Searching \ for '[EE] 2-level constant-current source?' 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/power.htm?key=current
Search entire site for: '2-level constant-current source?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] 2-level constant-current source?'
2005\08\05@143146 by PicDude

flavicon
face
Hi all,

I need to supply power to a large number of LED's in a vehicle, where all
LED's will be on or off simultaneously, and with 2 brightness levels.  I also
want to use a constant-current source for this.  With the matrix I'm planning
to use for the LEDs (14 parallel strings of 6 series LEDs each), my initial
high-level calcs are ~210mA normal and 350mA when the brakes are pressed.

To keep the whole thing relatively small and low-cost, the good old generic
LM317 comes to mind, but the question is how to set up a single LM317 for
dual current levels.  I came up with this circuit, though I'm not sure if it
would work properly...
       http://www.narwani.org/neil/electronics/Const-2Cur-Ckt.gif

Any ideas?  Is there a better way to achieve this?


Cheers,
-Neil.


2005\08\05@152249 by Spehro Pefhany

picon face
At 01:31 PM 8/5/2005 -0500, you wrote:
>Hi all,
>
>I need to supply power to a large number of LED's in a vehicle, where all
>LED's will be on or off simultaneously, and with 2 brightness levels.  I also
>want to use a constant-current source for this.  With the matrix I'm planning
>to use for the LEDs (14 parallel strings of 6 series LEDs each), my initial
>high-level calcs are ~210mA normal and 350mA when the brakes are pressed.

I'm guessing you'll find that to not be enough of a difference to be
visually acceptable, but that's another issue...

>To keep the whole thing relatively small and low-cost, the good old generic
>LM317 comes to mind, but the question is how to set up a single LM317 for
>dual current levels.  I came up with this circuit, though I'm not sure if it
>would work properly...
>         http://www.narwani.org/neil/electronics/Const-2Cur-Ckt.gif
>
>Any ideas?  Is there a better way to achieve this?

I suppose you need the low side of the LEDs to be grounded...

Do you have enough voltage for the LM317 to work (about 3V on top of the
LED forward voltage)?

The circuit you drew shows the 2N3904 carrying pretty much the entire
load current when it is on (350mA). I think the Vce saturation voltage
will be too high relative to the reference voltage (1.25V) to get very
consistent results, even if you put a fair bit of current into the base.

You could do it with a MOSFET, resistors, zener (for gate protection) and
a 2N3904, but you could also just build two LM317 current sources and
parallel them. If the voltage drop isn't an issue.

A tiny bit more complex, but you could make an accurate and very compliant
(say within 100-200mV of the supply rail) high-side current
source with a R-R in/out op-amp, an LM431 and a power BJT. Getting the
two levels could be done similarly to your circuit (2N3904 or whatever BJT),
but with an inverse diode on the base for protection, and it would need a
very small base current because it would only be working with the reference
voltage rather than the load current.

Or maybe you could just use resistors and forget the sophisticated stuff,
but careful that there's enough voltage if the engine fails to give a good
brake brightness.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
spam_OUTspeffTakeThisOuTspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com




2005\08\05@155452 by Wouter van Ooijen

face picon face
> To keep the whole thing relatively small and low-cost, the
> good old generic
> LM317 comes to mind, but the question is how to set up a
> single LM317 for
> dual current levels.

don't, those things are dirt cheap. use two :)

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\08\05@155741 by Dwayne Reid

flavicon
face
At 12:31 PM 8/5/2005, PicDude wrote:
>Hi all,
>
>I need to supply power to a large number of LED's in a vehicle, where all
>LED's will be on or off simultaneously, and with 2 brightness levels.
>
>  I came up with this circuit, though I'm not sure if it
>would work properly...
>         http://www.narwani.org/neil/electronics/Const-2Cur-Ckt.gif

What is the voltage at the top of the LEDs?  In other words, how much base
current will you have when the brake line goes HI?

