With my LED's (and I think that goes for most LED's) max pulse current is the same as max continuous current...
Are you still using the OSRAM LWP4SG?
Because that is not true for them nor most other LEDs - eg, the 20mA-continuous LWP4SG takes up to 200mA.
But it sounds as if you still don't understand PWM for controlling intensity (eg, as in fluoros which are similar to LEDs in that they do not tolerate "normal" dimmer switches etc that reduce voltage). Resistance has already been discussed.
But fear not - I won't reply further unless perhaps you are interested in understanding.
I originally wondered if you had a legitimate reason for using a variable analog device for brightness.
Meh, from personal experience the only disadvantage with PWM is that LEDs can switch at high frequencies and so does PWM. Sure it's easy enough to throw a capacitor in there large enough to smooth it out for some frequency range. Maybe the top or bottom. Kinda hard.
The thing is, the PWM is going to switch the input voltage. In a car it might be 11 or 15 or a load dump. That voltage is going directly into the LEDs, just switched. Sure, regulate it. With switching or linear? LOL Well, I don't regulate it and it's fine. There's a fuse on it, and fuses and LEDs are cheap.
Current regulation still needs voltage regulation, but current is what creates heat. The more heat the more wear or thermal overrun and all that fun stuff. So it's a more controlled environment for the LEDs.
It isn't voltage that is being regulating or limited the LEDs - it is current. (Over some time - whether 20mA continuous or 200mA for 200uS or 10% duty cycle for BES's former OSRAM LWP4SG 2.9-3.7V 20mA 2600i LEDs.)
We happen to use voltage for continuous operation (eg, 2.9-3.7V for 20mA for the LWP4SG), but that is irrelevant for PWM where it is the current's duty cycle that is varied with a constant voltage source...
Remember that resistors are used merely because current limiters are generally considered too expensive (hence a real overkill situation for most applications except where strings exceed a handful of LEDs at high voltages and reliability is paramount - that's not an issue for 3-4 series LEDs at ~15V or less).
And what will you fuse the LEDs to - the max 200mA that they can tolerate? Or is there a 20mA fuse in each string? (Not that I have ever seen LED strings fused!!)
I was also trying to point out how inefficient and limited analog current/brightness limitation is.
It doesn't matter how you look at it - whether resistors or transformers & linear regulators - switching circuits like SMPS & PWM will always be more efficient. (Overheads excluded, but they are quickly swamped by the mA for a single LED (string)!)
And switching of LEDs at high frequencies??? YES! That is why they are used for OF (Optical Fibre) communications - the new copper to your home to your communications needs. (Even if this www update is via satellite, it will still pass through LEDs!)
To paraphrase a previously attempted conclusion....
Is it clear that LEDs are current devices?
Is it clear that caps are not used for PWM or LEDs? (If so, are they used in LED muliplexed display situations like clocks? And do those clock displays have resistors in their 7-segment drivers (often anodes) that limit to the continuous LED current (ie, 10-20mA) or are they smaller (ie, allow 40-200mA etc)?) [I'm trying to get the horse back in front of the cart. Hence I'm using typical daily examples (probably long since forgotten lol!).]
PS - (missed replies)... I recall someone saying "robs current from another string" which is complete bullsh - except where overall current limiting exists - but resistors do not provide that, nor do typical automotive circuits.
And what is the lifecycle effect of the LWP4SG being driven at 200mA in accordance with their specs which is to the same lifecycle as 20mA continuous?
But years of experience are often independent of misunderstandings like that. Hence why even the old dogs can be wrong or learn something new.
Don't provide fodder for the youngies when they allege "oldies think they know it all...." when they don't! (Unfortunately many of them are right.... damn it!)
Alas despite my years of experience and leading-edge involvements, it has only become apparent to me relatively recently how flawed so many practices and beliefs are, and that's probably thanks to a greater non-involvement with the mainstream.
But I for one enjoy realising that I have been wrong, or have misunderstood something based on a limited application or downstream knowledge.
But that's the fun of being young enough to still learn and roll with the punches instead of rejecting and defending... Or is my ego to big to continue defending something that is wrong..?
But I'd better go and put my nappie on. Funny how one returns to childhood...
PS - I'm not interested in discussing. What I am saying is correct. If it isn't, throw some facts or details, not assumptions or misrepresentations. (If you do misrepresent something, I may take that as a deliberate obfustigation etc, else a genuine lack of understanding that I may address - or ignore for other priorities. So if genuinely puzzled or open, please say so....)
Well, the one thing I've noticed that I don't like about LED backlighting but tolorate is the color. Not the K, but the multiple colors. It's something about how they dope the LED with different phosphors to get the full spectrum.
If you've ever pointed a laser at frosted glass or even a wall you know about the glistening effect. Warning: Once you notice this you can't forget it ever! LOL Move left and right while looking at the screen. You can see all the colors glisten from the slight offsets of the different colors. Mostly just on white or solid colors though.
Yeah - I still think alternatives have their place....
I thought the simple diagram with some explanatory info below might help separate previously discussed topologies - ie, it separates voltage driven from current driven (limited) LEDs. Both have been mixed herein when discussing LED dimming....
It is KIS and doesn't expand into higher voltages (for LED pulsation) etc.
[ Though if merely driving from higher than the 3-LED voltage, just add the normal resistor to limit peak current as per normal (since this is about dimming - not multiplexing or pulsating currents). ]
Nor does it even mention average vs peak voltage nor even PWM frequency (usually a minimum of 400Hz), but as a first cut....
Since BES is using voltage control, I thought the preferred (economical) topology would fit right in.....
And thanks to izu for his help & resources. (No - he will NEVER live down his use of MS Paint for these primitive hi-tech diagrams )