I sure wish you would answer ALL my questions...

My apologies, I thought I'd answered earlier about what was driving what....

PIC PWM pin, driving a MOSFET gate, mosfet source is to ground, moseft drain is to LED cathode, LED anode connects to 3.3V

but here's a schem I've just knocked together for you....

 

PWM is the way to go for best efficiency if you have an inductor in series with the PWM output to the LEDs to average the current. Alternaltely, you can add a resistor in series with the LEDs at reduced efficiency. Without a resistor or inductor you will get high peak currents which gives a high apparent brightness even at the minimum PWM duty-cycle, which is what you experienced.

So add a resistor in series with the LEDs to limit the current and you will get fine resolution control of the LED brightness. Use a resistor value that gives the maximum desired LED current at 100% duty cycle (completely on). For best current balancing you should have an individual resistor in series with each LED.

 

thanks crut, that was very good info u just posted.

I am just wondering....or is it me..!

Silly me.... but the first post says that and it ended up with this.

This thread is way confusing..!

 

This thread is way confusing..!

Actually, the thread started out asking if such a thing as a low voltage voltage follower that can drive heavy loads exists, but it seems to have taken a detour with everyone talking about my present implementation (which I want to move away from). As it goes, I'm *really* pushed for pcb real estate (which rules out all the extra resistors) and an inductor in the circuit is a no no to, so I may be forced to live with the higher than preferred brigtness at 1/1024th PWM!

nevertheless thanks to all for your contributions.

 

Resistors on each LED are more than just a good idea.

Best practices are best for a reason.

If you would have had room for an Op Amp then you should have room for resistors nets.

 

If you simply lowered your PWM frequency, much of the dimming problem would go away. At 1/1024, the pulse width is 62.5nS, but the turn-on delay is 4.5nS, and the turn-off delay is 45nS. That tacks on another 40.5 seconds to the 1/1024 ON-time - not to mention extra time required to charge/discharge the 3.9nC total gate charge. Decreasing the frequency means that the charge/discharge time would not be as critical, and the difference between the turn-on and turn-off times would be less significant.

However, you still need a way to limit the maximum current through each LED. PWM alone won't take care of that.

 

This thread is way confusing..!

I concur. It is not at all what it looked like at first post.

 

If you simply lowered your PWM frequency, much of the dimming problem would go away. At 1/1024, the pulse width is 62.5nS, but the turn-on delay is 4.5nS, and the turn-off delay is 45nS. That tacks on another 40.5 seconds to the 1/1024 ON-time - not to mention extra time required to charge/discharge the 3.9nC total gate charge. Decreasing the frequency means that the charge/discharge time would not be as critical, and the difference between the turn-on and turn-off times would be less significant.

I can't lower the PWM frequency due to audible PWM bleed into a low level audio circuit (those who have ever dabbled with any circuit involving a low frequency clock & low level audio can probably empathize - this stuff permeates everywhere )... so I'm constrained to using a frequency at the upper end of the hearing range.

re using resistors....if you look at the chart for most high brightness 3mm LED, up until its rated FWD voltage the current is quite modest, so running them below their rated fwd volatge should be of no real concern. I realise everyone fixates about using a current limiting resistor...but the supply voltage really limits the current too!

 

The forward voltage of an LED changes when its temperature changes. With a voltage source then it heats itself which causes more current if it is missing a current-limiting resistor which causes more heating which causes more current which causes more heating which causes more current which causes .... It is called thermal runaway.

Why don't you use current-limiting resistors like everybody else?

 

Why don't you use current-limiting resistors like everybody else?

because I'm running the LEDs *under* their rated fwd voltage ...unlike everyone else!

Re thermal impact ....yes I know LEDs have a negative temperature co-efficient, but having run these leds for several hours, I didn't see any significant change in the current draw as measured at the beginning & at the end.

 

News flash: LEDs are not voltage operated devices. They are current driven. The proper use is to give them current proportional to the brightness you want and let the breakover voltage be what it turns out to be for that batch of LEDs.

 

Stop Press: LEDs get less bright when you supply them with less voltage vs. their rated fwd voltage - this is because they draw less current ....





now I agree...you wouldn't want to be pushing an LED at or over its rated fwd voltage without a resistor, but really, I'm not pushing the envelope with respect to LED 'in danger' territory.

Yes, I know, LEDs current driven, yes I know about the negative temp co-efficient, yes I know best practise = current limiting resistor ....but hey, sometimes you *can* actually configure/implement stuff in other ways ...and do you know what? ...yep, they still work just fine!

