Old Silver Mag 2C Deep SMO SST-50 thrower fully regulated

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
(White wall comparison beamshots in post #7)

I had a new-in-box old style Silver Mag 2C that I was saving to make a special light with. This is the beginning of that build log.

Very pleased with the SST-50 thrower that I made a short while ago, I decided to base this project on that one with a few improvements. I really like the 2C form factor better than 3C, the silver colored Mag better than black ('cause scratches aren't as obvious on Silver and it's a rare color for a 2C these days), neutral white LEDs better than cool white, and I'd like to try to keep the dome this time (even if losing it did help improve throw). Other than those differences, it will essentially be the same torch, unless current plans change.

I'll be using two DX 20330 2.8A 8.4V buck drivers in parallel again, and I'll be using two AW IMR26500 batteries that already fit perfectly in the old style Mag 2C with no honing required. That unfortunately is exactly where the old style C making things easer ends.

The deep Mag Rebel reflector, which fit's fine in a modern Mag C with it's lower shelf and lack of head-stop, would not fit in the old C head without removing a significant amount of metal from the head. The head-stop had to go completely and the edge of the shelf that remained had to be beveled to allow the reflector to sit all the way down so the bezel could be tightened completely. You can see in the pics that I removed a small layer of plastic from the outside of the reflector cup before I found a grinding bit long enough to let me bevel the edge of the shelf. The bottom of the 40mm deep reflector sits way down in the head, so the heatsink will have to be recessed into the neck.

FittingMagRebelDreflectorintoOldC.jpg




The brass pill from the DX reflector that I used last time was too fat to fit in the neck of the modern Mag C, so I had to hone the inside of the neck of that last torch to make it fit. The Old C has a much wider internal diameter and this time the brass pill is much too small. I considered shimming it with a sleeve made from duct-work Aluminum sheet metal, but I wasn't in love with the idea. I wanted to try making a heatsink from 1" copper tubing, but my go-to hardware giant only sells 10 foot sections which cost a fortune. Last week I came upon some copper bushings in Sears Hardware and hatched a new plan. The pic below shows how I cut the top from a 3/4" end cap, attached it to the top of a 3/4" bushing and nested it inside a 1" bushing. I cut down the 1" bushing first, then used a propane torch to silver-solder it all the parts together into a single piece. Finally, I cut it down the 3/4" bushing flush. I must have gone through about a dozen cut-off discs.

Makingtheheatsink.jpg


Being difficult again, the old 2C's much larger neck still wasn't large enough to accommodate my home-made without-a-lathe heatsink. I took out the brake hone again and honed the inside of the neck until my Cu heatsink fit.

Fittingtheheatsink.jpg


The top of the neck above the threads was razor thin after honing so I filed that small portion off completely. And, yes... that is the hammer I'll use to "tap" the heatsink into place after the drivers and switch are installed. The heasink fits very snughly.

That brings me to the other complication of the Old C, vs New C... The Old C switch can only be removed from the top of the torch. It cannot exit through the battery tube. My heatsink is so snug in the top of the neck that pushing it out from beneath will crush the drivers with the switch. I'll tap some threaded holes to mount a star with. Hopefully, I'll be able to use the same screws to pry the HS out the top if necessary. Copper is very soft though. #3-48 threads might strip if I try to pry the HS out with them. Here's hoping it never comes to that.

I haven't drilled any holes for screws or wires yet, 'cause the metal work alone has already taken several days. I want the emitter centering to be perfect, so I'll tap the holes at the outside edges of the star. The focus point of the reflector i just below where the cam is cut off so no pedestal is needed.

I have a 4500K GH-400 SST-50, WJ-N3 tint emitter that I'll probably use in this torch, after I solder it to a 20mm XRE star still on its way from DX. I might need to mount that emitter on a smaller board if it turns out that the UniqueFire HS-802 that I also have on the way will not fit a 20mm star. If that happens, I'll solder the premium tint emitter I had reserved for this torch onto a smaller round star for the HS-802 and order an SSR to use in this project. I'd really like to finish this project before modding the HS-802, but I'd also like to delay buying another Phatlight emitter for a while and don't really care for hot-plate soldering emitters to stars. I'm trying to avoid doing it any more than necessary, so if that means using the bare emitter for the HS-802 and waiting to buy an SSR-50 on star to complete this projec, then that's what I'll do.
 
