it still can be used to measure all the light poured out the end of your torch, regardless of what direction it`s going in.

Candlepower or Candela is a peak spot measure- the intensity of the very brightest point of light within the focussed beam, usually the very center. It does relate with Lumens, to the half angle of the beam- in other words, the angle from the centre at which the brightness is exactly half of the peak value at the centre. For an identical intensity light source (ie, lumen value is the same) the narrower the beam the higher the Candlepower rating. But the dispersion of the beam plays a big role. It depends largely on the width of the hotspot, what portion of the light goes into the main beam, and what portion spills off to the side.

Unless you know the precise dispersion of the individual beams, with an accurate 2 dimentional plot of intensity (or a simple 1-dimention graph if the dispertion is perfectly circular and not lopsided or oval), and then apply advanced mathematics that would make most highschool students heads spin, you cannot convert between the two. Can`t be done, sorry.

Regrettably there is not a simple home-friendly way to measure lumen outputs. A simple photographic Lux meter can be used to give you candela (candlepower) readings but Lumens require advanced methods. You need lots of space, lots of money and lots of Barium Sulphate paint in order to do it. Hence why most manufacturers just use candlepower. Additionally (and maybe more influencially), candlepower figures are often high and look great on fancy packaging and web sites- that thing about narrow beams again. Many Pelican lights and all Maglites for example, are rated quite highly as far as cp goes, but the overall light output is low compared to some, just the fact that they can focus pretty narrowly. A narrow beam is not always a useful beam so candlepower ratings cannot be soley relied upon. Similarly Lumen ratings cannot be soley relied upon either. A high-lumen beam could be so wide and dispersed that it does not throw more than a few feet, or a low lumen beam could be so narrow that it shines a very long way.

It`s all down to what you`re looking for in a lignt and until flashlight manufacturers get their act together and embrace a universal, standardised and informative rating method they will all use, the only real way to tell if one torch is better for you than another is to try it out yourself. Hence the official CPF motto "Get them both
". We reviewers try our best but however detailed we try to be in our writeups, you still cannot beat seeing the thing for yourself.

#### Darell

##### Flashaholic
Wow!

Chris made my "it is almost impossible to compare" look really concise, huh?

That Chris M. really knows his stuff!

#### rlhess

##### Enlightened
Hi, Chris,

I was suggesting that we could approach a ballpark total lumens output by, for example, taking illuminance measurements at every (let's say) five degrees, determining the area over which that is falling and do that until we get to a five degree ring that is at say 1% of the center output.

Yes, we'd have to do cosine correction (depending on how we held the light meter) and use s different radius value in the inverse square compensation (or move the light meter in a spherical path) for the outer rings, especially, and we'd have to come to some agreement as to the level/area for the center hotspot, but that would start to create a figure that could be considered "pseudo lumens" and a first-order approximation to the integrating-sphere measurements.

I wasn't meaning to say that we could approach laboratory conditions, but, for example, the illuminance charts that Craig does might be able to be used to provide an approximation of total light output.

I don't think the "integrating ceiling" approach is valid as there are too many other variables there.

But, what I'm proposing is a fair annoyance with trig functions and other calcs for only a rough approximation.

Since I "know" that my SF E2e puts out 60 lm, I thought it would be interesting to use that as a model for the hypothetical rough-order-of-magnitude conversion.

If I have time, I'll build the spreadsheet.

I doubt I'll have time to take all the measurements for a week or more.

Cheers,

Richard

#### treek13

##### Flashlight Enthusiast
Ok, first I had this whole lumen vs. candlepower thing to try to sort out and now I've got this ChrisM vs. Chris M. problem.

Now I'm really confused; heck I'm not even sure if I have their names straight here.

#### BugLightGeek

##### Flashlight Enthusiast
First, congrats on the web page contest.

Wow, Chris M, a most excellent response. Even though it just boils down to GTB! I was hoping for some sort of conversion formula but oh well.

#### Chris M.

##### Flashlight Enthusiast
I was suggesting that we could approach a ballpark total lumens output by, for example, taking illuminance measurements at every (let's say) five degrees, determining the area over which that is falling and do that until we get to a five degree ring that is at say 1% of the center output....

....I wasn't meaning to say that we could approach laboratory conditions, but, for example, the illuminance charts that Craig does might be able to be used to provide an approximation of total light output.