If these LEDs have the usual 1.7V forward drop each, your circuit should
work just fine.  Base resistor value needs to be fairly low or, better yet,
perhaps consider the use of a logic-level FET instead of the transistor.

dwayne

--
Dwayne Reid   <.....dwaynerKILLspamspam@spam@planet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 21 years of Engineering Innovation (1984 - 2005)
 .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
    `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

2005\08\05@160946 by olin piclist

face picon face
PicDude wrote:
> I need to supply power to a large number of LED's in a vehicle, where
> all LED's will be on or off simultaneously, and with 2 brightness
> levels.  I also want to use a constant-current source for this.  With
> the matrix I'm planning to use for the LEDs (14 parallel strings of 6
> series LEDs each), my initial high-level calcs are ~210mA normal and
> 350mA when the brakes are pressed.
>
> To keep the whole thing relatively small and low-cost, the good old
> generic LM317 comes to mind, but the question is how to set up a single
> LM317 for dual current levels.

Sounds like a job for a 10F204.  You provide an analog level, and it runs a
boost converter to maintain that voltage accross a low side current sense
resistor.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\08\05@163035 by Harold Hallikainen

face picon face
If you use two, you can just switch the DC to each of them to select the
current (you could get a total of 4 current levels, including off, if one
is set to twice the current of the other).

I recently used an LM317 as a current source and needed to turn it on and
off. I put a resistor between the sense pin and the "right side" of the
current sense resistor. Use an open collector to drive the sense pin. With
the collector open, the LM317 current regulates. When the collector is
pulled to ground, the LM317 regulates to 1.25V plus the transistor
saturation voltage. In this case, that was less than the Vf of the LEDs,
so they went off.

Harold

{Quote hidden}

--
FCC Rules Updated Daily at http://www.hallikainen.com

2005\08\05@170504 by PicDude

flavicon
face
On Friday 05 August 2005 02:29 pm, Spehro Pefhany scribbled:
> I'm guessing you'll find that to not be enough of a difference to be
> visually acceptable, but that's another issue...

I'm not sure until I get the LEDs, but used the ratio from another LED
taillight I did in the past (much less LEDs and only dropping resistors).  
Should be easily adjustable though.  The 3.3-ohm resistors shown are just for
this discussion, I am actually planning to use 2 series resistors for each of
those so I can better tune the resistance around that low value.


> I suppose you need the low side of the LEDs to be grounded...

Preferably.  One of the ideas I thought of was to put R101 at the common
ground of the LEDs instead with the transistor bridging that, but then
realized that a constant-current source would automatically adjust that out.  
Doh!


> Do you have enough voltage for the LM317 to work (about 3V on top of the
> LED forward voltage)?

Per the datasheet, each LED has 1.7V typical, which adds up to 10.8V per
string of 6.  I was looking at the dropout specs on the LM317 datasheet this
morning which showed a dropout of ~1.6-1.8V for this current and expected
temps.


> The circuit you drew shows the 2N3904 carrying pretty much the entire
> load current when it is on (350mA). I think the Vce saturation voltage
> will be too high relative to the reference voltage (1.25V) to get very
> consistent results, even if you put a fair bit of current into the base.

Oops/FWIW, I used the default 2N3904 in the Eagle library, but usually use a
PN2222A, which IIRC can handle much more current.  Interesting about the
VceSat -- never thought about that -- excellent point.  I did search Digikay
for small relays for this, but nothing small enough.


> You could do it with a MOSFET, resistors, zener (for gate protection) and
> a 2N3904, but you could also just build two LM317 current sources and
> parallel them. If the voltage drop isn't an issue.

I actually had it designed with 2 LM317's earlier, but was hoping this would
be easier and smaller.  I also since realized that the LM317 now comes in
packages other than TO-220 and TO-3 so 2 of them is not really a problem
anymore.  Just need to do the math for heat-dissipation to see if the SOT-223
or D-Pak versions will handle it.

BTW, first idea was the easy 28 resistors, but while doing that math I
realized I'd be in for problems when I remembered that at a buddy's shop
recently, he put one of those huge commercial battery chargers on the car's
battery and the voltmeter went to ~18V!

Other ideas were to use an adjustable boost regulator to say 20V or 30V and
have the brake signal change one of the reference resistor values, but I
thought it unnecessary to start adding inductors, large caps, etc.  I really
want to keep this board as flat as possible.


> A tiny bit more complex, but you could make an accurate and very compliant
> (say within 100-200mV of the supply rail) high-side current
> source with a R-R in/out op-amp, an LM431 and a power BJT. Getting the
> two levels could be done similarly to your circuit (2N3904 or whatever
> BJT), but with an inverse diode on the base for protection, and it would
> need a very small base current because it would only be working with the
> reference voltage rather than the load current.