Last edited:

vestureofblood

Flashlight Enthusiast
Joined
Jun 19, 2008
Messages
3,210
Location
Missouri
Re: Build Log: Old Mag 2C Deep SMO lathe-less heatsink SSR-50 SST-50

Looks like a good start. I enjoy making my own heat sinks as well when possible. I have used both copper and brass, and I like using them because if I need to I can not only solder parts together, they can be used for led + or -.

Thanks for sharing, and always bringing plenty of pictures to the table.
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: Build Log: Old Mag 2C Deep SMO lathe-less heatsink SSR-50 SST-50

Here's some update shots:

I painstakingly assembled the lowest profile possible dual DX 20330 driver stack. As you can see in the pics, I used vias to connect the +in and-in of each board, then replaced the supplied LED +/- wires with 20awg stranded Cu. I even made depressions in the bottom of the heatsink and top of the switch, and routed the switch+ beneath the switch top to make sure the stack sat perfectly flat between switch top and HS bottom. Not a mm was wasted.
dualdriverassembly.jpg


The result was this beautifully compressed assembly.
fullyassembleddriverstackHSstarswit.jpg


Well, the best laid plans of mice and men often go astray as they say. I tested everything with a DMM, made sure there were no shorts, hooked up the batteries and... nothing.:hairpull:

I spent the next 90 minutes or so double checking everything. Finally I unassembled the entire stack you see above and tested each driver individually only to discover that neither worked. I'm not sure what happened. Maybe the three vias grouped together that I used to connect +in to +in shouldn't have been connected to each other. Maybe because the two 20330 drivers were from different orders and had slightly different markings on the resistors they couldn't be paired together. Maybe the new one that I just received was a dud and took the slightly older one down with it.

Whatever the cause, the hobby gods were definitely not cooperating with me today and clearly demanded sacrifice to get this torch working.

Prior to the driver stack failure, I had a drill bit break off in the heatsink while making the screw holes. In the process of getting that chunk of bit out, I proceeded to break another two bits and a jeweler screwdriver that I've had my whole life. I made lemons out of lemonade by using the giant hole that resulted from this surgery to route both LED + and - through the same hole. That actually worked out well for the screw placement and wire routing. At least it looks like I did that on purpose.

starinstalledonrecessedHS.jpg



The saga continued when I replaced the two DX 20330 drivers with a single KD 3A 5-mode driver I had just received. I hooked it up exactly as prescribed and... Nothing! Again! I tested everything and clearly, this driver is DOA. That particular one has a bad rep for overheating, but I had a 100% success rate with the one I used before. Now I'm 50/50.

Today was a bad day for cheap Chinese drivers, a staple in my projects.

Not giving up and insisting to see this torch work today, I installed two 8xAMC7135 2.8A buck regulators in parallel. That also meant replacing the two IMR26500 2.3AH 3.7V batteries with two custom C-sized 2AH 3.0V LiFePO4 batteries that I bought from Download. They were prototypes and I got the only four of them ever sold. The batteries sag under load more than typical A123 26650s, but they fit perfectly in 2C torch and have considerably more capacity than LiFe 18650. Current provided to the emitter starts well above 5A on a fresh charge, but shortly into the discharge cycle, the battery resistance and regulators work together to provide a steady 5.0A. I couldn't have planned on that, but I'll take it.

Thank God for the good old reliable AMC7135 chips. I'm not thrilled about the reduction in runtime from 1 hour as I had orginally planned to less than 30 minutes, but considering that most of my torches are toys that barely get used for more than minutes at a time, I wont notice.

Here are some pics of the "finished" product.

businessend.jpg


2CSST-50deepreflector.jpg


The 4500K GH-400 SST-50, WJ-N3 tint emitter is surprisingly cool in color when pushed to maximum output. Unlike the 4500K WM-P4 SST-90 that I have which is much warmer, this SST-50 compares much more closely to a cool white than a neutral white. It's visibly cooler than my neutral white 4A tint MC-E torches and only slightly warmer than a CREE XR-E Q5 WC cool white emitter.

