Case positive devices

Why are most flashlights/lasers case positive? I always thought case-negative would be a better option but I must be wrong about this.

Can somebody explain why?


Thanks.

Well, on lasers it makes perfect sense flashlights however...

Can somebody explain why?

Click to expand...

^

Red lasers are usually case negative, IR (pump) diodes usually case positive.

I assume the difference arises from the technologies involved, with IR pumps often being side emitting (heterostructure) designs, and red lasers being VCSELs... but i'm by no means sure.

Benm said:

Red lasers are  usually case negative, IR (pump) diodes usually case positive.

I assume the difference arises from the technologies involved, with IR pumps often being side emitting (heterostructure) designs, and red lasers being VCSELs... but i'm by no means sure.

Click to expand...

While you're on the right track here, red lasers are NOT always VCSELs. VCSEL technology is an entirely new way of making laser diodes in general, of all wavelengths which eliminates the astigmatism and elliptical beam profile that standard laser diodes exhibit. Red laser diodes can be VCSELs, but usually they are edge-emitting just like the vast majority of other diodes. Whether or not a laser diode is case positive or not depends on the construction of the actual laser diode chip itself and whether or not the orientation favors case-positive construction or case-negative. In some cases, (violet diodes for example) they are case-neutral (floating).

Laser diodes are made up of layers of materials on a substrate. If the orientation of the chip favors case positive, then the positive contact will be on the bottom of the stack and the negative on top. Of course, the orientation could be reversed just by flipping the chip over, but this may not be the best orientation due to heat dissipation needs, ease of wiring and other factors.

Yeah every single red diode that any of us use are side emitting. VCSEL's are expensive, can be high powered, and are generally super badass.

To be honest, I'm not sure why they choose different case polarities. I dont think the die really cares (but I could be completely wrong here) because as far as I know I THINK red lasers and IR lasers use basically the same materials for the die itself, so I dont know why they are different. I think it just has to do with general applications. For us, we want case neg diodes because we use positive regulators. If we used negative regulators, everything could be case positive. Also, in the world outside of pointers the polarity really doesn't matter.

Requesting climbak and/or Frothy-awesome to jump in

Niko said:

Requesting climbak and/or Frothy-awesome to jump in

Click to expand...

You should REALLY be asking for pullbangdead, because he's the freakin' MAN when it comes to this stuff.

GooeyGus said:

[quote author=Niko link=1240799632/0#6 date=1240970677]Requesting climbak and/or Frothy-awesome to jump in

Click to expand...

You should REALLY be asking for pullbangdead, because he's the freakin' MAN when it comes to this stuff.[/quote]

Requesting pullbangawesome to jump in.






;D

Awww, shucks you guys...  :-[


You guys covered it pretty well, actually  ;D.  It really just depends on the most convenient way for your application, and really whatever diode contact scheme gives you the best performance and convenience.

Violets are generally easy to "float" (case neutral), or you can pick either polarity, because violet GaN-based diodes are grown on a sapphire substrate.  The sapphire carries heat away well, but is electrically insulating, so when you mount it, there is no continuity between the diode and the case.  In the future, this *could* change, because the industry will likely move to using GaN substrates instead of sapphire.  When that happens, conductivity through the substrate becomes possible.  Not required, but possible.  It could be advantageous to put the n-contact on the back of the substrate, or it might not be worth it and things could stay like they are; that part remains to be seen what will happen with backside contacts for GaN substrates.  Positive case wouldn't really be any more likely than now with GaN; as it is, you can make the case positive if you want, and the contact that would be made through the substrate with GaN would be the n-type/negative contact, because the n-side is always grown first for GaN diodes.

For red/IR, I'm not familiar with how they do contacts for those diodes.  But the gist of it is, whatever is most convenient for the diode and its substrate.  I really don't remember if you can make contacts through InP or GaAs, but my instincts say yes, at least in some cases.  It would make sense to be able to make the n-contact to the diode through the GaAs substrate, resulting in a case-negative diode, since that's what most of the reds are.  Or it could be possible to electrically float the diode as well, I really don't know.  But everything I've done and everything we do is GaN based, so i don't really have much experience in the other material systems for red and IR.