Wow, that actually makes sense to me, in spite of my amateur op-amp knowledge.  
But unnecessarily complex.


Cheers,
-Neil.


{Quote hidden}

2005\08\05@170655 by KY1K

picon face
I suspect I've done the identical project here already, so I might be able
to make a suggestion or two.

First, the difference between the low and the high current isn't high
enough. If you need 210 ma to make the running lights visible in daylight
at high noon, then you need about 3.5 times that current to make the brake
lights stand out when you step on the brakes.

I think you will find that the dual filament lamp used for running lights
and brake lights has about the same ratio between the running light and the
brake light current.

For my project, I used a modified 12 volt to 3.3 volt switching regulator
from our old cell phone. It had an MC34063 switching regulator in it, which
is already set up to run on 12 volts and can be configured as a step up or
step down switching regulator. It is good for 175 ma output at up to 40
volts WITHOUT using an external transistor.

In my case, I used 4 strings of LED's (only 4 limiting resistors are
needed, one for each string). I used a  very small mosfet to short out an
added resistor in series to the voltage feedback pin so that the power
supply supplies 29 volts for running lights only and around 33 volts for
brake lights and running lights together.

I'm running 40 ma for running lights only and 150 ma for running and brake
lights.

The switching regulator runs very warm, but it does run continuously. I'd
feel a little more comfortable if it was running a little cooler, so I
might add the transistor to the switching regulator to dissipate the extra
heat.

I was also wondering about using a photocell to adjust the operating
current for night time operation......the LED's are brighter than the lamp
type lights are in the daytime. At night, they are way to bright and could
be just as visible with much less current.

Enjoy.

Art


and At 02:31 PM 8/5/2005, you wrote:
{Quote hidden}

>

2005\08\05@171328 by PicDude

flavicon
face
On Friday 05 August 2005 02:54 pm, Wouter van Ooijen scribbled:
> > To keep the whole thing relatively small and low-cost, the
> > good old generic
> > LM317 comes to mind, but the question is how to set up a
> > single LM317 for
> > dual current levels.
>
> don't, those things are dirt cheap. use two :)

See, I was afraid that everyone was gonna say this, but was hoping the simple
arrangement I came up with would be smaller/simpler.  But I'm getting
convinced to just do the 2 LM317's now that I also found out that they're
available in packages other than TO-3 and TO-220's.

Cheers,
-Neil.




2005\08\05@172146 by PicDude

flavicon
face
Yes, the spec sheet says 1.7V typical, so 10.8V per string.  I had not
calculated the brake current yet, but I figured as long as there is enough
base current to saturate the transistor, there should not be a problem.  
Spehro brings up a good point re: VceSat though, as it is significant
compared to Vref of the LM317.

Cheers,
-Neil.



On Friday 05 August 2005 02:57 pm, Dwayne Reid scribbled:
{Quote hidden}

2005\08\05@172958 by Spehro Pefhany

picon face
At 04:04 PM 8/5/2005 -0500, you wrote:


> > I suppose you need the low side of the LEDs to be grounded...
>
>Preferably.  One of the ideas I thought of was to put R101 at the common
>ground of the LEDs instead with the transistor bridging that, but then
>realized that a constant-current source would automatically adjust that out.
>Doh!

;-)

> > Do you have enough voltage for the LM317 to work (about 3V on top of the
> > LED forward voltage)?
>
>Per the datasheet, each LED has 1.7V typical, which adds up to 10.8V per
>string of 6.  I was looking at the dropout specs on the LM317 datasheet this
>morning which showed a dropout of ~1.6-1.8V for this current and expected
>temps.

*Plus* Vref (1.25V nominal), in this configuration, yes?

<snip>
>Wow, that actually makes sense to me, in spite of my amateur op-amp
>knowledge.
>But unnecessarily complex.

< 15 parts counting series diodes and 3 caps, not insane. And very resistant
to crud on the 12V using a 100V transistor and shunt reference/power supply.
And 150mV minimum rather than 3V.