Beam shots will eventually follow.
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: Build Log: Old Mag 2C Deep SMO lathe-less heatsink SSR-50 SST-50

I played with this a bit more before going to bed this morning and realized that one of those LiFePO4 C-cells and an NiMH C-cell combined also push 5.2+A to the emitter. I'm considering feeding this torch that way to ease the load on the regulators and also to free up 25% of the worlds stock of these LiFe C-cells for another project.

I also noticed that three NiMH C-cells run through the DMM yields ~2.8A output which is obviously not nearly as bright, but the color tint at that current is much more comparable to the neutral white MC-E. Now I'm wondering if 4A looks as bright but less cool. Next time...
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A

Here are some white wall beamshots from my Luminus + Mag Rebel reflector creations compared to more traditional P7 and MC-E in 53mm MOP reflectors.

P7.jpg


MC-E4A.jpg


SST-50WJ-N3.jpg
<<<The torch this thead is about

SST-90WM-F3.jpg


SSR-50WJ-EH.jpg
 
Last edited:

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

I've been playing with this torch a little more each day. I just took a current measurement at the tailcap and measured 5.6A. I'm sure I only measured 5.2A the first time, although this time the NiMH C cell has been fully chraged. Maybe that accounts for the difference. Maybe burn-in had something to do with it.

It wasn't really my goal to overdrive the SST-50 this way. Luckily, it maintains brightness and color after several minutes of continuous use at this level, so the regulators and the heatsink are doing their jobs.

I mentioned before that the tint was cooler than I had anticipated, although not bluish. I've recently noticed that the instant-on tint is slightly warmer than one second later, but then the tint remains constant. Strobing the light reproduces the same effect with every press of the switch regardless of how hot the torch is or how long I've run it for. If heat alone was causing the tint shift then I doubt that it would cool instantaneously between strobes when the torch is hot. That leads me to suspect that the regulators and the emitter ramp-up to full current within the first second of powering on. (Maybe the batteries sagging under the intial load is contributing to the less than instantaneous ramp-up.) If my suspicions about that are correct, then I think that the ideal current to produce the slightly warmer tint that I prefer would be between 4-5A instead of 5+A.

Translation: There's just one more reason that I would have been happier with the two DX 20330 drivers in parallel (besides the runtime and more common battery options.)

It was such a pain in the butt to get the focus just perfect on this one torch that I don't think I'll be taking it apart again to swap the regulators for drivers. Maybe I'll build another one that way and convert this one from reflector to aspheric...
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

I'm an obsessed perfectionist. Here's a rapidfire photoshoot of the startup sequence of this torch:










What should I do? I'm tempted to order two new DX 20330 drivers to carry this out as I had planned it, that is unless there's some way to put additional AMC7135 based regulators in the tailcap so I don't have to take everything apart again. I don't think that can be done though, 'cause the regulators have three connections (a common positive, a negative input and a negative LED). I'm driving myself nuts on this one. :hairpull:

Screw it, I'm ordering the DX drivers :crazy:
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

:hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull:
"Item is temporarily sold out"
:hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull::hairpull:
I've never seen any item return from that condition back to in-stock again.
:crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull: :crazy: :hairpull:

Here's hoping that sku 20329 with the 5-mode PCB removed is the exact same thing...
 
Last edited:

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

:ohgeez:sku 20330 has changed status again, this time from "temporarily sold out" back to being available once again for purchse. I've never seen that happen before. Now I have two of those on the way too. I'm not cancelling the order for the two 20329s, just in case something else changes. I'm just going to keep throwing money at this until it works the way I want it to. (Don't tell my wife I said that.)
 

ma_sha1

Flashlight Enthusiast
Joined
Jan 10, 2009
Messages
3,042
Location
CT, USA
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

Tech,

That's a lot of hair you are pulling out there :crackup:

"LiFePO4 C-cells + NiMH C-cell combined also push 5.2+A to the emitter"
Are you mixing chemistry here? Is this safe? You were concerned with my DD 14250 previously. Do you have any reference that others' have done it this way?