And all the laser diodes we ("we" being LPF laser diode users) use from optical storage drives are edge emitters, and the contacts don't depend on whether the diode is edge emitting or VCSEL anyway.

Thanks for the information

All diodes grown on any of the GaAs (InGaAs or AlGaAs ie anything from far ir to 600nm) will favor a case positive configuration because of heat reasons.  The laser is 'grown' on one side of a slab of GaAs, and to power it you need to apply + to the side with the laser on it, and - to the other side of the GaAs substrate.  So you can solder the chip down with the substrate side down, but all of the heat generated by the laser structure has to be transfered through the substrate, which limits the maximum (electrical) power you can use, and also causes issues with the laser heating up depending on current (more important for telecom applications where they need the lasers to be stable to a tiny fraction of a nm).  The more logical choice is to solder the side with the laser structure to the substrate, because now the laser is soldered directly to a piece of copper and you can remove a heck of a lot more heat out of it--you just have to be careful when you design the encloser/driver.  Some manufactures get around this issue by putting a small block of something that is an electrical insulator, but very thermally conductive (better than GaAs at least), between the chip and the heatsink, which allows the diode to be floating but still have good thermal properties (but of course it costs more).

As the the 405nm stuff I have seen (which is made out of GaN, which is grown on saphire--you can't make a laser diode out of Al2O3), all of the diodes I have seen have used sapphire that is doped with something (no idea what) so that it is conductive--they just make it floating by mounting it on a block of what looks like undoped silicon? to isolate it from the case.  

... said:

All diodes grown on any of the GaAs (InGaAs or AlGaAs ie anything from far ir to 600nm) will favor a case positive configuration because of heat reasons. *The laser is 'grown' on one side of a slab of GaAs, and to power it you need to apply + to the side with the laser on it, and - to the other side of the GaAs substrate. *So you can solder the chip down with the substrate side down, but all of the heat generated by the laser structure has to be transfered through the substrate, which limits the maximum (electrical) power you can use, and also causes issues with the laser heating up depending on current (more important for telecom applications where they need the lasers to be stable to a tiny fraction of a nm). *The more logical choice is to solder the side with the laser structure to the substrate, because now the laser is soldered directly to a piece of copper and you can remove a heck of a lot more heat out of it--you just have to be careful when you design the encloser/driver. *Some manufactures get around this issue by putting a small block of something that is an electrical insulator, but very thermally conductive (better than GaAs at least), between the chip and the heatsink, which allows the diode to be floating but still have good thermal properties (but of course it costs more).

As the the 405nm stuff I have seen (which is made out of GaN, which is grown on saphire--you can't make a laser diode out of Al2O3), all of the diodes I have seen have used sapphire that is doped with something (no idea what) so that it is conductive--they just make it floating by mounting it on a block of what looks like undoped silicon? to isolate it from the case. *

Click to expand...

This is a great answer, guest, you should register, you'd be a great addition to the community

... said:

All diodes grown on any of the GaAs (InGaAs or AlGaAs ie anything from far ir to 600nm) will favor a case positive configuration because of heat reasons.
*

Click to expand...

There must be something more to it though.. 808 diodes are commonly case positive, while ~650 are case negative in almost all cases.

All laser diodes over 50mW are going to be "case positive" if mounted directly to the can, but some of them such as the infamous long die reds use a thermally conductive insulator to MAKE them negative case. The insulator materials are either diamond, aluminum nitride, or beryllium oxide. The reason they are case positive is that the junction is on the P side of the chips and is very close to the surface, that way the heat generated- READ: 10's to 100's of KILOWATTS PER SQUARE CENTIMETER power density- does not have to go through the bulk of the chip to get dissipated.

That sounds very reasonable, but brings us back to the original question: For some reason a 200 mW 808 diode is manufactured case positive, and a 200 mW 660 diode is made case negative.

Having it case negative has its obvious benefits for installation in a drive since the chassis is usually grounded and often only positive power rails are available... but what would compel manufacturers to go through the troubly only for reds and hardly ever for infrareds? Do you think this is only because of market demands, not for any technical reason?