Is there a simple circuit solution for this?

[ballen]

I want to feed a small laser diode, 3.5VDC @ 25mA.  My power source is a single 18650 Li-Ion cell, whose nominal voltage range is 3.7 - 4.2V.  Is there a simple circuit which will provide either fixed voltage @ 3.5V or fixed current @25mA, dissipating the extra (0.2-0.7)V x 25mA as heat?  My usual ways of doing a constant current source would be with a 3-terminal device, or a transistor, which always creates at least one junction of voltage drop.  I can't afford that here, since if the battery is nearly discharged, I only have 0.2V of headroom.

 So, given that the load is fixed, is there a simple way to create a constant current or constant voltage source which will work well over the dV range of 0.2 to 0.7V?

PS: if this is the wrong forum, please tell me to go and I'll move the post.

[wraper]

My power source is a single 18650 Li-Ion cell, whose nominal voltage range is 3.7 - 4.2V.

It's not. More like 3.0V-4.2V, unless you want to use only a small part of it's charge. Basically you need a buck-boost converter with constant current regulation. Or boost converter + resistor to drop excessive voltage/limit current (less efficient). You should not connect laser diode directly to the voltage source without current limiting, the same as LED.

[SilverSolder]

It probably isn't going to be possible with the simplest circuits, it seems to me.

It might be possible to design something around a pass transistor, low value current sensing resistor, and a low voltage op-amp?  -  that is still pretty simple, just not quite as simple! 

[ballen]

Wraper: I see your point -- here is a charge/discharge curve:



Almost all of the capacity is in the 4.2 - 3.5V range.

A buck upconverter would be a reasonable solution, given the small amount of power needed.  I suppose that I could use one of the tiny 5V output ones, in series with a 1.5v zener.  Is a simpler solution?

[wraper]

I suppose that I could use one of the tiny 5V output ones, in series with a 1.5v zener.  Is a simpler solution?

You can use 5V boost converter + current limiting resistor. Zener should not be used since laser diode is a current controlled device. Powering from 3.5V will not result in stable 25mA current. It can vary a lot depending on particular laser diode and temperature.

Almost all of the capacity is in the 4.2 - 3.5V range.

According to the curve you provided in 3.5-3.7V range. But it also depends on a cell, discharge current and ambient temperature.

[tunk]

Almost all of the capacity is in the 4.2 - 3.5V range.

Depends on the cell and discharge current.
You can find many examples here: http://www.lygte-info.dk/

[ballen]

Actually since an LDO current regulator has a drop of about 1.2V, I'd have just enough headroom for that plus a 5V buck upconverter.  The laser diodes will be under 3.8V.

[David Hess]

My usual ways of doing a constant current source would be with a 3-terminal device, or a transistor, which always creates at least one junction of voltage drop.  I can't afford that here, since if the battery is nearly discharged, I only have 0.2V of headroom.

There is no requirement for a full junction voltage drop in a regulator.  The collector-emitter saturation voltage can be considerably less than the base-emitter saturation voltage.

There are many small LED driver ASICs which can be used to drive laser diodes.  Alternatively any buck-boost or SEPIC converter can be combined with an operational amplifier to make a constant current source.

[ballen]

Whole question is moot.  I was just experimenting with the diode and discovered that it already has a built in controller.  Clue was, when I increased the voltage, the intensity did not change and at some point the current dropped slightly.  So no need for an external controller.   It runs happily on 3-5V.

[magic]

If the "diode" has three terminals and a small PCB with two transistors and a few resistors on its back than this is indeed the case.
The third terminal is feedback from an internal photodiode which measures output intensity and the regulator is a simple circuit with low dropout voltage.

There are "bare" laser diodes too and laser diodes with a PCB containing a single current setting resistor, intended to run from a particular voltage.

[ballen]

Device is packaged in a small metal cylinder with an adjustable focusing lens and a red and black lead coming out the back.  I don't think I can see what's inside without destructive disassembly 

[magic]

Well, nevermind then. Mine was like shown below and yours is probably similar if it maintains constant current.

Actually, with one difference. I think R2 was some sort of a thermistor originally, but I damaged it while replacing power leads and installed a fixed resistor instead.

[RJSV]

Let me see if I got this right, please:
   You present a question: "...need proper fixed VOLTAGE..."
"...for a FIXED load...".   NO, it's not a fixed load: it's a semiconductor.
   After which a likely very experienced reply comes:
   "...don't connect directly to a voltage..."

  So, you then proceed to another possibility: A VOLTAGE, this time created by a zener, followed by an inevitable reply, explicitly warning:  "laser diode is a current-controlled device.."
Generous with advice, that reply also mentions things (like necessary VOLTAGE) 'can vary a lot, with temperature..."
   THEN you introduce another LDO regulator option and focus on... wait for it: VOLTAGE output to laser diode.
Granted, of course you have to ensure the headroom is sufficient, there.
To which, the predictable reply ends with:
'...to make a constant current sourc, e...'.
   You then further 'experiment' with drive VOLTAGE, even mentioning  current NOT constant, a little. (Whatever that means), 'happily' working, (and again, at whatever the device temperature was then...who cares?)

And yet, yet, again REPLY comes: to focus on 'current setting', for a temperature-dependant semi-conductor load.
Ok, I'm done.
Set your sights on 'device current'!!!
  I made that mis-step a bunch of times, (tried voltage controls).

[NiHaoMike]

A MOSFET current sink can have close to zero voltage drop.