# LEDs in Series with LM334 Current Source

M

#### Mike H

Jan 1, 1970
0
I'm working on a circut design to replace the LED's that came in a
parking-light assembly on an automobile. The manufacturers design used
some HP designed circuit that simply used current limiting resistors to
drive the leds. For whatever reason something shorted out and things
became warm enough to take the isulation off the power connection at the
circut board, melting the board itself.

Time to build a new circut with new white LED's.

Considering the voltage variability of a automobile alternator I've
decided to build into the new circut a current source rather than a using
a current limiting resistor.

Digging around it appears the LM334 is a good solution. I have 20 LED's
to light in this array.

12 volts with an estimated 3.2 volt forward voltage on the LED's allows me
3 leds per series.
3 leds per series means 7 series assemblies.

As I'm completely new to this, the question that remains for me is what is
the value of the resistor I choose to set the current output of the LM334?

The LM334 has an upper limit of 10ma, but people in the model trains world
have been able to run it at 12ma. That is the goal for me as well, get an
output from the LM334 of 12ma.

Any comments or suggestions are welcomed.

Here is a link to the circut.
http://www.frontiernet.net/~miketoni/circut.jpg

The ideas for this circuit came from the circut on this website:

Thanks.

C

#### CFoley1064

Jan 1, 1970
0
Subject: LEDs in Series with LM334 Current Source
From: Mike H [email protected]
Date: 4/30/2004 10:55 AM Central Standard Time
Message-id: <[email protected]>

I'm working on a circut design to replace the LED's that came in a
parking-light assembly on an automobile. The manufacturers design used
some HP designed circuit that simply used current limiting resistors to
drive the leds. For whatever reason something shorted out and things
became warm enough to take the isulation off the power connection at the
circut board, melting the board itself.

Time to build a new circut with new white LED's.

Considering the voltage variability of a automobile alternator I've
decided to build into the new circut a current source rather than a using
a current limiting resistor.

Digging around it appears the LM334 is a good solution. I have 20 LED's
to light in this array.

12 volts with an estimated 3.2 volt forward voltage on the LED's allows me
3 leds per series.
3 leds per series means 7 series assemblies.

As I'm completely new to this, the question that remains for me is what is
the value of the resistor I choose to set the current output of the LM334?

The LM334 has an upper limit of 10ma, but people in the model trains world
have been able to run it at 12ma. That is the goal for me as well, get an
output from the LM334 of 12ma.

Any comments or suggestions are welcomed.

Here is a link to the circut.
http://www.frontiernet.net/~miketoni/circut.jpg

The ideas for this circuit came from the circut on this website:

Thanks.

Hi, Mike. As long as the car's voltage regulator is working properly, you
shouldn't see that much variability in your power supply. You'll see around
12VDC with the car motor off, and 13.6 to 14.3V with the motor on.

An LM334 is good for 10 mA, and possibly 12 mA, but I think you're going to
need more than 12 mA to get the type of brightness you want with the white
LEDs.