That`s an interesting idea- assuming you had a perfectly round beam it could work. Lopsided beams have too many variables that would make for some horrendous 2 dimentional analysis with hundreds of measurement points and enough calculations to fill all 65000 rows that MS-Excel provides per-spreadsheet!
All I can say is- Try it and see! The mathematics involved is waaaaay over my head though so you`re on your own there I`m afraid. I`m now a two-times university dropout resitting and failing the final year of my Electrical Power Engineering twice. The mathematical element of the course was my biggest fall-down point. That and the fact that now I know what it`s all about, I don`t actually want to be an electrical power engineer and was tired of the course
(don`t worry, I`m glad to be out of that place now and am in a job I quite enjoy for now so it`s all OK).

Since I "know" that my SF E2e puts out 60 lm, I thought it would be interesting to use that as a model for the hypothetical rough-order-of-magnitude conversion.

Here`s the big fall-down point, you would need a few lights that you knew the exact lumen output of, in order to "calibrate" and otherwise prove or deny your measurments and calculations. I am not 100% sure that SureFire`s MN03 lamp does put exactly 60 lumens. Think about it- the E2 light has always been rated at 60 "real out-of-the-reflector" lumens, but so is the new E2e. But the E2e uses a Pyrex lens with anti reflective coating that has some 97% efficient light transmission. The old E2 uses Lexan and has only 85% efficient light transmission. Both use the same lamp and probably the same reflector so which one is the 60 lumen output? I think probably 60Lu is only an approximation.

A friend of mine has access to a big proper Integrating Sphere that has recently been calibrated to Japanese industry standards. He said I could send a small amount of lights to him for measurment, so possibly I can get a few of my Surefires and a specific brand MR16 lamp properly lumen-measured, we could work from there then (assuming you had the same SureFires and batteries). No guarantee that I`ll be able to have this done any time soon, he is a very busy man and acess to the Sphere is probably controlled too.

Definately interested in your results though. If we could get this to work, approximate "home" lumen measurements could be taken, if only for ball-park figures.

I have plans here to build an integrating sphere. I`ll say no more at present but if it all goes to plan, I could build my own for more accurate measurements

First, congrats on the web page contest.
Wow, Chris M, a most excellent response. Even though it just boils down to GTB! I was hoping for some sort of conversion formula but oh well.

Thanks

I agree it is a shame that there is no easy comparison between the two, but that`s just the way things are- it`s kind of like equating gallons to feet or something. Photometrics is a complicated subject indeed. It`s enough to make your head spin

But you can never have enough torches so I`m all for an excuse to buy more!

#### hotfoot

##### Flashlight Enthusiast
With regards to an integrating sphere, maybe we can wing it. If a sphere can be set up with ordinary white paint, or perhaps completely mirrored/chromed inside, an initial measurement can be first taken with a reference light of known and reliable output, then the deviation noted and used to correct readings against other lights - ya know, kinda slide rule style. I'm just guessing here, so please ignore this post if you recognize it as rubbish

#### Chris M.

##### Flashlight Enthusiast
Mirrored, no. White paint, maybe- but it would have to be matte finish and very even. All`s I can think of is white primer spray- that has good bright-white matte properties, or at least the stuff I last had did.

The big troubles after that are:
- finding a large hollow sphere- or better, two hemispheres that can be bolted/hinged together. Minimum reccomended diameter is 10 times the diameter of the size of the lamp- in this case it would be at least 10 times the size of the lens of largest torch you plan to be able to measure. It needs to be openable for painting, cleaning (dust, etc) and to install an internal deflector that stops light falling on the sensor directly from the incoming lamp (all light on the sensor must be reflected from the surface of the sphere, none directly or it ruins the result)
- finding a suitable sensitive and linear light sensor. Maybe a Lux meter would be OK, it must have a remote sensor though (one built into the top or front would be impractical), and probably not be too big of a sensor.
- finding an adjustable iris for the "incoming" window through which you shine your torch. Different torches have different diameter heads of coruse so you need a variable iris that can be sized to fit the torch, stopping all external light getting in. Ideally white coloured on the inner surface but black apparently can work, it`s all down to the calibration.

Eventually I`ll try and build one (hopefully with Barium Sulphate paint). Anyone here who fancies a go, please feel free but I cannot guarantee the success or accuracy of anything you try. If it sort-of-works, it`s a step in the right direction at least....

#### brightnorm

##### Flashaholic
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Chris M.:
..... But the E2e uses a Pyrex lens with anti reflective coating that has some 97% efficient light transmission. The old E2 uses Lexan and has only 85% efficient light transmission....<HR></BLOCKQUOTE>

Chris M,

I am very surprised to see that the pyrex lens passes 14% more light than the lexan. I thought it was more like 8%.