You're all trying to fix something that isn't broken! (and coming across a tad patronising to boot)

So even though there's much gnashing of teeth around these parts at the audacity of it all ...the resistors are staying out!

it all reminds me of the well meaning teenager that grabbbed an elderly lady's arm who was standing on the kerbside & walked her across the road...he never thought to ask her if she wanted to cross the road at the outset (she didn't)

the irony here is I came here asking if a low voltage 'voltage follower' existed (ie that could drive heavy loads) & now I'm getting all this sarcastic 'news flash' crap.

 

and yet, you're still having trouble getting the brightness levels you want. I wonder what's causing that.

 

Your graph has a linear voltage scale and a linear current scale.
That is wrong because they should be exponential, not linear.

Datasheets show an exponential graph.
Maybe you did not look at their datasheet.

 

and yet, you're still having trouble getting the brightness levels you want. I wonder what's causing that.

I'm completely happy with my overall LED brightness levels, but I'd prefer the apparent visual brightness from PWM step 1/1024 to 0/1024 to be smaller. As much as you're wallowing in the patronising type posts, I actually think you are getting over fixated on the current ...these are high brightness 3mm led with optics arranged to hugely 'up' the apparent brightness even at the lowest of drive settings. I've read in many places that you need 16 bits to be able to fully tame an LED at lower brightness .....so that would be 1/65535th.

As a little diversion, I took the Voltage down to 2.5V (this is for a 3.4V LED)....even at 1/1024th the LED was still visibly on ...in other words...cack all current (just 1/1024th of a few mA), so I actually don't think it's current here that's my problem at all...more the characteristic of a 3mm domed top high brightness led....and since I can't go to 16 bits PWM (they don't do 8 pin PICS in any more than 10 bit PWM), I'm now seeking creative aways to remedy...and I'm pretty a resistor isn't it! (sure it'll likely curtail the low end brightness, but it'll take the top end brightness down too)

 

......................
the irony here is I came here asking if a low voltage 'voltage follower' existed (ie that could drive heavy loads) & now I'm getting all this sarcastic 'news flash' crap.

You came to us with a problem and we suggested the proper way to solve it. It's not crap. You can not reliabily drive LEDs with a voltage source.

As my old 8th grade teacher used to say, "You can lead a horse to water, but you can't make him drink".

 

You came to us with a problem and we suggested the proper way to solve it. It's not crap. .

No I came here asking if such a thing as a voltage follower with high current drive existed....that was my problem (because such a device will actually solve my underlying problem - a resistot won't) then had many an unrelated solution down stuffed my throat (& in a patronizing way too!) ...nice.

If you go into a shop asking for a loaf of bread would you not get a little fed up if the shopkeeper said "What you want it for? To eat? No, no you don't want bread it makes you fat...that's well known by everybody around these parts, are you stupid - what you need is a slice of celery"

As my old 8th grade teacher used to say, "You can lead a horse to water, but you can't make him drink".

Did he also say "you can ask a forum a particular direct question..but then have lots attempt to patronize you with a inappropriate replies"

Anyway, thanks to those who attempted to answer my original question early on (you know, the one about the voltage follower) & resisted the urge to go "no, no, no, you don't want to do it like that".

 

I will repeat that a small SIP Resistor Net can't possibly take more room on the circuit board than the components for your voltage follower.

You seem to have tunnel vision. You are in love with a circuit idea and will pursue it no matter how much simpler and better the other ideas are.

No point in bothering really because you don't seem to be listening, but it is amusing to see that I am not the only person who has ever assumed with full faith something that was off base.

 

Peskywinnets, you are new here. You have no idea how many noobs come here asking how to make what they think they need to solve their problem, when their problem is either something else altogether, or they have taken the wrong approach to solving their problem. We had no idea what your level of expertise was when you started this thread. Everyone was trying to be helpful.
I agree that current control is the way to drive an LED. However, I continued to pursue the high current, rail-to-rail op amp idea, running some simulations, and I don't believe you will find such an animal.
A switching regulator might solve your problem, but it wouldn't fit on your circuit board. I'm afraid you are up the proverbial creek without the proverbial paddle. Hopefully, I'm wrong.

 

No I came here asking if such a thing as a voltage follower with high current drive existed....that was my problem (because such a device will actually solve my underlying problem - a resistot won't) then had many an unrelated solution down stuffed my throat (& in a patronizing way too!) ...nice.
.............

The problem is, what you perceive to be the solution to your underlying problem isn't a good one. And you refuse to believe when you are told a resistor will work to solve your problem. I don't see how that's patronizing.

As obviously the proper solution can't be stuffed down your throat (hence my horse analogy), I'll leave you to those who still support your preferred path.