# 64625 vs 62138 Osram lamps

#### andrewwynn

##### Flashlight Enthusiast
Being fascinated by high-output lamps, and the Mag100 project (kicked off by Tom (NikolaTesla) and his desire to have a home-brew USL-like light and starting with a 100W westinghouse lamp that we could direct-drive from 12AA cells.. things naturally progressed to the osram 64625 lamp.

It has been quite a bit of a challenge to get this little puppy to run, but found a magic combination that works brilliantly well..

11xGP2000 cells and the hotdriver set to 12.5V for a nice round 2700L out the front and 94% efficiency.

I finally figured out why my spreadsheet numbers just weren't calculating just right.. by setting the regulator to 12.0V and actually measuring the wattage of the lamps rather than assuming when they say they are 100W they mean it..

The 625 lamp pulls 9.25A at 12V.. or 111W.. the 138 lamp pulls 8.76A at 12V or 105.1W..

now all the numbers are working out with the formulas properly.. It was bugging me that my math never worked out, but it was the BASELINE that was off..

I run the 138 at 13.2V and it always would measure in the 9.2A+ range not the 8.78A value that the formula was providing.. now the formula predicts 9.23A which is almost exactly what is measured.. all is right in the world.

So.. back on topic.. the comparison:

Doing a ceiling-bounce test confirms they

1) are very very VERY bright, oMG.. and
2) very close to the same output. what's a couple hundred lumen?

Beamshot:

The 625 on the left and 138 on the right.

The picture does NOT do justice to the fact you are looking at over 5000L of light bouncing off that wall.. it's pretty much like looking at the headlights of an oncoming bus!

The rest of the beams can be found here and here.

Looking at the comparisons..
Base output of 520more lumen out the front and 200 higher CCT really had me leaning in favor toward the 625, and the driving force to try out the lamp! Not to mention the extra 5.8L/W efficacy.

Since they are rated both for 50 hr lifespan, i am very surprised that we can barely get away with 16 and 7% overdrive.

So.. after blowing up a few bulbs we settled in on 12.5 or 12.6 being very happymaking for the 625..

The Vbat and LDO figures represent the voltage required at the battery to get the Vbulb to the lamp and the LDO is the voltage drop on the regulator FET to keep things happy.

Notice that both lamps run at about 120W ± 2W.. amazing.

Notice that the overall output is slightly higher on the 625, but that the CCT is 146 higher and definitely noticeable.

what makes me like the 625 more than the 138... is that the spot doesn't have that tightly defined circle that you can't eliminate even with a stippled reflector in the 138.. you get a big monster spot of just about 50,000 lux of light.. so has about the throw of a typical Mag85 but with 3-4x the spread!

As shown in the beamshots on the Mag625 thread.. the throw is good for a half a block EASY. Basically.. you can't see things w/o binoculars as far as the lamp will illuminate.. so you really don't need more spot than it has, so it's the best of both worlds and my new favorite lamp for sure.

The 138 has been running a good long time.. i've never blown one, though mine is frosted something fierce inside.. it's milky-white.

Time will tell how close the hour life figures work out, but i'm kinda liking the idea of double the bulb life... me thinks i'll be very happy with my Mag625.

I've decided that it only makes sense to use the similar nomenclature we all use for the likes of the "mag85'.. but to avoid possible confusion with the WA lamps and point out the lamp is different.. using 3 digits instead..

So the Mag100-138 is now called the "Mag138" and the Mag100-625 is not called the "Mag625".

When i figure out how to shoehorn 11 x 4/5A cells into a smaller light than a 3D host which is a nobrainer-fit btw.. fits great.. then i'd use my 'shrinkydink' nomenclature like the M85 or the M66 and call it the M625 or the M138.. or possibly the M120 (watts).. using either the 138 or the 625 lamp..

Ok.. snubbed the other major problem i had with the 625 and have #2 running, and more and more i'm in-favor of the 625's bigger, bolder beam. Spotting 500' is fun once.. but lighting up everything from here to 300' away is far more useful.

beamshot comparison: 250' to a building

Here is the 62138 running at 13.2

Here is the 64625 running at 12.8..

