DDL Evolution - The quest for a single-cell linear RED driver IC

With the excitement brought by some of the new LPC diodes, I'm inspired to want to create an updated DDL design that can run RED diodes from a single li-ion cell.

There's not a ton of wiggle room here. While li-ion cells may charge to 4.1 or 4.2V, they won't sustain that voltage under a load - it will drop a bit. And while LPC diodes may lase (laze?) at ~3 or ~3.2 V, they'll hit somewhere aroudn 3.6 or 3.7 at a current of 600-650mA.

So - not a ton of wiggle room. A dropout of 0.3V would seem a reasonable target - 0.2 even better - and here is my first candidate.

It's 5 pin, not 3. However, one of those pins is merely an enable pin, while the other gets tied to the (-) rail. At least, this is my interpretation of the datasheet. Input?

LD39080 - About $2 from digikey
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00150600.pdf

Does it work in a constant current regulator configuration? No CC application example given in the doc.

Just for you, I checked the voltage at a LPC-815 running at 360mA:2.9V.
Plus, a good freshly charged 18650 will only drop to around 4.0V unless you are
demanding MASSIVE currents from it.

Looks like you have more wiggle room than expected.

It looks like hitting the real high end of a LOC's current, we might see 3.7 or 3.8 V max (http://laserpointerforums.com/f50/question-about-lpc-815-loc-forward-voltage-46572.html#post631700). Since I'm really approaching this endeavor with an eye towards pushing the envelope, I'd like to be able to hit that max.

So we're looking at the differential between ~3.8 V and ~4.0V (0.2V). With a bit of wiggle room to perhaps say ~3.7V and ~4.1V (0.4V).

Either way, very small margins. The IC I linked up has a 0.15V dropout. It doesn't explicitly state that it can be used in current regulation mode. However, if it functions by maintaining 1.25 V across the ADJ and OUT pins, won't it effectively accomplish that, regardless of whether that is an intended use?

So I already did this.

It's running off of one 14500 li-ion, and I am using a home-made, SMD LM1117 chip. So... yeah.

Check it out in my siggy.

Of course, it's not running at incredibly high power (for a small host, the heat dissipation is sort of minimal), and instead is just being pumped with roughly 381mA, but meh.

Ya - and that's a good approach. But it won't be able to push a LOC (or 826) to its limits.

At 380mA, you've got a Vf of around 3.2V, which leaves plenty of room for an LM1117 to work it's magic. But you won't be able to exceed about 400mA with an LM1117. At about 580mA for example, the vF hits 3.6.

So we really do need something other than an LM1117 to accomplish high (relative) current driving of red diodes from one cell.

OK update:LPC-815, Vf = 3.19V @ 490ma.
It's fun to solder and to measure real quick.
I personally wouldn't want to push the current much further than that...

This is true - however, what's the point of using a single-cell build, besides theoretical? Because if your host is only large enough for one cell, you're probably not going to have enough room in there for a sizeable heatsink, either. And at 390mW at 580mA, the power dissipated is roughly 4.2V*.58A which is 2.4W-390mW, or roughly 2W of heat needed to be dissipated by a tiny, single-cell host (if my math is right? still not that comfortable doing the calculations).

Can we fit a good enough heatsink in a single-cell host?

I think the real goal here may be making pen-style hosts using two regular alkaline AAAs (1.5V nominal each) in order to power it.

That would give us 3V, so maybe we could do a small, DDL build with roughly 280mA (130mW) of power. That would be impressive - an alkaline powered, DDL built pen-style build. Plus, the power would be small enough that the host as a heatsink would probably be enough to dissipate the heat.

I don't see any benefit to using two alkaline cells to power anything... why not just use the equivalent 10440 or 14500 li-ions, and get higher voltage. It's a pretty easy build to power ANYTHING with two li-ions.

Regardless, I disagree.

Here's a 1-cell build I did using a 14500 and a boost driver to run a 445.
http://laserpointerforums.com/f51/f...00mw-build-njg-18-modes-62024.html#post880361

Yes- the duty cycle is pretty short. About a minute is my feeling. But we're talking 40% less current for the proposed RED build above. The duty cycle couldn't be any worse - and I don't have any problem with 1-minute duty cycles.

I suppose it depends on your intentions. Sure - if you want to push red diodes beyond their "wise" level, it would be good to do that in a host with a large heatsink if you expect any sort of meaningful duty cycle. But that's not my intent here My intent here is to find a way to squeeze the most out of a red diode, in a small single-cell host, with a DIY driver we can build for under $5 bucks

Perfectly understandable. Actually, after reading this, I was thinking about buying about ten or so of those ICs, a few in that DFN8 (4mm x 4mm) package and then the rest in the PPAK package... even though I can't seem to find any DFN8 LD39080s anywhere....