Perhaps it is the antireflective coating that makes light transmission so efficient. Since many plastic prescription lenses have such a coating why not apply it to Lexan? It should raise costs nominally, and I would think that Lexan lenses might be an advantage during very rough usage or LEO action where even Pyrex could be vulnerable.

Any info on this?

Brightnorm

D

#### **DONOTDELETE**

##### Guest
I could try to check whether it is possible to convert between the candlepower and lumen readings using that ugly photometric math Chris referred to.

All I need is the following data for any of the surefire lamps:
- lumen output of the lamp
- lux readings at the hotspot of the beam at the distance of 1 and 2 meters
- diameter of the reflector
- diameter of the beam at the distance of 2 meters (ideally, such "diameter" would correspond to the lux reading being 50% of the maximum, but an eyeballed value would suffice). Is such data avilable anywhere? If not, does anybody out there have a Surefire AND a lux-meter?

If the computed lumen value lies in the neighborhood of the actual known lumen value, it may be possible to implement a lumens<->candlepowers converter (as a Java applet, for example) that would perform the conversion based on the specified beam profile.

D

#### **DONOTDELETE**

##### Guest
I wish someone would produce a cheap (mebbe coupla hundred bucks?) integrating sphere. Exact accuracy not necessary, mebbe just within a +/- 10% margin of error?

I'd even go in with a few someones locally and then we'd have a reference machine -- and everybody could send us their beloved lights temporarily and we'd measure'em all and have a handy-dandy reference chart from which to decide on our purchases. In fact I wouldn't mind sending in \$10 or \$20 bucks to one of the CPF administrators, and if we had enough participation, we could have our very own CPF IS (Integrating Sphere) which we could send our addiction items for testing.

Whatsay, a Group Buy of one IS? Anybody?

#### Chris M.

##### Flashlight Enthusiast
I wish someone would produce a cheap (mebbe coupla hundred bucks?) integrating sphere. Exact accuracy not necessary, mebbe just within a +/- 10% margin of error?

If only it were that simple. The specialised Barium Sulphate paint (extremely pure white) that is used to coat the inside of integrating spheres costs some £400 per litre (US\$600) and it needs re-painting every couple of years as it degrades over time. It also needs constant re-caliabration against a reference light source, and such precision-made lamps and their associated power supplies cost big bucks too. Every year, the sensor and display/measuring equipment will need proffessionally recalibrating too- \$\$\$

Got \$1000 spare? If about a hundred of us got together with that amount each we could probably buy a properly constructed "CPF Sphere" and have some put aside for maintenece and calibration. Unfortunately it doesn`t seem practical. They can also be quite large.

DIY Lumen measuring is currently not too simple, which is a shame.

#### rlhess

##### Enlightened
Hi, Jonathan,

That took time!

The integrating sphere is actually calibrated, as I understand it, from a standard lamp which is calibrated using a photometric goniometer or some such named device.

One needs to keep the sensor perpendicular to the light source to eliminate the need for reverse cosine correction.

It is convenient if the sensor moves in a sphere and not a plane, then you don't have to do inverse-square law distances.

I think we're in agreement here.

Great summary!

Cheers,

Richard

#### Jonathan

##### Enlightened
Hmm. I suspect that if one uses the integrating ceiling approach, that you explicitly _wouldn't_ need to bother with inverse cosine correction nor inverse square law correction, and furthermore, you would want the sensor in the plane of the ceiling.

My reasoning is that as you move further off axis, the solid angle covered by a unit area patch of the ceiling will get smaller and smaller. This will be precisely matched by the error introduced by having the sensor further away as well as off axis.

The net result will be that the off axis lux measurement will no longer correspond to the candela in that direction, but that the _total_ lumen number that you calculate will be correct. (Well, for the total lumen going toward the wall.)

So if your goal is to figure out the spatial distribution of candela, then you will need to point the sensor at the light source and correct for inverse square law...and it would be easier to simply move the sensor about on a spherical surface, but if you just want total lumen output, then the 'integrating ceiling' will work with pretty simple maths.

-Jon

#### rlhess

##### Enlightened
Jonathan,

As I recall the setup described (by Mr. Bulk (tm) I believe) for the integrating ceiling, you aim the light at the ceiling and you measure the light falling off the ceiling with an illuminance meter at the light location (but no direct light shining on the meter).

Could you please try one more time to explain how the rays at 0 degrees and 20 degrees (for example) become integrated at the meter.

I'm not visualizing what you are--and since you're in Boston, I'm starting to worry that you're from MIT and I'm just not getting it.

Richard