Sold/Expired Boost driver for 'D' Mag and miltiple series LEDs

I have designed a boost driver circuit for my 2D Mag light which I installed the Tri-star LED heatsink in. I wanted all 3 LED's to be in series (equal current to all of them). I couldn't find anything that I knew for sure would work. It seemed there were some that were close but I wasn't sure about all technical parameters so I designed one from scratch. It's easier to do one of these if the voltage is higher than 2 cells but I wanted one that would run off of the standard 20,000 mAhr alkaline D cells and down to a voltage of 1.8volts. (It will also work with 3D and 4D mag lights too)

When it came to putting in an aftermarket LED such as Mag's or Terralux, I wanted something that would wow people. Since the D size batteries have more capacity than AA's why not utilize the D Mag light into something that really shines. At least that was my justification.

The problem with boosting the voltage up to around 10volts and 6oomA of current to run these 3 leds in series is that power in = power out, minus efficiency of the converter. This means around 4 to 5 amps of input current on 2 D cells. I'd expect to get 4 to 5 hours continous use on the D cells but I'm not sure about current handling capability on the stock Mag light switch. Maybe someone can enlighten me there?

I decided to use Linear Technologies LT1619 low voltage current mode pwm controller because of its ability to work down to 1.8 volts. I sized up a n-channel mosfet with a balance of low RDSon resistance and also low Vgs threshold to work at 1.8 volts. The inductor is probably a little bigger than needed but wanted to make sure it could handle a variety of current capability. Anyhow I've simulated it on Linear's switcher cad, have layed out the pcb to a 2-sided 1"x1" board. I know its not the smalest thing but its robust and there is lots of room within the cavaty of the 'D' size Mag light.

I'm curious to know if anyone would be interested in this driver board. For me to get the fab (NRE) done on the gerber files is $140 and the unpopulated pcb boards are $8 ea in quantites of 12. So if I buy 12 of them my first batch is $20/ea. All of the parts can be bought at Mouser and Digikey. I'd provide a schematic and parts list but I don't have the resources to assemble. With the schematic you'd always be able to repair if needed. I did choose parts that can be hand soldered for the most part. Resistors and caps are 0805 instead of 0603 and the mosfet and pwm converter are SO-8 size. I think it's workable by hand. I know at work I have a gal do all of my small soldering becuase I can't. There is one small sot-6 transisor which I use as a current to voltage converter to save efficiency instead of a power hungry current sense resitor.

Now I haven't built this but I have worked with DC-DC power supplies and I believe it will work fine. Let me know if anyone is interested as I'm interested not in making money but in amortizing the cost to me.

One other thought. A lot of you guys probably have that LaCross BC-900 battery charger. The required input power to it is 3Vdc with 4amp capability. The power supply adapter that comes with it plugs into 120vac. I have always wanted to make an a dc-dc converter so that I could plug into my 12v outlet on a vehicle. If I don't get much interest in the boost converter mentioned above I may actually combine this project (which is also already designed) onto the same pcb fabrication so that I share the NRE charges.

Comments, suggestions, or questions welcome.
Jeff

Sounds very cool. Seems to me that since you are planning to complete this project for yourself (the driver and some sort of 3 LED setup for your 2D mag), that if you could figure a way to offer the completed drop-ins, ready to plug and play so-to-speak, that you would get a great deal of interest.

Anyway, it sounds like you've done a good job with the design considerations, good luck with the project.

Hi,
This sounds like something I would like to have, I think.

Ideally this would be IMO:
- driver for mag 2D. with 2 NiMh D batterys (so 1.2V each)
- to be used with three LEDs. (for example SSC Q4)
- able to have two output modes; 350mA and 1000mA for each LED.

Would this be possible? If yes, I would be definately interested.

