Searching \ for '[EE] Challenge:: White LED driver' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/displays.htm?key=led
Search entire site for: 'Challenge:: White LED driver'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Challenge:: White LED driver'
2007\12\03@182910 by Apptech

face
flavicon
face
I'm in the process of implementing the lowest possible cost
white LED driver able to drive 6 LEDs at up to 20 mA from 3
x NiCd/NimH cells.
This isn't rocket science and there are many circuits 'out
there' but few if any meet my needs. I don't THINK I can do
better than what I'm doing, but please prove me wrong :-) .

Output will be close to constant brightness across battery
life with some intentional taper at the end.

There are any number of LED driver ICs available. Most are
priced at too much, to too too too much.
I'm looking at a non conventional solution (or course ;-) ).

Pricing need only be approximate - the brilliance and
manifest low cost of the circuit is the key thing.

**** MUST **** be.

> 80% efficiency across 3V-4V battery range. More better.

All parts in spec at all battery voltages above 3V.

All LEDs in series.

Constant and fully designable LED current within 10% across
3V - 4V battery range.

Low low low parts cost.
Assembly cost not stupid but not as important as low parts
cost.

Always starts when power applied.

___

Other:

Brightness taper in say last 10% of battery charge to
prevent sudden extinguishing from full brightness when flat.

Multiple or variable brightness a bonus.

Always on-power with enable line OK if standby current vvv
low. (< 10 uA?)

Low voltage shutdown to prevent battery being flattened
dangerously a bonus. Costed separately if extra cost.

________

High volume production.
Assembly costs not counted unless liable to be extreme.
Surface mount or through hole OK for this exercise.
Digikey lowest listed pricing for parts used as costing
guide unless better prices can be shown elsewhere.

Glue parts can be assigned sensible price estimates. (eg
small resistors < 1 cent, small zeners ~1c, small bipolars a
few cents or less.
More special parts (eg decent FET switch) at Digikey prices
(at as high a volume as you can find listed).

I will share the essential portions of my final circuit
****IF**** I get enough competitive suggestions to make me
feel it worth the exchange.



       Russell

2007\12\03@211822 by Richard Prosser

picon face
On 04/12/2007, Apptech <spam_OUTapptechTakeThisOuTspamparadise.net.nz> wrote:
{Quote hidden}

How about sharing some of the proceeds instead?   :-)

RP

2007\12\03@215937 by Jinx

face picon face
> > I will share the essential portions of my final circuit
> > ****IF**** I get enough competitive suggestions to make me
> > feel it worth the exchange.
> >
> > Russell
>
> How about sharing some of the proceeds instead?   :-)
>
> RP

I'll cover that if Russell doesn't. I'm excitedly working on
something that could be very lucrative, and for minimal effort.
If it all works out, it's drinks for everyone. Or I might have
you all whacked. Dunno, see how my Christmas goes .....

2007\12\03@220119 by Apptech

face
flavicon
face
>> I'm in the process of implementing the lowest possible
>> cost
>> white LED driver able to drive 6 LEDs at up to 20 mA from
>> 3
>> x NiCd/NimH cells.
>> This isn't rocket science and there are many circuits
>> 'out
>> there' but few if any meet my needs. I don't THINK I can
>> do
>> better than what I'm doing, but please prove me wrong :-)
>> .

> How about sharing some of the proceeds instead?   :-)

The proceeds for the actual design are liable to be minimal
for various reasons :-).
More on that anon after certain things are finalised.

And, as I said, I think my existing design will do what I
want as well as any alternatives. But, I may be wrong [tm].

I saw it more as an idea to share ideas that may be useful
to everyone, but I'd like other people to contribute rather
than just publishing a circuit.

My "low power switching regulator challenge" of some years
ago resulted in the now somewhat famous minimalist parts
count Black converter. "My" 3 transistor GSR has
substantially better performance specs but his does OK for
many applications and uses 2 transistors. And it in turn was
based on a brilliant (albeit only retrospectively obvious) 2
transistor constant current relay driver circuit.

