My first 62138 build

[TigerhawkT3]

I just finished my first superbulb Mag build. It uses a 4D Mag, an Osram 62138 (USL bulb), aluminum reflector from KD, FM high-temp socket kit, a Boro window(eventually, when they arrive), a hand-soldered 12AA NiMH composed of Rayovac Hybrids, a pair of custom-"machined" battery end caps, and a squeezed tailspring. Enough talk; here are pics:

The following pics are of the battery endcaps. I got a 1.25" poplar dowel from OSH and used my abrasive-wheel chop saw to get two little rounds (that explains the darker, burned areas). I then dremeled the proper shapes to fit over the battery's ends, drilled a small hole through the proper cell-holding location, inserted a wire through it, and put huge solder gobs on both ends. The result is a little rickety, but works fine.

The top- and bottom-left are of the +ve (near the head) end's flashlight and battery contacts, respectively. The top- and bottom-right are of the -ve (near the tail) end's flashlight and battery contacts, respectively.




I built two hosts, so I may buy four D Li-Ions from KD for an easy drop-in battery solution. The beam is a bit swirly because of the mirror-smooth reflector, but that's okay. I haven't even given the cells an initial charge, and they're performing all right anyway! Output should improve after I top them off.

 

[TigerhawkT3]

The Boros came in today! Hooray.

 

[jimjones3630]

Like those creative battery caps. Nice mod

 

[cmacclel]

Hmm I thought 12 cells on the 62138 = Poof Instaflash........even 11 cells need to be rested for 1+ hours or it will instaflash.

Mac

 

[jimjones3630]

My 62138 running now on 3x Emoli=12.6v, have ran on 4xA123=14.4v. Flash 62138 on bench at 15v, maybe this batch more hearty.

 

[cmacclel]

My 62138 running now on 3x Emoli=12.6v, have ran on 4xA123=14.4v. Flash 62138 on bench at 15v, maybe this batch more hearty.

Nimh's fully charged can top 1.5v

1.5v x 12 = 18v

Mac

 

[jimjones3630]

interesting to see his vbat, vbulb.

 

[TigerhawkT3]

When I first built the batt, it measured 15.5V OC. As pictured, these are LSD-type Rayovac Hybrids, which I purchased a few months ago. They didn't instaflash the 62138 when I first tried them, nor after I charged them with a 5-10 cell pack charger (6 cells at a time). I didn't think to measure the OC voltage right after charging, but I haven't had an instaflash yet. I believe it was modamag who told me (at one of his BBQs) that a 62138 could handle 12xNiMH easily, which I was slightly worried about.

If that was incorrect information, the issue might be that the Hybrids can't handle the current, and so aren't able to pop the bulb. If that's the case, using Eneloops (which Silverfox has had good results with at 8+ amps) or any other high-current NiMH would result in failure. If that's so, however, it would mean that I'm not getting full power from this build (horrors). Or, these 62138s may simply have been built stronger - that would be nice to discover. Alternatively, my build may have a lot of resistance.

All in all, I'm inclined to believe modamag's advice that 12 cells is fine, but if not, the next likely candidate is shoddy, high-resistance connections.

If you tell me how exactly to go about measuring Vbat and Vbulb, I'd be happy to do so.

I can't wait to take this sleeper for a spin!

 

[jimjones3630]

Fairly easy measure vbat with tailcap off and switch on DMM probe to batt end and other probe to exposed aluminum wall end of body.

vbulb place DMM probes to both bulb pins while plugged into socket, can pull bubl out partly allowing room for probes, or using wire loops one on each pin then plug into socket, DMM probes to each loop and turn on.

the difference shows how many volts getting from battery thru your circuit to bulb. If resistance high will show up here. Gives you an idea from one mod to another if built as efficently.

 

[TigerhawkT3]

Vbat fully charged was 16.0V

I learned that it's much easier to find Vbulb in a gutted Minimag (2.7, BTW) than in a superbulb Mag. The insulative handles on the test probes started smoking immediately (as in, within a fraction of a second) when the thing was turned on in close proximity to them. I didn't want to completely melt them, so I only took readings for a couple seconds, and the numbers jumped between 0, 4, and 8. It's hard to believe that resistance is that high; the low reading could be caused by intermittent contact of the probes... and I'm not really itchin' to try it again.

I'll work on the battery endcaps and test the firestarting time. How long should it be in an ideal setup?

The best thing would probably be to bring it to a get-together and have it looked at, and compare it to an actual USL.

EDIT: It pulls 8.1A at startup, and the beam is nice and white, if that helps the analysis.

 

[jimjones3630]

Somking probes! that gets expensive. measuring vbulb in a bi-pin touched pin and rim of bipin holder melting the metal tip. It's an education.

Fire starting time depends on vbat for that bulb, 62138. Did some timing in this thread, don't recall times off hand.

http://www.candlepowerforums.com/vb/showthread.php?t=162190

Whiteness can go by in bulbs that have enough experience. Thinking of 64623, once over 16v it loses some yellowish coloration looks whiter. 62138 is white at 12vbulb, 14v seems is as white to me but more volume of light, throwing further, more useful side spill. But not as much as 64625 which to me is more flood than throw.

Hope that helps
Jim

Vbat fully charged was 16.0V

I learned that it's much easier to find Vbulb in a gutted Minimag (2.7, BTW) than in a superbulb Mag. The insulative handles on the test probes started smoking immediately (as in, within a fraction of a second) when the thing was turned on in close proximity to them. I didn't want to completely melt them, so I only took readings for a couple seconds, and the numbers jumped between 0, 4, and 8. It's hard to believe that resistance is that high; the low reading could be caused by intermittent contact of the probes... and I'm not really itchin' to try it again.

I'll work on the battery endcaps and test the firestarting time. How long should it be in an ideal setup?

The best thing would probably be to bring it to a get-together and have it looked at, and compare it to an actual USL.

EDIT: It pulls 8.1A at startup, and the beam is nice and white, if that helps the analysis.

 

[TigerhawkT3]

Firestarting time was ~2.5min. I have the feeling I'm not doing it right.

The Boro didn't like it.

 

[jimjones3630]

Oooh, first seen do that. I broiled a reflector if that is any consulation.
The type of paper makes a difference. Newpaper inserts with colors kind of slick feeling or Hershey wrappers catch quickly.

Firestarting time was ~2.5min. I have the feeling I'm not doing it right.

The Boro didn't like it.

 

[TigerhawkT3]

Took it out with my Sam's 35W HID for Halloween night. I hadn't charged it since the 2.5min firestart, so it seemed a bit dim. Ceiling-bounce lux readings went up by about 50% after a full charge, so it was basically at 2/3rds power.

What I like about this mod is the fact that I actually chose the host body before the guts, which means that I have a usable light. I usually get caught up messing with the internals and have a pile of wiring and connections showing. This build may not be as powerful as it could be due to the high resistance, but I'm happy with it.

 

[jimjones3630]

Yeah, that broke window speaks volumes.