A little too hot to handle.. the bulb did not last too long, but i think that was due to another problem, so i will re-investigate running that hot once i get some more time on the lamp/host. It's very nice to have a WORKING light!

Tom, myself and many others are beginning to realize that that softer focus, flood-like beam is much more useful in most cases, and i would love to see some feedback what people think.

Realize, that the 625 with a stippled reflector still has more LUX than a Mag85! and nearly 4x the light output to boot! So consider that with your feedback.

-awr

#### LEDcandle

##### Flashlight Enthusiast
Wow Andrew!!

Very nice comparo information. These babies rock! Which bulb is used in the USL? 62138? Will they be changed to the 625?

Any Singaporean buying a USL? I think KevinL?
I would like to shine one of these into my litebox to see what kinda figures it's giving out

As for flooder vs thrower, even though the 625 looks like it collimates less light, I'm sure by itself it already throws like hell (as you said, 50,000 lux big hotspot!!). So I would say it might be the more useful of the two.

Last edited:

#### andrewwynn

##### Flashlight Enthusiast
The USL has the 138.. it's a beautiful lamp.. can't run in the USL or any other light w/o regulation.. not possible at all.. they can't handle the over-voltage.

Well you could maybe run it from 10 cells but what fun is not over-driving a light?

We measured just under 50,000 lux with the 625 and about 54,000 lux with the same exact reflector setup with the 138 lamp.

(we have measured over 80,000 lux with a smooth reflector and the 138 though!)..

-awr

#### Flakey

##### Enlightened
i dont think that there is any way for a USL to drive the 625 lamp. the usl is a true hotwire that direct drives the bulb, the 625 needs 12.5V or less and on start up the USL would no doubt flash it.

#### andrewwynn

##### Flashlight Enthusiast
yes the USL will flash an 625 bulb instantaneously.. i tried to run it regulated at 13.2 and it blew. I ran it at 12.88 and it blew on the second run of the cells.

-awr

this does not diminish the awesomeness of the USL, quite a milestone light!

#### Delvance

##### Enlightened
Awr,

Great work on the two insane [email protected]. Those two fieldshots tell more than a thosand words (or lumens in this case heh). I'm naturally a throw/LOP junkie at heart, but with multi thousand lumens firing out of a torch, i'd have to agree that the stippled reflectors turn those beams into true art. The fact that they can throw as far as a [email protected] is just plain amazing...As to 80,000lux vs 50,000lux...i really can't think of much need that even comes close to 50,000lux (for a HANDHELD portable light anyways). I'm sure many here would agree it's much more useful to have a flood beam than a spot beam when using the light...

I'm actually currently building a 100w+ project with a 4D host. Plans are 4D host, 12 cbps, and the Osram HLX 64623 on DD. I wonder how that will compare to your two beasts...

#### LEDcandle

##### Flashlight Enthusiast
Ah I see... I think I've been reading too much about the USL and Mag100R at the same time that I've kinda confused them all.

So the USL doesn't have the hotwire driver. Nonetheless :naughty:

80,000 lux in a mag body in insane! The Brightstar 24w HID with its bigger reflector is doing only *slightly* (well, relatively) better at 92,000 lux (variances noted). Impressive!

Delvance, the 64623 is *only* rated at 2800 lumens so it should perform similary to the 62138. But it has 2000 hrs of bulb life and might be able to take some overdriving (as in 14.4v from your setup) and churn out a lot more. But the T4 bulb size will require some opening of the reflector hole.

Let us know how it goes! A supermag shootout with all these lights would be amazing Would love to see some of bwaites 250w mods..hahah...

Last edited:

#### Delvance

##### Enlightened
Ledcandle,

Yup 2800 lumen bulb, although i think awr did some calculations on it in another thread. He came to 3200 torch lumens if the lamp can get 14.2V which is 12 cells DD, although i'm still trying to find another battery solution as those cbp 1650's are going to sag pretty well at 9A+. I read that Wilkey had tested the 623, and concluded that the lamp can take 16V or so but i'm trying to keep to 4D.

