*safe* and small dc-dc converter 1.5v -> 3.8v, @180mA?

Hi,

i bought three UV-LEDs from Roithner. Being (relatively) expensive, i´d like to build me a circuit to drive them, where i can be 100% sure nothing nasty happens. I dont know how delicate they are, i heard the early blue LEDs wouldnt even survive wrong polarity!
it would be nice to drive them off a single cell, and to get all of it into a tiny light. but again, top priority would be security.
i can solder SMD too, and guess i dont have any chance to get all of it in, say, a AA-cell light otherwise.

The three LEDs are each:
RLT360-1.5-30-3CC, 360 nm, 1.5 mW at 3x20 mA, 3 chips
there is no datasheet. the datasheet of a one-die-version says 30mA and 3.8v.
<a href="http://www.roithner-laser.com/All_Datasheets/LEDs/RLT360-0.5-1x_series.pdf"> Datasheet RLT360-0.5-1x</a>
would add up to 3.8v at 180mA.

Unfortunaly i have no access to all those nice cheap flashlights with converters, as i live in germany. there arent even any fenix lights on our ebay!


i hope you have hints for me, be it about burning UV-leds, shematics, or sources for tiny converters!

thank you in advance!

Manuel

disregardhttp://theledguy.chainreactionweb.com/index.php?cPath=48_49_61

Hi,

i looked around a bit.
a nice IC would be the LT1932 from Linear Technology, its a step-up converter, accepts down to 1v input, can output 500mA+, and, best of it all, limits the current.
no, wait, the best of it all actually is that its available here too!

Product Page
Datasheet


i still am unsure about some details though.

[3 LEDs, each 3 chips, 3.8v and 20mA, each chip has a single connector]

- can that IC even deliver 684mW from a single cell?

- how would i wire them? in parallel would mean 3.8v and 180mA,
in series would mean 34.2v at 20mA. in series surely is better, but i dont
know if that circuit can handle it.. "absolute max LED voltage 36v", no idea
at what current that would be.

perhaps it would be just better to use three circuits, one for each LED, but that would be even more expensive than the LEDs themselfes, and much larger too.

and finally:
the other parts, especially the inductor. how critical are they? would, for example, any 6.8uH inductor work?


i still dont feel too comfortable to discharge an inductor into my LEDs... they really dont get too high current/voltage?
sounds more like PWM'ing 30v into a 3v LED with a 10% ratio.. doesnt sound healthy to me, anyway.. can anyone comment on that?

thanks again

manuel

Hi,

just had a look at the data sheet:

- 1V input might work, but efficiency will go down significantly
- for your setup, I'd use either the circuit on page 15 (8 white LEDs) or on page 16 top (10 white LEDs), but you're going to have to try for yourself which configuration will work better in your case, see above.
With single cell input and 9 LEDs on the output, you're at the edge of the spec.


Bye
Markus

The chip is only set up to deliver 40ma max to the leds so you have to go with a series connection or series parallel connection. The series parallel (two led strings @ 20ma) would keep you away from the max voltage limit. But its going to be hard to put 4.5 leds on each string :laughing: , so you'll have to balance the currents by tweeking the series resistors.

The inductor choice is fairly critical - "any" 6.8uh will not do. If you look at the datasheet it says the switch current limit is as much as 780ma, which means the inductor must be rated for that current. You will be running the thing flat out at 40ma out and 1.5v input or less.

You can guesstimate the inductor current by working back from the led output power - 9 leds times 3.8 volts times 20ma = .684 watts out, ignoring the power lost in the series resistors. If the circuit is 80% efficient you need .855 watts in. At the 1 volt limit this is an AVERAGE input current of 855ma - peak inductor current is even higher. ie the chip will not be able to deliver this much power at 1 volt input because it will current limit. Even with a fresh 1.5v battery (.57A average input current) you are probably exceeding what the chip can deliver.

The Zetex ZXSC300/310 need an external transistor but they can deliver much higher power levels and it will work down to .8v input. Do a search here for "zetex" and you'll come up with a bunch of good stuff.

