Maglite Switch Modified for High Performance Bulbs

After modifying the reflector head and lamp holder for powerful halogen lamps, the third component to take care of is the switch unit. Heat is not that much of an issue in this case but the current rating of ~1A definitely needs some upgrading. What we are looking for is a permanent load of approximately 10A to cover every bulb up to 100W@12V and a soft-start function to increase the average life of our bulb and to avoid problems with powering up high performance types (30+lm/W). Furthermore, we want to maintain the waterproof (well, more or less) design and achieve a professional appearance.

The basic idea, Figure 1, is rather simple: To avoid overloading the electrical contacts the bulb current is switched by a MOSFET between battery pack and bulb. The MOSFET works like a voltage-controlled resistor. When a voltage is applied between Gate (G) and Source (S), the Gate-to-Drain resistance is low (lamp on), for zero control voltage it is high (lamp off). Now, the original switch is used only to control the MOSFET.


Figure 1: Electronic Circuit

The peripheral components R1, R2 and C1 provide the soft-start function: When S1 is closed the Gate voltage gradually increases and as a result the gate-to-drain current rises slowly, giving the filament some time to heat up.

The main advantage over other solutions, i.e. Willie Hunt's LRVs, is the low cost. However, in our case you don't get a voltage regulation. The electronics doesn't require much space and is easily installed on the original switch assembly (Figure 2) or on the battery pack (Figure 3). For the higher current bulbs (35W/6V in Figure 2, 100W/12V in Figure 3) I'm using two MOSFETs in parallel to improve power dissipation. Since in both cases we are using the original switch, no machining of the body is necessary and the modification remains invisible.


Figure 2: Electronics installed on Original Switch Assembly


Figure 3: Electronics installed on Battery Pack

In Figure 3 you can see how the electrical contacts from the lamp holder and the switch are soldered to the PCB.

That's all for now, folks.

Regards,
Snoopy

Re: Maglite Switch Modified for High Performance B

Nice work Snoop. I like the idea of using MOSFETs for assisted-switching the stock actuator. I find that a single IRL1104 is quite sufficient for handling 10A loads.

Some questions if you don't mind.

1) What's the time constant of the soft start?

2) With the electronics installed on the battery pack, it looks like you have to drop the entire innards out of the light to get the batteries out. is this the case? Or do you employ a charging jack at the rear?

3) What socket are you using int Figure 1? I haven't found a Gy6.35 socket that will fit inside of the Mag bulb ring

Thanks and fantastic work! This is some truly original thinking and I think builds nicely on the body of knowledge.

Wilkey

Re: Maglite Switch Modified for High Performance B

First off,

NICE WORK!

Second, I too have a question about the start-up time-constant. I looked into doing something like this with the TigerLight's and I found that I couldn't because of the power dissipation limits and capacitor size limits--and you're using only a 100 nF of capacitance(!). I haven't worked the numbers on your set-up yet, (although I just might do so today if you don't do it for me), but at first-glance it would seem to me that the time constant will be very short. And if it is NOT that short, how is it that you are not burning out your MOSFETs during the time when they are only partially on? I know that Willie Hunt told me that if the IRL3803 in his LVR was ever in that state that the FET would burn out almost instantly. And it's a TO-220 case, just like the MOSFETs you are using.

Also, how about the heat issues from the lamp? I would never want to put MOSFETs that close to something like a 50 or 100 watt Osram lamp, and what is holding the lamp in place? What is the pedestal or support structure made of?

Anyway, as you say, this is a great solution in that you maintain the water-tightness of the switch while increasing the current carrying ability at the same time. The only other thing that comes to mind is the same one that always comes to mind with the mag switch core: wasted space.

Even so, really nice work, Snoopy. Very impressive.

Re: Maglite Switch Modified for High Performance B

Ah,

I think I just answered the support question by looking at your other thread. Sweet! Nice solution to the problem, Snoopy. I assume that second picture in your first post shows the different configuration with the high-temp lamp holder. And in that case I can see that the MOSFETs are back behind the plate which holds the ceramic socket. Nice.

So how long have you run this? I was wondering about the wires coming from the ceramic socket and if they will ever get hot enough to melt their solder joints at the board end. I know that those wires get damn hot even several inches away from the socket.

Re: Maglite Switch Modified for High Performance B

OK. Wait. I'm still unclear on this.

When using your high-temp lamp holder in conjunction with this high performance switch mod, how are the componnents placed?

Is the plastic switch core right up against the metal support plate (which holds the ceramic socket)? And if so, won't it melt? And if not, how is the support plate being held by the snap-ring?

And is the electronics at the rear of the battery pack? And if so, how is the wiring done to connect to the modified switch and lamp?

Sorry to ask so many questions but I can't see how this all goes together.

Nice! Your circuit looks somewhat similar to the circuit(found here) that I'm planning on using in my mods.

Re: Maglite Switch Modified for High Performance B

The BUZ11 can easily handle loads up to 30A but with a thermal resistance of 75K/W between junction and ambient a single unit would get way too hot even at 0.04Ohm Source-to-Drain Resistance.