250W ???

#### andrewwynn

##### Flashlight Enthusiast
delvance.. sounds like a neat host.. which cells are you using? 1650s?

the 623 lamp can take a brutal punishment.. you can definitely direct-drive from 12 cells.

the 623 at 14.4 re-rates to over 3000L (300 more than my 625 solution) if you can hold 14.0V at the lamp (cells need to hold 1.2V/cell at TEN amps!

that combo is 140W btw..

If i was doing that mod i would put in 13-14 cells.. maybe even 15 if you could get away with it, but don't know if anybody can tri-bore a 4D light.

keep us posted on the outcome.. check out my battery holder solution at http://rouse.com/circuits if you need ideas for holding the cells.

The USL is a direct-drive light.. with extremely low resistance.. it actually outputs very similar light to my first Mag100, but does it with 11 cells vs 12.

one of the people on the hotdriver list is going to make a light with the 623.. with the regulator set really close to 14.4V.

In the 4-D host.. you could probably spin the thing and hone out the inside to get just the bit of extra you need to get 4-wide vs 3-wide of the 1650s..

=awr

#### Delvance

##### Enlightened
Thanks for the info Awr! Much appreciated.

I was planning to use the cbp 1650's but at 10amp draw, we all know they won't hold 1.2V . Even at 1.1 that only comes upto 13.2v and that's before going to bulb. Making the diameter bigger certainly sounds tempting, but i havn't been able to find anyone local that can perform it. Maybe i'll see if some of the known CPFer's that bore lights would be willing to bore a 4D and send it to me hrmm. If not, 5D with 15 cbp's maybe. In any case, when the thing is done, i'll definitely post it.

Back to thinking about possible configs!

#### andrewwynn

##### Flashlight Enthusiast
well there are bored bodies out there.. 3-bore and 4-bore. I would look into the possibilities.. 12 GP2000s will run the heck out of the 623.

#### NikolaTesla

##### Flashlight Enthusiast
Yep having been the guy holding the light and seeing that 625 light up that whole field like an X990 is amazing. The 11 cell / LDO st at 12.5 v is the trick to make it live though. The driver makes up for diminishing voltage holding brightness till dead but the battery setup is still important too. It took 2 of us 8 hours to jig up the super holder which Andrew has shown how to build it on his website. It is kind of a 4 handed deal to final assemble but once together is very low loss and stable enough to power that 625 and live without smoke ( Like my first attemp with stock spring loaded Modamag holders- they can not handle 9 amps). Either or is a choice you have to make. More light, whiter out of 625 or more controlled beam of 138 is very nice with #5 light stipple reflector- perfectly clean round beam like SureFire but way brighter. I am making another in a 4D with Li-Ion D cells so I have a 1/2 hour 100 watter which is nice run time for that kind of power. 9 amp light bulbs definitely are hungry for juice. All of this is doable now in several variations with Hot Driver which is easily adjusted. Insta-Flashing with hot batteries is a thing of the past. If you get/make one, you WILL be spoiled- I guarentee it.

Last edited:

#### bwaites

##### Flashlight Enthusiast
The USL can drive the '625 nicely, just not hot off the charger.

Leave it overnight, though, and it does fine.

I have one running with the 3" FiveMega head and it is awesome.

Bill

#### andrewwynn

##### Flashlight Enthusiast
More like 25 minutes of runtime, and as demontrated last night.. maybe 10 straight maximum before the over-heat sensor will kick in.. thanks to Bill (ala USL project) for putting the seed in my head to incorporate that! He is very correct about longer runtimes being a problem with heat. Haven't had a problem at all with either of the 625 lights, now they are finally working properly... the Big 4D model does get hotter than is comfortable to hold about the time it decides 'well that's just about hot enough' and starts to cycle off and on as it maintains the temperature.

-awr

#### andrewwynn

##### Flashlight Enthusiast
So... my buddy LarryK called me tonight with the suggestion: Put the 625 lamp on 3D LiON..

What a great solution..