If you go with the obvious LM317s, careful about sneak currents back through
the regulators. The series diodes will prevent it.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
.....speffKILLspamspam.....interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\08\05@173140 by Russell McMahon

face
flavicon
face
>>I need to supply power to a large number of LED's in a vehicle,
>>where all
>>LED's will be on or off simultaneously, and with 2 brightness
>>levels.

>>  I came up with this circuit, though I'm not sure if it
>>would work properly...
>>         http://www.narwani.org/neil/electronics/Const-2Cur-Ckt.gif
>
> What is the voltage at the top of the LEDs?  In other words, how
> much base current will you have when the brake line goes HI?
>
> If these LEDs have the usual 1.7V forward drop each, your circuit
> should work just fine.  Base resistor value needs to be fairly low
> or, better yet, perhaps consider the use of a logic-level FET
> instead of the transistor.

Above is good, but more thought is needed yet.

If this is automotive you have say 12v available. usually more a motor
running when brakes are usefully used but allow some wiring drop etc.

LM317 is far from being an LDO alas - drop across regulator is 1.75v
semi-worst case at 200 mA, possibly 2v in some applications (see data
sheet) . (Depends on how cold you want the mornings to be able to be
;-) ). lets say 1.7v v. Then there's the 1.3v worst case drop across
the sense resistor (see data sheet). These two add up to 2.05v
available for LEDS. Next look at your LEDs and see what the WORST CASE
drop is at rated current. May LEDs you can run is < (12-2.05)/Vled.
You mention series strings of 6 LEDs which would allow (12-2.05)/6 =
1.7v / LED which MAY be enough. If you give the LM317 some headroom
you may need 5 or even only 4 per string. Or select a good LM317 on
test :-)

The switched resistor idea looks good, although there can always be
something one misses.

Using two LM317's makes the job very simple.

The idea of using a boost SMPS adds complexity but makes most other
aspects simpler. Art's MC34063 is cheap and versatile. Also
inefficient compared to some modern alternatives but good enough for
this task, as he says. A modern White LED series-LED IC would do the
whole task BUT may want a low side current sense resistor (as do
several other suggestions) which your scheme avoids. The '34063 has
high side current sense capability which would suit your application.



       RM

2005\08\05@173433 by PicDude

flavicon
face
On Friday 05 August 2005 03:10 pm, Olin Lathrop scribbled:
> Sounds like a job for a 10F204.  You provide an analog level, and it runs a
> boost converter to maintain that voltage accross a low side current sense
> resistor.

I had considered a boost converter such as an MC34063, which with sufficiently
high enough voltage, the current would be low enough not to require any
external power transistor/fet.  But I really didn't want to add inductors or
caps on this board for space reasons.  And since it was for someone else, I
didn't want to give him anything with high voltages.

BTW, this circuit is also a quick project for someone else's custom truck so I
don't want to write any code for it right now, but I think it might be an
interesting exercise to write code for a DC-DC converter with a PIC.  Is this
similar to regulating current for a stepper motor where the voltage across a
sense resistor is controlled with PWM?

Cheers,
-Neil.


>
>
> *****************************************************************
> Embed Inc, embedded system specialists in Littleton Massachusetts
> (978) 742-9014, http://www.embedinc.com


2005\08\05@175445 by KY1K

picon face

>But I really didn't want to add inductors or
>caps on this board for space reasons.


The inductor is neither large or expensive, and the voltage is not a major
safety issue.

When I first started out doing this, I found that just using current
limiting resistors caused a tremendous difference in intensity depending
whether the motor was running or not.

I didn't want to operate like that, so I changed to a boost converter-now I
get the same intensity whether the motor is on or not.


>  And since it was for someone else, I
>didn't want to give him anything with high voltages.
>
>BTW, this circuit is also a quick project for someone else's custom truck
>so I
>don't want to write any code for it right now


Neil, I think this is a non-PIC project unless you really want it to be
fancy and/or massive overkill.

To some degree, my boost converter is also overkill.....but I really didn't
like the idea of having the LED's go dim when I shut the car off.

I'm aware of lower power (and smaller physical size chips), but the boost
regulator I had came in an 8 pin dip, so it was easy to build and a few
extra milliamps supply current wasn't going to make much difference to a
car battery. It's cheap and works well.

The LED's are amazingly bright and cheap too. I thought about using
expensive single (mega buck) LEDs, but couldn't resist the cheaper and
smaller LEDs in series.