Rich

Hello again!

sorry i didnt answer earlier, i moved to my mew flat and just am online again.

i still look for a way to safely drive my leds.
unfortunaly it looks like impossible to get all those nice led-driver-ics here (in germany).

soo.. i guess i´ll have to use >4v to drive them, and make sure they never get more than their 180mA (together).

top priority is still to drive them safely, under any condition.
it would be nice to keep it small, and it would be nice to have them at constant brightness before i swap the batteries.

how much can those AG13/LR44 button-cells deliver anyway?
i read "150 mAh" per cell somewhere ..
how about using 3 or 4 of them in series and some discrete electronics to regulate the output? "regulating" like with a FET, no converter or similar.

any suggestions or hints?

thanks in advance!

manuel

Look at the zetex circuit. ZXSC310 The current can be adjusted to 180mA easily. The other IC I'd recomment is the LTC3490. It requires a much smaller inductor

That 360nm LED spec looks the same as the Fox one -- and since they make a 3-chip version at 360nm, I would bet they are the same things. I don't have the Fox datasheet for the 3-chip one handy, but if I find it I will post it.

Piers!

finally!

thanks for flashlightguy i searched again for those ICs, found "new" vendors, and one of them has the xetex ZXSC310! they have the required transistor FMMT617TA as well, just not the other things like the special zener diode and the coil. i guess i´ll find some alternatives there too.
fyi, its www.buerklin.com. propably the only shop in germany selling these..

soo.. i´m surprised that there is so little information in the forums about designing a driver? can that IC deliver 0,7 watts from low voltage? would it work with a single cell? sure, they have graphs, and high efficiency, and work down to 0.8v, but no hint about single-cell usage.

i guess the IC would prefer to drive 3.8v at 180mA (all parallel) instead of the LEDs in series, or parallel-serial configuration? i never "designed" a step-up-converter yet..

and finally, the question i began with:
is it safe? how sturdy are 360nm UV LEDs, is that "pulsed" mode from a step-up-converter fine with them?

thank you xenopus for your hint, i´ll try to dig some info about Fox LEDs!


thank you all very much for the help, wouldnt even have gotten here by myself!

Manuel

MAX1795

hmm, no MAX1795 available here.. my list of electronic-vendors is now at six..

thank you anyway!

I can buy directly from www.maxim-ic.com, or even get samples for free.

To my knowledge, MAX1795/6/7 circuits are the most efficient when sourced with 1.5V or less.

Just do a search for Schottky diode. They have some there. Then you just need some low ESR capacitors. Some decent tantalums might be fine for what you are doing.

krutzbeuazen said:

thank you xenopus for your hint, i´ll try to dig some info about Fox LEDs!


thank you all very much for the help, wouldnt even have gotten here by myself!

Manuel

Click to expand...

Depending on how long you have this chip lit up, I'd be concerned about heat -- LEDs get hot when packed together running at 25mA, and having 3 chips in one die makes me nervous. In a small flashlight it's probably fine.

Piers

Aaahh..

I learned two things:
-the shop with the Zetex chips only sells to companys
and, worse
-the zetex chips are voltage-driven..

So I´ll forget about Zetex alltogether.

I guess I´ll end with a D-Mag-Size flashlight instead of a Minimag-Sized one as I hoped. With a horribly inefficient FET-driver. That´ll be the ugliest flashlight ever, with three LEDs in it, barely visible.

Anyone any hints for a *safe* current-controling LED-driver without ICs, the smaller the better?

Manuel

Try a SHARK from Wayne : http://theledguy.chainreactionweb.com/product_info.php?cPath=48_49_61&products_id=721

of maybe another of the Converters, maybe an MJ-mini converter, or a Badboy set for Your output (should be something with 0.7 Ohms, dont have my data sheet here)


DOH!!! You typed 1.5 Volts. hmmm
then the option is a madmax, if it is possible to have it set to that small output. Ask Wayne is this is possible.