The cells hold the 8.65A with cells holding 10.8V.. that gets 10.46 to the lamp with a KIU solution.. I was holding 8.70 and 10.73V at the lamp...

So.. it is just about the same CCT as the 138 lamp at 13.20.. looks very nice and white.. not as white as the 625 normally of course nor the 1185 or 1166.. but it runs no problem at the peak charged voltage of 3xLiON of 12.6 so no problems expected with insta-flashing a direct-drive solution.

Output of 2431/1580 lumen at 93.3W.. CCT of 3281.. but the neat part.. with under-driving the lamp so much.. bulb life re-rates to nearly 200 hours! talk about a nice combination.. nearly 1/2 hr runtime too!

Obviously the standard disclaimer of how to deal with LiON cells applies.. do not use them if you don't know what you are doing... and personally i would never run any series combination of LiON w/o a regulator that shuts down on low-voltage.

With the way these lamps are pushed.. if a single cells fails even on NiMH with 12 cells.. the regulator will not start up .. and if already running will shut off or 'blink' rapidly if a single cell is low.

-awr

#### Delvance

##### Enlightened
Ahhh interesting Awr! ~1500 lumen output is still nothing to sniff at, but seeing as you already had the 625 functioning at 2750 output regulated, safe to say you prefer the latter ? I certainly do! Not much point in trying to get extra bulb life out of what originally is short, may as well go for full ahead

And yep, those big D li-ions are scary! Although i've no doubt you know exactly! what you're doing with them. Still, be careful.

#### andrewwynn

##### Flashlight Enthusiast
oh yeah.. for me.. 11 pack in the 3D host for full blast is preferred.. but a total gimmie for a 3D vs 4D host using LIONs..

I don't run LiONs w/o the protection of the hotdriver.. it worked really well to shut down when one cell just was running lower than the rest in a 4D situation.. and i charge the cells parallel to make sure they are balanced charge.

-awr

#### OddBall

##### Newly Enlightened
Hmmm, do I need to get a dummy cell then? :thinking: I was gonna run 12 x AA nimh batts - whatever is going to fit and suit - but you recommend 11. :shrug:

"one of the people on the hotdriver list is going to make a light with the 623.. with the regulator set really close to 14.4V."

Andrew, if this isn't me, could it be? I want a high and a low-ish...

#### andrewwynn

##### Flashlight Enthusiast
it's listed in the chart actually. if you want to use thye 625 lamp it needs either 11 NIMH or 3 2/3 D LiON!

12x1650 cells will work great with the 138 lamp... possible that 12xgp2000 is still too much voltage for the 138.. but the high-temp ckt would probably just deal with it and it only means that you'd not be able to run 'full burns'.. it should work ok... it's not been tested yet though.

So far.. the successful combos:
4D with the 138
3D with the 625 (reduced output but 200 hr bulb life).. twice the output of mag85
3D with 11xGP2000 and 625
3D with 12xCBP1650 and 138

*ANY* use of 100W lamps in such a small host are a 'do at your own risk' situation.. they are a hazardous combination, to be used with extreme caution.. example: if one of these lights were to get turned on in a backback or on the floor of a car.. it's a virtual certanty to cause a fire within seconds! They are serious high-power devices, consider the caution you'd exibit with a firearm and you should be ok.

-awr

#### NewBie

##### *Retired*
Andrew, are you sure you didn't get something mixed up, or optimized the 62138 for close up focus? The 138 has a much hotter spot, but this does not show at all in your outdoor photos.

A re-processed photo of your wall shot, which clearly shows the 62138 with a much hotter spot, which should throw alot better...:

andrewwynn said:
Being fascinated by high-output lamps, and the Mag100 project (kicked off by Tom (NikolaTesla) and his desire to have a home-brew USL-like light and starting with a 100W westinghouse lamp that we could direct-drive from 12AA cells.. things naturally progressed to the osram 64625 lamp.

It has been quite a bit of a challenge to get this little puppy to run, but found a magic combination that works brilliantly well..

11xGP2000 cells and the hotdriver set to 12.5V for a nice round 2700L out the front and 94% efficiency.