Regards,

Art



2005\08\05@181631 by PicDude

flavicon
face
Wow, I thought of the MC34063 as well, but only because I picked up a couple
last week and they're really cheap.  Also thought of using the LM2585, but
again only because I have a couple here.  The problem is that I don't want
large caps and inductors on the board.  The truck is a low-rider with airbags
that can (and will) slam the truck straight to the ground.  And the chassis
hits the ground hard.  I'm worried about the structural integrity of
larger/heavier components.  Also want to keep it as flat (thin) as possible
for an easier fit.

Thanks for the data re: dim-vs-bright ratios.  I'm sure the LEDs I've chosen
would be a bit different so I'll experiment.  40mA to 4 strings would be 10mA
each, which seems reasonable, but then 150mA would be 38mA to each string.  
Can your LED's handle that much?  Since this app has 83 red LED's (per
side!), I am going with a slightly dimmer (2400mcd), but noticeably
lower-cost LED.  They are rated at 30mA absolute max, so I'll keep them at
25-27mA for the brakes.

BTW, did you use any heat-sinking on the regulator?  If I ever did go this
route, I would want to avoid a separate heatsink, and just leave a large pad
on the PCB to dissipate heat.

Speaking of night-time brightness, I've noticed that on some new vehicles, the
brake lights are overly bright and here in Texas where everything is an SUV
or truck except for my cars, I have to squint or cover my eyes when at a
traffic signal behind some of these vehicles.  IIRC the Nissan/Infinity FX is
one of these offenders.

Cheers,
-Neil.


On Friday 05 August 2005 04:06 pm, KY1K scribbled:
{Quote hidden}

2005\08\05@183414 by PicDude

flavicon
face
On Friday 05 August 2005 04:30 pm, Russell McMahon scribbled:
> If this is automotive you have say 12v available. usually more a motor
> running when brakes are usefully used but allow some wiring drop etc.

C'mon Russell, I know ya'll also speak English down there. :-)

Yeah, I figure a *healthy* vehicle when running have ~12+V at idle, and up to
13.8V during driving, but I'm allowing for much more since I've seen a
battery charger bring this up to +18V.  I expect the wiring drop to be low
since the current through these wires will be pretty low when these LED
lights go in.


{Quote hidden}

Ack!  I forgot to add the 1.25V drop!  I knew I asked here for a reason.  Will
be easy to adjust to 5 LED's per string.


> The switched resistor idea looks good, although there can always be
> something one misses.

> Using two LM317's makes the job very simple.

I've pretty much decided to go this route with all the convincing, and it adds
another nice factor -- I will have 8 regulators -- 2 for these park/brake
lights, 1 for the turn signals (3 strings of 3 LEDs), and 1 for the reverse
lights (3 strings of 3 LEDs).  Double that for the other side brings it to 8
regulators.  Keep them all the same part is a big plus.


> The idea of using a boost SMPS adds complexity but makes most other
> aspects simpler. Art's MC34063 is cheap and versatile. Also
> inefficient compared to some modern alternatives but good enough for
> this task, as he says. A modern White LED series-LED IC would do the
> whole task BUT may want a low side current sense resistor (as do
> several other suggestions) which your scheme avoids. The '34063 has
> high side current sense capability which would suit your application.


So who is Art?

Cheers,
-Neil.

>
>
>
>         RM


2005\08\05@184038 by PicDude

flavicon
face
On Friday 05 August 2005 04:36 pm, Spehro Pefhany scribbled:
> *Plus* Vref (1.25V nominal), in this configuration, yes?

Yep -- Russell's post also pointed that out to me.  I'll change to 5 LED's per
string.

> < 15 parts counting series diodes and 3 caps, not insane. And very
> resistant to crud on the 12V using a 100V transistor and shunt
> reference/power supply. And 150mV minimum rather than 3V.

If he has a problem keeping 3V on that line, he's got much bigger issues :-)



> If you go with the obvious LM317s, careful about sneak currents back
> through the regulators. The series diodes will prevent it.

Are you suggesting additional diodes at the outputs of the regulators?  You've
brought up another thought -- with parallel voltage regulators, I know a
small resistance (< 1 ohm) is usually required, but would this be the case
for parallel current sources?