And, if I recall correctly, the original relay driver
circuit was your suggestion. So, lets see your latest LED
driver idea :-).

Mine uses the almost lowest price IC that I'm aware of,
fwiw.



       Russell


2007\12\03@221511 by Richard Prosser

picon face
part 1 3347 bytes content-type:text/plain; charset=ISO-8859-1 (decoded 7bit)

Hi Russell,
How about something like the attached.

Based on the same circuit that the "Black Regulator" derived from.

I've simulated it at about 90% efficiency but real components will give you
less.
The transistors can just be general purpose types.
Some of the values could possibly be optimised a bit further, and a few
extra components can stabilise it better over a wider voltage range. I'd
have to dig out some old notes on this but it was pretty simple. Stabilising
for temperature was more difficult but is less of an issue here I think.
I'm not sure about EMI issues - it worked OK as a basic relay driver with
some additional filtering but if the LEDs are distant from the driver their
wires are carrying fast-ish edges.

More LEDs can be added iin series as required up to the voltage limit of Q1.

Richard

On 04/12/2007, Apptech <.....apptechKILLspamspam@spam@paradise.net.nz> wrote:
{Quote hidden}

> -

2007\12\03@225318 by James Newton

face picon face
My coworkers are relieved to find that it was only my sense of humor and not
an epileptic fit that had me sliding out of my chair when I read that...

Jinx, you almost saved the cost of having me whacked; I about died laughing.

Midway through the first bout, it hit me that Jinx could play the janitor on
"Scrubs"

--
James.

{Original Message removed}

2007\12\03@232056 by Nate Duehr

face
flavicon
face

On Dec 3, 2007, at 7:59 PM, Jinx wrote:

> I'll cover that if Russell doesn't. I'm excitedly working on
> something that could be very lucrative, and for minimal effort.
> If it all works out, it's drinks for everyone. Or I might have
> you all whacked. Dunno, see how my Christmas goes .....


Jinx Soprano?  :-)

--
Nate Duehr
natespamKILLspamnatetech.com



2007\12\03@235504 by Apptech

face
flavicon
face
> How about something like the attached.
>
> Based on the same circuit that the "Black Regulator"
> derived from.
>
> I've simulated it at about 90% efficiency but real
> components will give you
> less.

Richard

That's an excellent start to the challenge.
Gives my IC based circuit a run for its money.
I'll try it after I have mine working fully.

I suspect that the current regulation may be poor, but
adding eg a reference diode (more cost) in place of the
4148's would help that.


       Russell




{Original Message removed}

2007\12\04@021050 by W. Jacobs

flavicon
face
Hello All,
I have seen several question about driving LED's in series and was
wondering what the failure condition for them was.  If one led in a
series string of several quits, the whole string will quit.  Would it
not be better to have several strings of fewer led's.
If you have 2 strings of 12 and 1 led fails, you loose 1/2 of your
array.  If you had 8 strings of 3 and 1 went out you would still have
87% operational.  I was wondering if this was important

Thanks
bill


Richard Prosser wrote:
{Quote hidden}

2007\12\04@023550 by William \Chops\ Westfield

face picon face

On Dec 3, 2007, at 11:11 PM, W. Jacobs wrote:

> wondering what the failure condition for them was.  If one led in a
> series string of several quits, the whole string will quit.

Not necessarily true.  I have a series string of LEDs in different
colors as a night-light in one of the kids rooms.  Individual LEDs
have failed (output of a "purple" LED went to essentially zero), but
the string as a whole stayed lit...  (I concluded that I don't really
understand the failure modes of LEDs... :-)

BillW

2007\12\04@024417 by Richard Prosser

picon face
Bill,
I haven't seen that many LEDs fail without them being abused. If it is a die
(overtemperature / overcurrent / static ) failure then I guess they'll fail
like most diodes - short circuit. Otherwise a bond wire failure will leave
it open.  Either way you can loose unless you double up on the drive circuit
also.

