DIY lm317 driver for 445nm 1W diode?

[anselm]

I have been using such a lm317-based driver for my 445nm diode.
The 2nd link in my signature is the thread to the build, and in that thread should be a link
to another thread talking about the driver.

 

[Benm]

Yes - but you need at least two, potentially three li-ion batteries to power an LM317 build with a 445nm diode.

Count on 3!
Even with 3 cells you are on the edge of guaranteed operation of the LM317:

5-6 volts for the laser diode
1.25 volts across the resistor
3 volts input-to-ouput drop for the LM317*

Worst case that would end up with a requirement of 10.25 volts. Depending on the exact batteries 3 will probably suffice, but 2 definitely will not.

* the 3v is specified by the manufacturer as minimum dropout required, but it -may- work with a lower dropout in real life.

 

[anselm]

I don't mean to step on your toes, Benm, but do you have practical experience to back that up?
My DIY laser works fine off 2x18650.
Sure, you want them to be charged, I suppose 3 batteries would be even better.

But I can tell you that 2x18650 = 8.4V is plenty for this config, and I have pushed 1.6~1.8 Amperes.

5-6 volts for the laser diode

more like 3.5V - 4.7V

this:

1.25 volts across the resistor

is already acounted for in this:

3 volts input-to-ouput drop for the LM317*

 

[Cyparagon]

5-6 volts for the laser diode

I don't think it is quite that high. 4 point something plus <3V total for the regulator means 2 lithiums will be fine. If the voltage isn't quite enough towards the empty battery state, the current will be just a little lower. No harm done.

If you add a third, you might garantee it operates at full current until battery death, but it is at the cost of a LOT more heat. 3-5W more for 1A+. It is also difficult to find a battery casing that takes three. You also run the risk of over-discharging the batteries.

 

[rhd]

There's also the option of going LM1117 - less of a voltage dropout.

I haven't yet received my LM1117s to give this a shot, but that should put us well within the two-cell range, even for a 445nm diode. Max current of 1A per IC though, so you'd have to parallel two of them to get higher currents. Though they're smaller than an LM317 anyway, so that shouldn't be a problem in most cases where an LM317 would otherwise work.

 

[jib77]

From my tests 6V is plenty to drive a 4V load with a LM1117. Charged li-ions deliver 3.9V under load, so 7.8V is plenty for a 445. The ON Semi NCP1117 I am using in my drivers actually go up to 1.33A when properly heatsinked.

 

[rhd]

445nm is more than a 4v load.

 

[jib77]

445nm is more than a 4v load.

4.7V, so theoretically 6.7V input should be enough, thus 7.8 from 2xli-ions is plenty.

 

[anselm]

445nm is more than a 4v load.

The diode's Vf is a function of the current (or maybe the other way round)
If you drive it at <500ma (roughly of course), the voltage drop will be <4V.

 

[rhd]

I was under the impression that 445nm diodes could require upwards of 5V in some conditions, and also that LM317 could drop as much as 3V on the edge of their (1.5A) current supply limit.

If that were the case, you could see how an 8V requirement would be just at the edge of what two (2) fully charged lions could provide.

But some of my numbers must be off? If 445nm diodes really don't pull in more than 4.7V and the LM317 is under 3V drop, then you could get away with 2x lions as long as they were fairly fresh, and above the 3.6 nominal each.

I'm excited to try LM1117s though - sounds like they'd put this entire debate to rest if they really drop 1/2 the voltage of LM317s.

 

[jib77]

This is a great read, especially how they point out the difference in topology between normal voltage regulators (LM317 type) and Low Drop Out regulators (LM1117 type). This is what influences the overhead needed to saturate the internal transistors to begin regulation.

>>Linear and Switching Voltage Regulator Fundamentals<<