Need help with designing driver for 445nm diode

Hello,
I am new to lasers but not to electronics...these drivers are also somewhat new to me in application. I was hoping I could get some specific instruction in designing a driver for a 445nm laser diode, specifically one of DTR's 1.5-2w diodes.

I want to build my own using something like the LMxxx series of regulators.

Can someone here recommend the best regulator IC for the application? What I want is to run the diode at or near max power output. I have read that is somewhere around 1.8A? Not sure what voltage though....

So my specific questions are the following:

1) what IC should I use?
2) what is the maximum current I can run the 445nm 1.5-2w diode DTR sells?
3) what voltage battery/batteries can I or should I use for maximum output?
4) what resistor value do I need to set the IC to the max current above?

The reason I am not using a FlexDrive or similar prebuilt driver is cost primarily....I plan to build a laser with three diode heads in one giant MagLight heat sink. jayRob is building the sink custom for me. Since it's a MagLight host I plan to use three separate Cells, one for each driver/diode. Plenty of room in the long host to make a module/sink for the three drivers and batteries....that's my plan anyway.

One click...three beams of light each close to the max 2w output...

Why do this? Because it sounds cool, why else! Maybe down the road I will switch out the diodes and adjust the drivers to have three different beams, maybe red, violet, and blue....who knows.....

I really appreciate the instruction, you guys have been great since I started this crazy hobby.

Sincerely,
Jeff

Hey,

Nobody will spoon feed you the basic driver instructions because there are SO many threads on here dedicated to EXACTLY the type of driver you're trying to build. Just do a bit of reading.

BUT, I'd be happy to weigh in on the trickier part of your build -

If you're building your own linear driver, for three diodes, go with ONE driver and four batteries feeding it in series. Similarly, wire the diodes in series. That way, if one diode fails, you won't lose the other two the way you might if all three were parallel wired. Plus, it makes the driver essentially as simple as a single diode driver.

One way to do it, is to use a Maglite 'Stubby' build... (single cell)

And use three FlexDrives in parallel. (even though you said you want to build your own drivers, this way is easiest)

if you want a linear driver i have one they are 2 for 15$ but the 2 for 15$ is limited one set per person.
But i can sell you 3 drivers for 25$.
http://laserpointerforums.com/f67/new-linear-driver-st-drive-67549.html

rhd said:

Hey,

Nobody will spoon feed you the basic driver instructions because there are SO many threads on here dedicated to EXACTLY the type of driver you're trying to build. Just do a bit of reading.

BUT, I'd be happy to weigh in on the trickier part of your build -

If you're building your own linear driver, for three diodes, go with ONE driver and four batteries feeding it in series. Similarly, wire the diodes in series. That way, if one diode fails, you won't lose the other two the way you might if all three were parallel wired. Plus, it makes the driver essentially as simple as a single diode driver.

Click to expand...

Since I first posted I was able to find the info needed to make my driver...it's here but not easy to find.

I think I have settled on the LM1085. With a sense resistor of 0.68R I should yield just a tad over 1800mA. Please check my calculation.

Running three diodes in series at 1800mA each...what's the drawback of this? If any? Since they are in series the whole circuit should be set to deliver
1800mA correct? So just tie the diodes in series, connect to the DDL, and set sense resistor to deliver 1800mA to the circuit...same calculation I made before, correct? Something seems intuitively off with this...I would think that if I want each diode in the series to draw 1800mA then the total circuit would need to deliver 5400mA. How am I seeing this wrong? If the DDL IC were to require a 5400mA output then what IC is up to that task?

I am very interested in this series idea but I am going to need some specific instructions here, would you mind terribly explaining this in greater detail, exactly how do I set this circuit up to deliver 1800mA per diode...

Thanks,
Jeff

If diodes are in series, and each uses 1800mA, they all together still use 1.8A.

It's the voltage that gets added up when you put things in series. When in parralel, then you add up current.

So each diode at 4.5 V, you've got three, you need ~13.5 V of output at 1.8A. That's around 25 Watts of electrical power output. Make sure your driver can regulate that, and your batteries supply that.

It's a very tricky task you've put before yourself. I'm not saying it's impossible, I'm just saying - don't get disappointed if you fail for the first time. We've all been there.

The 1085 can handle the 13.5 voltage delivery.

I'm using one in my tent melter light. It's got 4x cells powering a 1085 (1084 actually, because I'm running higher current), which powers LEDs with a Vf of around 13V. Similar specs.

Just heatsink the IC, and you'll be fine.