RP

On 04/12/2007, W. Jacobs <EraseMEwh2007spam_OUTspamTakeThisOuTcitlink.net> wrote:
{Quote hidden}

> -

2007\12\04@024809 by Richard Prosser

picon face
On 04/12/2007, Apptech <@spam@apptechKILLspamspamparadise.net.nz> wrote:
{Quote hidden}

more stable reference would certainly help.
Turning the circuit upside down is also quite feasible & would let you use
an NPN output transistor - or an N channel fet.

Richard

2007\12\04@031032 by Jinx

face picon face
> Jinx Soprano?  :-)

Hehe. I see myself more in the Joe Pesci mold -

"What do ya mean, funny? Let me understand this cause, I don't know
maybe it's me, I'm a little ****** up maybe, but I'm funny how? I mean,
funny like I'm a clown, I amuse you? I make you laugh... I'm here to
*****' amuse you? What do you mean funny, funny how? How am I
funny?"

Scary scary man

But I shouldn't be hobnobbing on Russell's thread. Sorry R

"This...Basil"
"This...Basil's wife"
"This...smack on head"

2007\12\04@034211 by Justin Richards

face picon face
I too had a giggle, then my thoughts turned to Jinx having a fantastic
Xmas.  Heres to a very very merry Xmas Jinx.

How do I remove myself from that PICLIST map ...

But back to the LED driver, I have tried to get an understanding of
how these 'simple' ccts actually work.

I tried to follow what would happen when the cct was powered up and I
cant see how the LEDs would ever light up.

Are the LEDs around the wrong way.  Can anyone do a quick walk thru of
how this cct would work.  Any enlightening will be greatly
appreciated.

Cheers Justin

On Dec 4, 2007 5:09 PM, Jinx <KILLspamjoecolquittKILLspamspamclear.net.nz> wrote:
{Quote hidden}

> -

2007\12\04@043013 by Apptech

face
flavicon
face
> I tried to follow what would happen when the cct was
> powered up and I
> cant see how the LEDs would ever light up.

> Are the LEDs around the wrong way.  Can anyone do a quick
> walk thru of
> how this cct would work.  Any enlightening will be greatly
> appreciated.

LEDs are as Richard showed them because it works by
"ringing" inductor negative.

STARTUP
R3 D12 D11 applies about 1.2V to Q2 base via R8.
As no current in R2 then Q2 emitter at ground so drive via
R8 turns it on.

Q2 on pulls Q1 base low via R5 and C2. C2 is a "speedup" cap
which causes voltage change at Q2 C to appear
instantaneously on Q1b.
Q1 on applies V to L1 top.
Current in L1 rises approx linearly with time.

Current in L1 flows in R2 so Vr2 rises.
Q2 will turn off when Q2Vbe falls below about 0.6V so as Vr2
rises Vbeq2 falls until it reaches 0.6V. [It will actually
clamp at about 0.6V when there is enough drive and excess
voltage appears across R8 providing base drive].

Q2 turning off turns off Q1 interrupts current in L1.

The universe demands that the current in L1  must not be
interrupted so polarity of L1 reverses and rises until
current flow continues unchanged - this will cause VL1 top
to ring negative until IL1off = IL1 on. The LEDs are the
sink for this current.

When energy in L1 is transferred to LEDs such that Ir2 falls
low enough the cycle repeats with Q2 turning on.

R6 is the magic bullet without which this will tend to turn
into a linear regulator.

Hysteresis is introduced by R6 which lowers Q2b due to the
R8-R6 divider when L1 top is negative. This means that the
circuit will conduct until Il1 falls below its original
value by an amount which can be defined by this divider.