I finally figured out why my spreadsheet numbers just weren't calculating just right.. by setting the regulator to 12.0V and actually measuring the wattage of the lamps rather than assuming when they say they are 100W they mean it..

The 625 lamp pulls 9.25A at 12V.. or 111W.. the 138 lamp pulls 8.76A at 12V or 105.1W..

now all the numbers are working out with the formulas properly.. It was bugging me that my math never worked out, but it was the BASELINE that was off..

I run the 138 at 13.2V and it always would measure in the 9.2A+ range not the 8.78A value that the formula was providing.. now the formula predicts 9.23A which is almost exactly what is measured.. all is right in the world.

So.. back on topic.. the comparison:

Doing a ceiling-bounce test confirms they

1) are very very VERY bright, oMG.. and
2) very close to the same output. what's a couple hundred lumen?

Beamshot:

The 625 on the left and 138 on the right.

The picture does NOT do justice to the fact you are looking at over 5000L of light bouncing off that wall.. it's pretty much like looking at the headlights of an oncoming bus!

The rest of the beams can be found here and here.

Looking at the comparisons..
Base output of 520more lumen out the front and 200 higher CCT really had me leaning in favor toward the 625, and the driving force to try out the lamp! Not to mention the extra 5.8L/W efficacy.

Since they are rated both for 50 hr lifespan, i am very surprised that we can barely get away with 16 and 7% overdrive.

So.. after blowing up a few bulbs we settled in on 12.5 or 12.6 being very happymaking for the 625..

The Vbat and LDO figures represent the voltage required at the battery to get the Vbulb to the lamp and the LDO is the voltage drop on the regulator FET to keep things happy.

Notice that both lamps run at about 120W ± 2W.. amazing.

Notice that the overall output is slightly higher on the 625, but that the CCT is 146 higher and definitely noticeable.

what makes me like the 625 more than the 138... is that the spot doesn't have that tightly defined circle that you can't eliminate even with a stippled reflector in the 138.. you get a big monster spot of just about 50,000 lux of light.. so has about the throw of a typical Mag85 but with 3-4x the spread!

As shown in the beamshots on the Mag625 thread.. the throw is good for a half a block EASY. Basically.. you can't see things w/o binoculars as far as the lamp will illuminate.. so you really don't need more spot than it has, so it's the best of both worlds and my new favorite lamp for sure.

The 138 has been running a good long time.. i've never blown one, though mine is frosted something fierce inside.. it's milky-white.

Time will tell how close the hour life figures work out, but i'm kinda liking the idea of double the bulb life... me thinks i'll be very happy with my Mag625.

I've decided that it only makes sense to use the similar nomenclature we all use for the likes of the "mag85'.. but to avoid possible confusion with the WA lamps and point out the lamp is different.. using 3 digits instead..

So the Mag100-138 is now called the "Mag138" and the Mag100-625 is not called the "Mag625".

When i figure out how to shoehorn 11 x 4/5A cells into a smaller light than a 3D host which is a nobrainer-fit btw.. fits great.. then i'd use my 'shrinkydink' nomenclature like the M85 or the M66 and call it the M625 or the M138.. or possibly the M120 (watts).. using either the 138 or the 625 lamp..

Ok.. snubbed the other major problem i had with the 625 and have #2 running, and more and more i'm in-favor of the 625's bigger, bolder beam. Spotting 500' is fun once.. but lighting up everything from here to 300' away is far more useful.

beamshot comparison: 250' to a building

Here is the 62138 running at 13.2

Here is the 64625 running at 12.8..

A little too hot to handle.. the bulb did not last too long, but i think that was due to another problem, so i will re-investigate running that hot once i get some more time on the lamp/host. It's very nice to have a WORKING light!

Tom, myself and many others are beginning to realize that that softer focus, flood-like beam is much more useful in most cases, and i would love to see some feedback what people think.

Realize, that the 625 with a stippled reflector still has more LUX than a Mag85! and nearly 4x the light output to boot! So consider that with your feedback.

-awr