If you really want to use a current source for the LEDs, you might want to try
something like this (view in fixed font or M$Notepad): Multiple Current Sources VCC o----------o-----------o-----------. | | | | | | | | V .-. .-. .-. - | | | | | | | | | | | | | | '-' '-' '-' | | | | V | | | - |< |< |< | .---| .---| .--| o-. | |\ | |\ | |\ | | | | | | | | | | | | | | | | And so on... | '--o-----------o------------o---------> | | | | | | | | | | | | .-. V ~ V ~ V ~ | | - ~ - ~ - ~ | | | | | ~ '-' V ~ V ~ V ~ | - ~ - ~ - ~ | | | | | V ~ V ~ V ~ | - ~ - ~ - ~ | | | | | | | | o----------o-----------o-----------o | === GND created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de The diodes can be 1N4001s, the resistor in series with the diodes can be 1K, and the current sense resistors are chosen as 0.7V/I(LED). For 20 mA, that would be 0.7V/.02A = 33 ohms. The transistors can be any TO-92 PNP with V(ceo) of 60V or greater, your choice. Good luck Chris M #### Mike H Jan 1, 1970 0 [email protected] (CFoley1064) wrote in Hi, Mike. As long as the car's voltage regulator is working properly, you shouldn't see that much variability in your power supply. You'll see around 12VDC with the car motor off, and 13.6 to 14.3V with the motor on. In general, the concern is what happens when the voltage regulator dies. The existing circuit, that failed, relies on resistors to limit current flow. In my case something catastrophic happened that damaged the PCB. But more commonly, the LED's in the original circuit simply die to to excess current. An LM334 is good for 10 mA, and possibly 12 mA, but I think you're going to need more than 12 mA to get the type of brightness you want with the white LEDs. I haven't received any LED's yet to verify brightness. My goal was to run them below their rated current as what I have read thus far leads me to believe that their rated current is generally a maximum and to be avoided. This application uses the LED's with a reflector and a lenses that disburses the light in a wider range than a LED normally would. It is this reflection of light that is visible in the marker lens. Since there is a right and a left, and I only need to replace a right one. The goal would be to adjust the current so that the right side would match the left side. The stock circuit (on the left side) was developed in 1993, so I'm guessing the stock design is not all that "bright" in comparison to the LED's available today. But I may be wrong. Regarding the circuit you proposed. If've done some further research and am wondering if all of the components are necessary. Here is a second circuit diagram I drew where the LED's are placed into an array, and then the current into that Array is controled by a single device. http://www.frontiernet.net/~miketoni/circut2.jpg Would that work? Could that be translated to what you proposed? I.E. reducing the number of components in the circuit. And can you clarify the calculations for me in your diagram. Where did the value for .7 come from for the LED? Thank you for your assistance. J #### John G Jan 1, 1970 0 Mike H said: [email protected] (CFoley1064) wrote in In general, the concern is what happens when the voltage regulator dies. The existing circuit, that failed, relies on resistors to limit current flow. In my case something catastrophic happened that damaged the PCB. But more commonly, the LED's in the original circuit simply die to to excess current. I haven't received any LED's yet to verify brightness. My goal was to run them below their rated current as what I have read thus far leads me to believe that their rated current is generally a maximum and to be avoided. This application uses the LED's with a reflector and a lenses that disburses the light in a wider range than a LED normally would. It is this reflection of light that is visible in the marker lens. Since there is a right and a left, and I only need to replace a right one. The goal would be to adjust the current so that the right side would match the left side. The stock circuit (on the left side) was developed in 1993, so I'm guessing the stock design is not all that "bright" in comparison to the LED's available today. But I may be wrong. Regarding the circuit you proposed. If've done some further research and am wondering if all of the components are necessary. Here is a second circuit diagram I drew where the LED's are placed into an array, and then the current into that Array is controled by a single device. http://www.frontiernet.net/~miketoni/circut2.jpg Would that work? Could that be translated to what you proposed? I.E. reducing the number of components in the circuit. And can you clarify the calculations for me in your diagram. Where did the value for .7 come from for the LED? Thank you for your assistance. Unless I missed something your new circuit requires 70 milliamps thru the LM334 and its maximum allowed is only 10ma. Surely there are larger constant current regulators that would work here but I have not got time to research this morning. The LM334 was most likely used in the model train because he needed to get the ONE led and regulator in a small space. You should not have this restriction. C #### CFoley1064 Jan 1, 1970 0 Subject: Re: LEDs in Series with LM334 Current Source From: Mike H [email protected] Date: 4/30/2004 4:04 PM Central Standard Time Message-id: <[email protected]> [email protected] (CFoley1064) wrote in In general, the concern is what happens when the voltage regulator dies. The existing circuit, that failed, relies on resistors to limit current flow. In my case something catastrophic happened that damaged the PCB. But more commonly, the LED's in the original circuit simply die to to excess current. I haven't received any LED's yet to verify brightness. My goal was to run them below their rated current as what I have read thus far leads me to believe that their rated current is generally a maximum and to be avoided. This application uses the LED's with a reflector and a lenses that disburses the light in a wider range than a LED normally would. It is this reflection of light that is visible in the marker lens. Since there is a right and a left, and I only need to replace a right one. The goal would be to adjust the current so that the right side would match the left side. The stock circuit (on the left side) was developed in 1993, so I'm guessing the stock design is not all that "bright" in comparison to the LED's available today. But I may be wrong. Regarding the circuit you proposed. If've done some further research and am wondering if all of the components are necessary. Here is a second circuit diagram I drew where the LED's are placed into an array, and then the current into that Array is controled by a single device. http://www.frontiernet.net/~miketoni/circut2.jpg Would that work? Could that be translated to what you proposed? I.E. reducing the number of components in the circuit. And can you clarify the calculations for me in your diagram. Where did the value for .7 come from for the LED? Thank you for your assistance. Hi again, Mike. The second circuit shows several strings of three LEDs in parallel, driven by a current source. That won't work too well, because minor changes in the Vf of the LEDs will mean that the current will get hogged up by the strings of LEDs that have the lowest combined forward voltage. That means major differences in brightness between LED strings. If you match forward voltages, it might work pretty well, but as a paractical circuit, no. (I guess the business with having two LEDs in one string is a typo -- that will limit forward voltage to 6.4V, and only those 2 LEDs will light. Not to mention that the LM334 has no chance of pumping enough current for all those LED strings.). Let's take a look at the original "HP" circuit for best and worst cases (view in fixed font or M$ Notepad):