Current accuracy is limited by the accuracy of the
comparator formed by the VD12 + VD11- VbeQ2 reference in
which is compared with the VR2 input signal. Vd12 and Vd11
and indeed Q2vbe are temperature dependent and also somewhat
soft kneed (if not actually lily livered)

Haven't tried it yet but hope to real soon now.
The Black voltage regulator, based on this circuit, has
somewhat poor voltage regulation  with Vin so this may too.
We'll see.


My circuit is using a CD40106 Schmitt inverter package
($US0.08 ish in high volume) and works OK although so far
slightly less well than I'd like. About 80% efficiency. I
want more than that. Using an N Channel MOSFET as the
switch. Schottky output diode. Glue will cost more than IC.
Playing ... .





       Russell



2007\12\04@044758 by Richard Prosser

picon face
Justin,
1. Q1 & Q2 are initially turned on, and current flows though the inductor L1
& R1
2. As the current increases, the voltage across R1 increases. Once a
threashold voltage is reached (about 1 diode drop below the base voltage of
Q2) then Q2 starts to turn off.
3. This turns Q1 off. The voltage on the collector of Q1 drops pulling the
base of Q2 lower via R6 and turning it off completely.
4. Since the current through L1 can't immediately stop, the voltage at Q1
collector drops below zero volts to whatever voltage is required to turn the
diodes on. This is a straightforward flyback type effect.
5. The current though L1, R1 and the diodes then decays as the energy stored
in the inductor is used until the voltage across R1 drops sufficiently for
Q2 to start tuning on again.
5. This turns Q1 on, the collector voltage rises, turns Q2 on harder, and
starts the supply current flowing through L1 & R1 again.
6. The process repeats.

Now - how best to stabilise the average diode current with changing input
voltage?

RP


On 04/12/2007, Justin Richards <RemoveMEjustin.richardsTakeThisOuTspamgmail.com> wrote:
{Quote hidden}

2007\12\04@054511 by Apptech

face
flavicon
face
> Now - how best to stabilise the average diode current with
> changing input
> voltage?

TL431 does a good job. Cheap but the most expensive part
except maybe the inductor and the LEDs.

Chained zener-resistor shunt regulators sort of work.



       Russell




2007\12\04@142240 by Richard Prosser

picon face
Ah, but that's only part of the problem. The positive side of the hysterisis
also changes with input voltage.

Coincidentally, I'm looking at trying to design a "constant power" circuit
to discharge a capacitor (33mF, 75V) quickly . Somewhat similar requirement.

RP


On 04/12/2007, Apptech <TakeThisOuTapptechEraseMEspamspam_OUTparadise.net.nz> wrote:
{Quote hidden}

> -

2007\12\04@202044 by William \Chops\ Westfield

face picon face

On Dec 4, 2007, at 1:28 AM, Apptech wrote:

> LEDs are as Richard showed them because it works by
> "ringing" inductor negative.

Huh. Could you improve efficiency by connecting the
LED's Anode to +Vin instead of GND, or doesn't it matter?

Is that induction 3.3 millihenries (rather large) or 3.3
microhenries (seems too small.)  At 3.3millihenries, I
suspect a large part of your efficiency would be limited
by finding an inductor with an appropriate low resistance.

(so, I've struggled with this switching regulator concept
for a while now - How does one pick an inductor value?  My
current understanding is that if you have ideal components,
the inductor value only affects the frequency of operation
of the circuit, and in theory "any" value would work.  With
"real" components you have to match of the frequencies and
maximum currents and such to fit the actual components, but
it still doesn't matter as long as you can stay within those
bounds.  Is that approximately correct?)

BillW

2007\12\04@223843 by Richard Prosser

picon face
I started with a 10mH choke and then realised I "had" some SMD 3.3mH chokes
available. Except I later realised that they were actually 3.3uH and didn't
work in the circuit as shown. The inductor is probably the weak part of the
circuit in that is has to be reasonably large for the circuit to operate. I
guess that a smaller inductor could be made to work but it would require
faster transistors - or biasing them so they were not saturated - which
would increase losses.