Cheers,
-Neil.


>
> Best regards,
>
> Spehro Pefhany --"it's the network..."            "The Journey is the
> reward" EraseMEspeffspam_OUTspamTakeThisOuTinterlog.com             Info for manufacturers:
> http://www.trexon.com Embedded software/hardware/analog  Info for
> designers:  http://www.speff.com ->> Inexpensive test equipment & parts
> http://search.ebay.com/_W0QQsassZspeff


2005\08\05@191247 by Spehro Pefhany

picon face
At 05:40 PM 8/5/2005 -0500, you wrote:

>Are you suggesting additional diodes at the outputs of the regulators?

Nope, just that you don't decide to leave out the two at the inputs.

>You've
>brought up another thought -- with parallel voltage regulators, I know a
>small resistance (< 1 ohm) is usually required, but would this be the case
>for parallel current sources?

Nope. It's analogous to series voltage source, you don't need to do anything
special, just maintain enough voltage across them for them to regulate and
Robert's your uncle.

Sounds like a plan!

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffspamspam_OUTinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\08\05@193748 by Russell McMahon

face
flavicon
face
> battery charger bring this up to +18V.  I expect the wiring drop to
> be low
> since the current through these wires will be pretty low when these
> LED
> lights go in.

Remember that any 'extra' voltage will be dropped across the LM317's.
At 12v you have 350 mA x ~~2v = 700 mW.
This assumes load just uses up all available voltage.
Any headroom allowed for regulator will make power dissipation higher.

At 14v you have 1.4w
At 18v ...

Use the TO220 packages!
Heatsink to a tab that bolts to metal body somehow and all will be
well.
TO92 cases may go into thermal shutdown at some stage.
This "may" be unacceptable for brake lights :-)



       RM

2005\08\05@205138 by KY1K

picon face
At 06:16 PM 8/5/2005, you wrote:
>Wow, I thought of the MC34063 as well, but only because I picked up a couple
>last week and they're really cheap.  Also thought of using the LM2585, but
>again only because I have a couple here.  The problem is that I don't want
>large caps and inductors on the board.  The truck is a low-rider with airbags
>that can (and will) slam the truck straight to the ground.  And the chassis
>hits the ground hard.  I'm worried about the structural integrity of
>larger/heavier components.  Also want to keep it as flat (thin) as possible
>for an easier fit.
>
>Thanks for the data re: dim-vs-bright ratios.  I'm sure the LEDs I've chosen
>would be a bit different so I'll experiment.  40mA to 4 strings would be 10mA
>each, which seems reasonable, but then 150mA would be 38mA to each string.
>Can your LED's handle that much?  Since this app has 83 red LED's (per
>side!), I am going with a slightly dimmer (2400mcd), but noticeably
>lower-cost LED.  They are rated at 30mA absolute max, so I'll keep them at
>25-27mA for the brakes.


Mine are rated for 30 ma max as well, so with the running lights only, they
are well within ratings. With the running and brake lights on together,
they are running over their max ratings.

I powered up a few on the bench here and ran them at 60 ma for a month. At
the end of the month, it had the same apparent intensity as a brand new
diode did (compared them side by side).

So, I figured they could be overloaded for short periods (since they are
only overloaded when the brake lights are on).

Honestly, they are so much brighter than they have to be.........I will
probably reduce the number of diodes in each string (currently have 15 in
each string) and/or run them at even lower current. It's overkill, even in
the bright sun here. At night, they would surely be a pain for tailgaters
who follow to closely....maybe I should leave them as is:>:


{Quote hidden}

I noticed this as well.....I'd probably be creating a similar situation.

Regards,

Art


2005\08\06@174639 by olin piclist

face picon face
PicDude wrote:
> But I really didn't
> want to add inductors or caps on this board for space reasons.  And
> since it was for someone else, I didn't want to give him anything
> with high voltages.

The inductor would be rather small physically and also inexpensive.  Since
more LEDs would be in series, the current would be low and easily manageable
with a cheap transistor.  I wasn't talking about "high" voltages that would
be dangerous.  The voltage just needs to be enough above the 15V or so top
end normal operating voltage that the LED string doesn't come on without a
boost pulse.  20V sounds like a nice target, and there's no safety issue
with that.  The cap can be similarly small and cheap, maybe something around
220uF 50V (I haven't worked it out).