My order of 20330 haven't shipped yet (ordered on the 9th), now it says
expect 3/31/ in stock date.


Also, I wonder if you could wire 20330 with 20329 together in parallel & get 5 AMP with 5 modes? The description looks similar,
That's going to be awsome if it works.





:ohgeez:sku 20330 has changed status again, this time from "temporarily sold out" back to being available once again for purchse. I've never seen that happen before. Now I have two of those on the way too. I'm not cancelling the order for the two 20329s, just in case something else changes. I'm just going to keep throwing money at this until it works the way I want it to. (Don't tell my wife I said that.)
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

Tech,

That's a lot of hair you are pulling out there :crackup:

"LiFePO4 C-cells + NiMH C-cell combined also push 5.2+A to the emitter"
Are you mixing chemistry here? Is this safe? You were concerned with my DD 14250 previously. Do you have any reference that others' have done it this way?

My order of 20330 haven't shipped yet (ordered on the 9th), now it says
expect 3/31/ in stock date.


Also, I wonder if you could wire 20330 with 20329 together in parallel & get 5 AMP with 5 modes? The description looks similar,
That's going to be awsome if it works.

Ma_Sha1,

Mixing chemistries is always a no-no, however, I've written myself a pass to do this for the following reasons:
1) I'm not running any cells in parallel, just one of each in series, so neither should leech power from the other
2) both are "safe chemistry" cells
3) both can candle the 5A load easily without internal resistance causing heat
4) both are approximately the same capacity and so will discharge at approximately the same rate
5) I'm a fanatic about topping off half spent cells
6) There's another CPFer who did quite a few builds several years ago with the same "safe mixed chemistry". I forget who, but IIRC, he used to use a LiFePO4 18650 and an NiMH AA of the same capacity and never reported any issues. I'm not pioneering this, I'm copying his formula. (I'm not recommending it either, by the way.)

I was concerned for your safety with the 14250 cell because it was unprotected Lithium Cobalt (unsafe Li) chemistry and the XPG emitter would pull beyond the usual safe 2C limit for LiCo.

If the 20329 are both the same beneath the mode-module on top, then I suppose you could remove the mode module from one and pair up the two bottom halves beneath the single top half. Whether or not the mode module can handle the extra amps coursing through it remains to be seen. (It probably comes down to which FET is used on the mode board.)

After my experience with trying to mix two 20330s that were nearly identical but not completely (resistor values, markings on one chip) and having both boards fail, I wouldn't recommend mixing a 20330 and a 20329 unless you removed the mode board from the 20329 first and confirmed that everything beneath was exactly the same as on the 20330.

If the mode board could be reprogrammed for just hi/low, and if there were room in this build (clearly, there isn't), then I'd consider doing that too. A Hi/Lo 5A driver would be sweet, especially for ~$10.
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

Well, while I wait for the DX drivers to come in, I've found an interim solution that works pretty well. Through both LiFe cells and one of these resistors, I measured a 4.2A draw. I put one in the (de-ano'd) tailcap in place of the spring. Now I can use the torch for a while before I rebuild it.

interimsolution.jpg


At 4.2A current and a ~1.5-2V drop across the resistor, I'm running it way above the 5W spec, so this is obviously not a permanent solution, just something to play with.
 
Last edited:

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

Two DX 20330 boards came in today. Instead of using the vias to connect the inputs, I wired around the boards (there was just enough room without even having to file the edges). A clamp meter between the drivers and the emitter measured a perfect 5.0A.

This torch is finally done! Yay!

I also finished the 4.3A driven 1C version with the same reflector, same bin and tint SST-50. The difference between driving the 4500K SST-50 at 4.3A and 5.0A appears to be only the tint and the heat produced. 5.0A being hotter temperature-wise and cooler tint-wise. Other than the color and the heat generated, output appears exactly the same.

I smell beamshots and an 8-way shootout coming this weekend...

Contentders:
5 SST-50 torches, an SST-90 torch, an MC-E torch, a P7 torch and a 5*XR-E torch.
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
Turns out that I should have filed the edges after all... After getting good and hot, it developed a very slight short between the drivers and the emitter. :ohgeez:Luckily, everything survived the incident, which just resulted in some smoke from where the insulation smoldered.