Returning the LEDs to the +ve line doesn't appear to improve efficiency.

RP


On 05/12/2007, William Chops Westfield <RemoveMEwestfwspamTakeThisOuTmac.com> wrote:
{Quote hidden}

> -

2007\12\05@023517 by Apptech

face
flavicon
face
> Ah, but that's only part of the problem. The positive side
> of the hysterisis
> also changes with input voltage.

Split R6 into two. Take a diode from it's midpoint (anode)
to the reference diode (D11/D12). This clamps the midpoint
at 1 diode drop above reference and "stabilises" the on
drive through R6_left. Two diodes or lower R6_left = more
drive.

BUT somewhere about here you say enough is enough and
Replace Q2 with two transistors in a long tailed pair. The
reference feeds Q2_left_base  and the sense signal drives
Q2_right_base with any hysteresis added to (probably)
Q2_right. This was done by ?Natsemi? long ago. Different
than either 'my' GSR or your/Roman's approach. Proper
comparator should be much more regulated.

*BUT* unfortunately, the thing that makes this circuit
almost certainly unsuited to me need is that, as didn't
occur to me in my initial enthusiasm, the "regulation" at
present occurs on the input and not the output side. ie the
ocillation occurs when Iin reaches a limit set by the
voltage across R2. At that stage the inductor energy is
dumped into the LEDs (and a series diode with a capacitor on
the KLED side of it would make the LEDs' life more gentle)
and the cycle then repeats. You COULD add LED current
sensing to make the process more defined but for now it's a
bit open loop. I need something with reasonably well defined
LED drive current over the battery operating range and
suspect that this would be hard to tame. Maybe not
impossible though. My CD40106 circuit uses a $US0.08 (25000
volume IC) and 3 transistors (2 are FET gate drivers, one is
for LED current sensing) so is certainly dearer. Efficiency
is excellent.

> Coincidentally, I'm looking at trying to design a
> "constant power" circuit
> to discharge a capacitor (33mF, 75V) quickly . Somewhat
> similar requirement.

Why "constant power"?
Would constant current do?


       Russell

2007\12\05@045352 by Richard Prosser

picon face
On 05/12/2007, Apptech <apptechEraseMEspam.....paradise.net.nz> wrote:
{Quote hidden}

Fair comment re R6 but it's not the complete story either - there is another
mechanism involved I've yet to track down.

Re the load on the LEDs - the maximum current is only what current is
passing through the inductor at the trip point, so if tjis is set at a
sensible level, there should not be a problem.

Constant resistance is simplist but Since the current drops during the
discharge, the peak power is higher for a given discharge time.
Constant current is better, but the power disapation is shared between an
active device and  a load resistor (in the simplist case). The  power
dissipated also drops as the voltage falls - so a larger load capability is
still required.

Constant power minimises the peak power requirement, offers the chance to
use smaller active devices as they are operated `in switching mode and the
load power is constant so doesn't need to handle a high peak.

It's just an idea at present but looks like I could use a 10watt load
resistor instead of a 25watt one - or something like that. A big price
difference.

One of the specs for the product is an MTBF of >100years so I'm a bit wary
of overrating things too much, even on current pulses. Cost is less of a
problem in this case but profit is profit so a cheaper build price for the
same performance is preferred.

I should have time tomorrow to look at some of the details again - was too
busy today with an urgent quote.


Richard

2007\12\13@193025 by Justin Richards

face picon face
Thanks Russell/Richard for taking the time to explain its operation.
I assume R1 is R2.

Thanks Again

Justin

On Dec 4, 2007 8:17 PM, Richard Prosser <EraseMErhprosserspamgmail.com> wrote:
{Quote hidden}

>

2007\12\13@202929 by Richard Prosser

picon face
Yep - sorry about that, R1 seems not to be there.

RP


On 14/12/2007, Justin Richards <RemoveMEjustin.richardsTakeThisOuTspamspamgmail.com> wrote:
{Quote hidden}

> >

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...