Current adjustment is a matter of changing the value of a resistor. I would correlate this resistor value with output current. Nimh cells will work just fine. Minimum input voltage is 1.8v (0.9 x 2 cells). With 1A output current to 3 SSC P4 Leds amounts to around 10.5 volts across all 3 leds in series. This is 10.5 watts. The driver will handle it no problem, but this amounts to about 12.35 watts input power if you consider 85% efficiency. (I'm not sure exactly of the efficiency until I build it up and measure). 12.35 watts using 2 'D' cells is 6 amps or so when the batteries are down to 1v each. The wire off of the switch would need to be awg18 preferrably. Awg20 would probably work. The switch needs to be able to handle this current too.

Derek,
I saw on CPF some time ago, someone that had a picture of what I remember to be a rectangular driver board mounted on the end of the switch in one of these D size mag lights. I think they used silicone to hold it in place as it was standing on end. This gave me the idea that it could be done. Unfortunately it doesn't look to be a drop in replacement of course but very doable.

The tri-star aluminum heat sink housing has a round cavity where a board could be placed but the area was too small for what I was trying to do. That is drive all 3 LED's in series and do it with 2 cells. When you do this with a higher input voltage such as the adapters that allow AA cells in place of the 'D' cells then with the higher voltage means lower input currents.....which means smaller magnetics and components.

I just thought it would be fun to design and build something that exactly meets my needs. But of course I'd like it to be somewhat universal for other applications. It definitely is adjustable as far as the output current goes. Most of the issues I think are with mechanical form and fit.

Thanks for the comments.
Jeff

That is a driver I've been looking for myself. So yeah if you manage to get production going on it, I would definitely be in for a few myself.

I'm trying to figure out how to get the photoplots done cheaper. If I thought it was something that would have a lot of interest I'd just do it because in large quantities the boards would be really cheap.

But anyhow I'll keep you in mind. I'll probably just bite the bullet regardless of cost because I consider these projects a good learning tool also and I'm dying to get my tri-star 2d mag up and running.

hiuintahs said:

The problem with boosting the voltage up to around 10volts and 6oomA of current to run these 3 leds in series is that power in = power out, minus efficiency of the converter. This means around 4 to 5 amps of input current on 2 D cells. I'd expect to get 4 to 5 hours continous use on the D cells but I'm not sure about current handling capability on the stock Mag light switch.

Comments, suggestions, or questions welcome.
Jeff

Click to expand...

Here's my take on your numbers:
600mA x 10V = 6000mW
Assume 85% efficiency, so ~7100mW input required

2D @ 3.0V start draws 2.4 amps
2D @ 1.8V end draws 3.9 amps

4D @ 6.0V start draws 1.2 amps
4D @ 3.6V end draws 2.0 amps

Even using 4 alkaline D cells probably wouldn't give you 4 hours runtime, and there's no way you could get that from 2 D cells.

At a constant 2 amp draw, 2 D cells in series will be down to 1.8V in ~2 hours. Since the required current starts around 2.4 amps and increases as discharge progresses, I'd guess you could expect maybe an hour total runtime on 2 alkaline D cells.

I think it'd be really cool to see this built, but don't expect too much as far as runtime goes. You'd be better off getting some of the cheap 2AA-to-D (parallel) adapters from dealextreme and using them with NiMH cells in a 2D light. A fairly constant 2.4V from the NiMH cells at around 3 amps draw should give more than an hour of runtime.

I would say that a driver that can provide 1-2 hours runtime driving 3 emitters on 2 cells fairly brightly would still be good to me. It's always cool when someone comes up with a driver that performs well with standard batteries and standard battery configurations (i.e. 2-6 cell AAA-D flashlights). Kind of like Gene Malkoff's approach with his Mag mods and dropins--max performance on regular batteries.

I was going by Energizer's web site specification of the D cell alkaline where it says that it is 20500maH or 20 amp-hr. http://data.energizer.com/PDFs/E95.pdf

At a 4 amp constant discharge that is 5 hours. However a little further looking at the graph when you discharge at a higher rate the total capacity drops. Thanks for pointing it out. One thing about this driver is that if it doesn't give satisfactory results with one combination of cells, I can change the battery configuration like you mentioned.
Regards,
Jeff