1.) The time constant is 0.47s. From my subjective estimation it takes about one third to half a second from pressing down the button to full output.

2.) Yes, since there is no room left for a connector it all goes into the lamp in one piece. Therefore, I have a charging jack on the rear battery plate.

3.) That is a G4 tube socket. It has an outer diameter of 7.6mm (0.30") and fits nicely into the standard PR-base (bulb and cement removed). Unfortunately, it is made from polycarbonate and can't handle bulbs of more than 12W. I found them here for $3.20 a piece. Please let me know when you're interested and have problems with ordering from the US.

Thanks again for the high praise. I very much appreciate that.

Regards,
Snoopy

Re: Maglite Switch Modified for High Performance B

Right. It's pretty much of a standard application. I only needed to add R1 in order to unload C1 after opening the switch.

Re: Maglite Switch Modified for High Performance B

Jim:

Yes, I've set a fairly short time constant of 0.47s. So, no problem with a physically large capacitor or burning out the MOSFET.

Regarding your question on the placing of the individual components: From front to rear we have 1) the lamp holder with ceramic socket and base plate, 2) the switch core, 3) the electronics on the front plate of the battery pack, and 4) the battery pack itself. The base plate is attached to the snap ring as described here. The distance between base plate and switch core is 0.2". The (+) and (-) battery contacts are on the rear side of the PCB that carries the electronics (Figure 3). The "wiring" is just two short copper bolts going through the PCB and directly into the electronics. When mounted in the body tube the switch core lies between the MOSFETs. So, there isn't much wasted space here.

The longest run time so far was between five and six minutes. No problems with the soldered joints so far.

I hope this has pretty much answered your questions. If not, please let me know.

Best regards,
Snoopy

Re: Maglite Switch Modified for High Performance B

Where di you get that circuit board?

Re: Maglite Switch Modified for High Performance B

Thanks, snoopy!

I realized last night that your set-up, and the one I was looking at were completely different and also that it wasn't so hard to calculate the RC time constant, and that with a large R in the mix, C could be smaller.

Anyway, you can probably do with a much shorter time constant when driving 15W or greater lamps. Something like .1s would be fine.

As for the socket plate lamp holder and switch core, how is it that there's a .2" space bewteen them? I thought that the plate was secured by the snap ring, and I know that the switch core presses right against that self-same snap ring. I obviously misunderstand the lamp holder. No big deal.

A .2" space is probably enough, but I look forward to hearing how your set-up does with a 10 or 12 minute run of a 100W lamp.

Again, nice work Snoopy. Good stuff.

Re: Maglite Switch Modified for High Performance B

Jim:
Sorry, if that has not become clear. Figures 2 and 3 show two different designs. For the lamp with the high-temperature holder I'm not using the whole switch assembly as seen in Figure 2, but only the switch core (blue & black cylindrical part on the right in Figure 3). That's why there is a 0.2" space.

Re: Maglite Switch Modified for High Performance B

mike2004:
It's homemade. I'll make a couple and offer them here if people are interested.

Re: Maglite Switch Modified for High Performance B

[ QUOTE ]
Snoopy said:
Jim:
Sorry, if that has not become clear. Figures 2 and 3 show two different designs. For the lamp with the high-temperature holder I'm not using the whole switch assembly as seen in Figure 2, but only the switch core (blue & black cylindrical part on the right in Figure 3). That's why there is a 0.2" space.

[/ QUOTE ]

AH YES! Of course. Now I see. Thanks, Snoopy.

[bump for searching with the new board software]

i started with nearly an identical circuit but it evolved into something much nicer.. http://softstart.rouse.com With the upgaded design you have much more control over the starting delay and how slow the startup ramp is.. in addition it will shut off instantaneously.. It uses only 2-3 more components and could be built for under $10.

This design is beautifully made.. do you have a start delay like i would project? i..e from the click on does it take a fraction of a second BEFORE it even STARTS turning on? this is what we experieneced testing almost an identical circuit.

Have you run the circiut on an o-scope to measure the turn on time? this post shows the startup of my soft-start in the hotdriver regulator.

-awr

hey.. ps is that a rubber band holding the cells together in the middle? depending on the bore of the light, you should be able to shrink-wrap the pack.. it will look something like this:


pps.. what size nuts/rod are you using, and did you thread it yourself? i used 4-40 nuts re-threaded to 6-32 on my mod to modamag's holder.

Is there another guide somewhere, the links for the pictures are broken.

Hey guys, I'm in the same boat, not really into resurrecting posts from the dead, but I'm really interested in this mod. unfortunately all the pictures are broken and the search hasn't produced an alternate thread

Maybe PM snoopy and ask him to update the links to his pictures. I'm sure more people must be interested in his solution.

Icarus said:

Maybe PM snoopy and ask him to update the links to his pictures. I'm sure more people must be interested in his solution.

Click to expand...

I will pm Snoopy as soon as I get my post count in