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\08\08@093516 by JAMES J VANNES

picon face
The perennial appeal of the LM317 one-resistor current-source has always
been tempered by the large dropout voltage of this part.

Has anyone managed to construct a similar design with a low dropout
linear regulator?



{Original Message removed}

2005\08\09@121238 by PicDude

flavicon
face
On Friday 05 August 2005 07:51 pm, KY1K scribbled:
> Mine are rated for 30 ma max as well, so with the running lights only, they
> are well within ratings. With the running and brake lights on together,
> they are running over their max ratings.
>
> I powered up a few on the bench here and ran them at 60 ma for a month. At
> the end of the month, it had the same apparent intensity as a brand new
> diode did (compared them side by side).

One month may be too short to determine even a significant decrease in their
lifespan.  To each their own.  Where the ones I'm building will be mounted,
it will be a nightmare to pull out and replace LED's when they are blown.  I
won't be convinced into running more continuous current than recommended.

> So, I figured they could be overloaded for short periods (since they are
> only overloaded when the brake lights are on).

When/if you sit in traffic for any length of time, an in Texas heat, those
"short" times can get quite long.


> >Speaking of night-time brightness, I've noticed that on some new vehicles,
> >the
> >brake lights are overly bright and here in Texas where everything is an
> ...
>
> I noticed this as well.....I'd probably be creating a similar situation.


There must be a law limiting the brightness of these.  I know there is one for
headlights, but haven't found one for taillights yet.

Cheers,
-Neil.



2005\08\09@124903 by PicDude

flavicon
face
On Saturday 06 August 2005 04:46 pm, Olin Lathrop scribbled:
> The inductor would be rather small physically and also inexpensive.  Since
> more LEDs would be in series, the current would be low and easily
> manageable with a cheap transistor.  I wasn't talking about "high" voltages
> that would be dangerous.  The voltage just needs to be enough above the 15V
> or so top end normal operating voltage that the LED string doesn't come on
> without a boost pulse.  20V sounds like a nice target, and there's no
> safety issue with that.  The cap can be similarly small and cheap, maybe
> something around 220uF 50V (I haven't worked it out).

I had thought of the higher voltages to reduce the number of resistors.  But
there is perhaps the advantage of using more strings, that if one LED blows,
less go out overall.  (Not sure if LEDs go open or shorted when dead).

Relatively, the cap (and inductor) would still be taller than the
multiple-LM317 options.  Both solutions would work, but the latter is simpler
and lower cost.

Cheers,
-Neil



2005\08\10@180549 by KY1K

picon face
Hi Neil,

One month of solid operation is more than your brakes will be applied for
the lifetime of the vehicle.

LED's DO NOT  violently self destruct (generally), they become dimmer over
time.

I've used LED's in place of lasers for data transmitters and have pulsed
them at ma 100 (50 percent duty cycle) for months and months (while doing
long term remote receiving several miles away). I've never had one blow up
and have never had one get soft.

The ratings for most LED's is number of hours at maximum ratings before the
intensity is reduced by 10 percent. There was a nice article some months
back about this in EDN.

It's possible you might be right, but I think worrying about higher than
rated current for brief periods of time is not warranted.

I fired up my LED's last night and determined they can be run at 3 ma at
night and 8 ma in the daytime to give adequate brightness........probably
because I have so many of them in series.

I was thinking about a photocell to determine night time and daytime
operation and set he current accordingly.

GL to all.

Art




At 12:13 PM 8/9/2005, you wrote:
{Quote hidden}

2005\08\11@045601 by Alan B. Pearce

face picon face
>I was thinking about a photocell to determine night time
>and daytime operation and set he current accordingly.

Why not just determine if the headlights have been switched on? Then run the
brightness accordingly.

2005\08\11@120215 by KY1K

picon face
At 04:55 AM 8/11/2005, you wrote:
> >I was thinking about a photocell to determine night time
> >and daytime operation and set he current accordingly.
>
>Why not just determine if the headlights have been switched on? Then run the
>brightness accordingly.


Wow, a great idea, thanks. I'd rather have the system be independent, but
sensing power to the headlights makes alotta sense.

I'll probably do it.

Thanks,

Art





More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...