I've since filed a flat edge onto each driver for the wires to pass around and also replaced all the wires. While I was at it, I changed the sense resistors on the drivers to bring the output down to 4.1A, verified with my clamp meter.

It's interesting to note that the tint of this 4500K emitter is cooler when driven only to 4.1A than it was when driven to 4.2A by the tailcap resistor. I'm certain of this, because as a reference, I have been comparing it to its twin driven to 4.3A by 8135 regulators in the cutdown Mag 1C with the same reflector.

I wonder if the cooler tint has to do with the way the DX buck drivers work. If they're not really performing DC-DC conversion, but just reducing mean current by PWM like a hot driver would do (in which case, they really shouldn't be used for LEDs), then that might explain this phenomenon.

Disregard everything in silver italic font above. I didn't realize that the batteries were partially spent and not driving the resistored or 8135 regulated configurations to top current draw. On a fresh charge at 4.3A draw, the 8135 regulated shorty is a dead match for the 4.1A two cell silver torch. I'm very, very happy with both of these.

IOU one set of beamshots...
 
Last edited:

ma_sha1

Flashlight Enthusiast
Joined
Jan 10, 2009
Messages
3,042
Location
CT, USA
Re: (Beamshots post7) Mag 2C Deep SMO SST-50 @5A+

Tech,

Interesting set-up. I have two of these RadioShack resistors & was thinking about putting it with 2x LiFeO4 to drop the volt & driver SST-90,
just worried that it's not going to last.

However, I am finding that there's Metal High Power resistor can handle up to 20W, I am wonder if 0.5 Ohm could drop 2v on the 6.6V set-up with SST-90.

I might hook one up & see if I can get 8 Amp. As I am only getting 4 Amp with 1x IMR 26650 on SSR-90. Would prefer 2xLiFeO4 with a drop.

Just wondering if you have tried on SSR/SST-90?
I think it might be better than using the Linear Board 8 Amp,
as a 20W metal resistor is simple, beefy & will last.

With 2V over resistor, it'll be 16W of heat at 8 Amp from the resistor alone :eek:. Considering roughly 4 v drop across the SST-90. it's around 60% "driver" efficiency.









Well, while I wait for the DX drivers to come in, I've found an interim solution that works pretty well. Through both LiFe cells and one of these resistors, I measured a 4.2A draw. I put one in the (de-ano'd) tailcap in place of the spring. Now I can use the torch for a while before I rebuild it.

interimsolution.jpg


At 4.2A current and a ~1.5-2V drop across the resistor, I'm running it way above the 5W spec, so this is obviously not a permanent solution, just something to play with.
 

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
I hated the idea of all that wasted power, using a 6V power supply for a 3.6V emitter and no DC-DC converter. For my 8.4A SST-90, I used a 4.8V supply that sags to around 4V under load. In that case, instead of resistors, I used regulators. Not a lot of power was wasted in the AMC8135 chips.

If you really wanted to drive an SST-50 from two LiFePO4 26650 with just these resistors, I think the formula would call for four of them wired as 2s2p to keep the resistor wattage in spec. For the SST-90, I think you'd need 9 as 3s3p. (Not bothering to really do the math here.) I guess you could do it in a 3D Mag, but who'd want to waste all that battery power and runtime on resistors?
 

ma_sha1

Flashlight Enthusiast
Joined
Jan 10, 2009
Messages
3,042
Location
CT, USA
I hated the idea of all that wasted power, using a 6V power supply for a 3.6V emitter and no DC-DC converter. For my 8.4A SST-90, I used a 4.8V supply that sags to around 4V under load. In that case, instead of resistors, I used regulators. Not a lot of power was wasted in the AMC8135 chips.

If you really wanted to drive an SST-50 from two LiFePO4 26650 with just these resistors, I think the formula would call for four of them wired as 2s2p to keep the resistor wattage in spec. For the SST-90, I think you'd need 9 as 3s3p. (Not bothering to really do the math here.) I guess you could do it in a 3D Mag, but who'd want to waste all that battery power and runtime on resistors?


Good point, I might be better off waiting for the 10A CC driver that sectorclear is making.
 
Top