Ahhhh, yes....how could I have forgotten about the series voltage! Now that makes perfect sense. So, 4x 18650 cells would work well...is there another 3.7v battery that is narrow enough to get two side by side into the barrel of a Maglight? Would sure be nice to have x8 cells in a series/parallel arrangement to net my 14+v. Anybody ever done this before?
What about RC hobby batteries? I think they make some high capacity 14.7v batteries...would that voltage work fine?

So, if I were to series four 3.7v cells that should net me 14.8v...isn't that too high? Is it ok to set my DDL to deliver 1800mA at 14.8v to three diodes in series? Or should I adjust the current down a little, say 1600mA.

I saw a thread where someone used two 445nm diodes in one head and somehow joined the beams into one beam...can someone refer me to some information on how one achieves a single beam from multiple diodes? I imagine these are some expensive optics involved but it sounds like something fun to try...this guy made a 3.5w 445nm laser somehow but he didn't go I to the detail of how he conjoined the diodes into one beam.

Lastly, the LM1084. Would you recommend this over the 1085 for my proposed application? I can order them just as easy from digikey.
Thanks a million!

LM1085 is just fine for under 3A of current.

For your battery issue, it's pretty easy to figure that out. Just take the diameter of the batteries your host was meant for (26mm for C or 34mm for D) and figure out what type of lithium ion cells could fit in that space.

For example, an 18650 =
- 18mm diameter
+ 65mm length
+ 0 means round

So you (probably) won't fit two 18650s (36mm) side by side in either a C or D host.
But you might be able to fit 2x 17670s side by side in a D host.
And you could certainly fit a few 10440s side by side in either C or D.

It doesn't make sense to do it that way though. If you can get a 4x C or 4x D Maglite, go that route, and use 4x 26500 or 4x 32600

Again, please clarify this .....
If I have four 3.7v batteries my net voltage is 14.8v. The diodes are 4.5v x3= 13.5v. Will I run into problems serving my DDL to 1800mA to drive tue three diodes in series with a 14.8v source? If not, what adjustment would you recommend?

One other thing. You need to isolate the diodes one from another. Meaning you do not use the case pin in any circuimstances, you need to obtain the battery negative connection through some other way, which may prove difficult.

Note that each lithium ion cell also is 4.2 V when fully charged, so that's 16.8 Volts, and at 1.8 Amps of current and 13.5 Volts of output, you've got 6 Watts of heat to deal with on the driver, in addition to ~20W of heat from diodes.

So, the case of the 445 is the negative pole of the diode huh? Well, that will be very problematic...looks like I will be going back to three spereate drivers with two cells per diode...six cells. I don't think it will be possible to effectively sink the diode in a head drilled for three diodes AND keep them electrically isolated from one another. Unless someone here has an idea...seems to me though any idea would be difficult to implement. It was a good idea for a while though...

If you take my suggestion and run them in series, with four cells, and one 1085 based linear driver, then you don't need to worry about isolation. In fact, you can just grab the battery negative from ANY case pin, and ignore the other two.

Plus, you save $75+ that would otherwise be wasted on flexdrives.

rhd said:

If you take my suggestion and run them in series, with four cells, and one 1085 based linear driver, then you don't need to worry about isolation. In fact, you can just grab the battery negative from ANY case pin, and ignore the other two.

Plus, you save $75+ that would otherwise be wasted on flexdrives.

Click to expand...

If the case is the negative pin and I place it in a heatsink and mount all three diodes into one heat sink then there is no way to run them in series...all the negative pins will be tied together because they are touching the heatsink...you cannot tie + - + - + - in series AND have all the - - - connected to each other too....tell me I am wrong here and show me how to do it, believe me, I want to do as you say but I don't see how it can be done without isolating the cases of each diode from one another...

445nm diodes are isolated. the pos and neg are all isolated from everything.

Jmillerdoc said:

If the case is the negative pin and I place it in a heatsink and mount all three diodes into one heat sink then there is no way to run them in series...all the negative pins will be tied together because they are touching the heatsink...you cannot tie + - + - + - in series AND have all the - - - connected to each other too....tell me I am wrong here and show me how to do it, believe me, I want to do as you say but I don't see how it can be done without isolating the cases of each diode from one another...

Click to expand...

I feel like this is the twilight zone. What am I missing here? Were it not for all the Thanksgiving drinking, I'd put my life on the fact that there's no isolation issue.

445s are case isolated, why would there be any issue with diode isolation?

Am I having a brain fart? Eud is more of an authority than almost anyone here, so I must be missing something.