yes what he said ^
the AMC7135 things are what i like to think of as a voltage clipper :-)
if the amps goes to high VIA the battery voltage vrses the LED voltage, then it trims it down a bit, taking the extra Heat in resistance into itself.
this type of driver is PERFECT IMO for only very specific applications.
Like: say you were gonna DD anyway, these will trim the top of the DD overdrive off, then after that they do nothing.
SO
take the example of running a DD li-ion into the led, with this glorified resister, it will clip while the battery voltage is to high, then once the voltage drops down a bit, the only inefficency is the slight resistance still existing via the curcuit being there.
when the voltages (under the load) are much higher like say 6volts , then the chip has to take a lot of load doing the clipping to keep the current right for the led. basically blowing power out its backside.
with the higher voltages you end up with the more "regulated" type of output, because the voltage is high enough to NEED clipping, to maintain current.
seems to me if you know that, and understand that, then you can pick a time when that would be a best item for your application.
something like this would be actually more efficient, when it comes to energy savings, than a real converting power driver, as long as the voltage is not to high, and you enjoy the advantages of long declining output, which could mean severly declining.
It becomes terribly inefficent, if you want to be so fully regulated, and arent trying to get in the ZONE for its special properties.
I always wanted to write that :-)
take the example of running say 3xD Alkalines with a "driver" like this, starts out at 4.5v , for about 5 minutes :-) then basically that combo could direct drive ANYWAYS , pulling 2.8amps from the alkies they will choke down and you will be in DD. because the alklines wont really handle being slammed for big amps anyways, as the voltage droops , and its basically in DD, the load on the alkalines lightens up, and the weak alkalines deliver what they can, and the light runs lower and keeps on going.
BUT
Jam the 4th cell in, and you have to drop a good Volt for a while (20% loss of power efficency) and the poor thing is going to heat up a bit, and it might need a thermal path for that heat to go. the alklaines will drive it ok after a while, still be only capable of a max amp output anyways, and your into DD.
THEN
we jump over and shove 4x NI-?? D cells in the same thing, and now they maintain the voltage a lot stronger, and the poor leetel driver chip thing has to keep dropping a volt about and its really heating up, and your loosing efficiency, and then you got it Enclosed in a case, and it cant cool down, and mabey it doesnt do so well, and your wasting more power probably than a power converting driver.
THEN
switch to a 3Xni-?? or li-ion and your right back to clipping off the top a little while the voltage is to high, and then dwindling output forever.
Same IDEA
have it working barely OK with those little 3V lithiums primaries at the 6+volts, dropping well more than a volt, then switch up and try and use 3.7V (type) little li-ion rechargables, and the thing is going to Break completly.
the one thing i like to remember about Drivers, they A. sure as heck arent 95% efficent very often, and some can be less than 80% efficent, so when we drop 20% with a resister, WHO CARES :-) so do they. so its more about your preferances for regulation type and the input output.
its cheap, it has one singular purpose, it is efficent only if you like direct drive long run declining output, anything else a GOOD efficient driver becomes more power efficient, might handle better voltage ranges, acts more regulated, and depending on the driver , might not have a heat issue as much , due to converting instead of resistive clipping.