+12V from regulator ~ ~ ~ GND
___ ~ ~ ~
o---|___|--->|---->|---->|------o

Let's assume you've got a series resistor of 200 ohms +/- 5%, an LED forward
voltage between 3.0V and 3.4V with a 3.2V nominal, and a regulator voltage that
varies between 11.7V (low battery, motor off) and 14.5V (high regulator, motor
on).

Nominal LED current: (13.8V - (3 * 3.2V)) / 200 ohms = 21 mA
High LED Current: (14.5V - (3 * 3.0V)) / 190 ohms = 28 mA
Low LED Current: (11.7V - (3 * 3.4V)) / 210 ohms = 7.14 mA

As you can see, high case to low case is almost a 4:1 ratio. Assuming the
motor is on, you can assume a minimum of 13.5V, which would give a low case of

Low LED Current: (13.5V - (3 * 3.4V))/210 ohms = 15.7 mA

Which gives you less than 2:1 when the motor is on. If you can get better
control over the LED Vf (after all, HP has been making LEDs since the 70s --
they know what they're doing), that means a better ratio.

Now let's assume your voltage regulator has gone south. In that case, your car
voltage may go as high as 16V or so. When that happens, you're killing your
battery, and bad things would happen to incandescent lights too because of the
higher power. In that event your nominal current would be:

Nominal LED current = (16V - (3 * 3.2v)) / 200 ohms = 32 mA

That is 50% over nominal, which isn't a problem if you're at 20 mA. If you're
at or over 40 mA, it might be fatal, depending on the LED.

That gets us back to the current source. The idea is a good one -- keep the
same current going through the LEDs despite variations in voltage. Here's a
segment of the circuit:

12V
.-o-.
| |
| |
D1 V .-.
- | |33 Ohm
| | |
| '-'
D2 V |
- | Ve
| |<
Vb o-|
| |\
.-. |Vc
1K| | |
| | V ~
'-' - ~
| |
| V ~
| - ~
| |
| V ~
| - ~
| |
'-o-'
GND

created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

You're probably familiar with the assumption that a forward-biased diode will
have about 0.7V across it. Two diodes in series, then, will have about 1.4V
across them. So, Vb is 1.4V less than the 12V. By the same token, the forward
biased PNP transistor will act so there is 0.7V from the emitter to the base.
That will mean that the transistor will tend to act so that Ve is 0.7V less
than the 12V. Your job is to choose the emitter resistor so that 0.7V across
it will give you the current you need. That would be (0.7V)/0.02A = 35 ohms,
or 33 ohms to the nearest standard value.

Of course this is a first cut approximation. Collector current is emitter
current plus base current (which should be less than 1/100th of emitter current
with a stiff base voltage source). More importantly, Vf of the diode and Vbe
of the PNP transistor is somewhat dependent on junction temperature. But under
normal circumstances in an automotive environment, Vbe will only drift by about
0.2V, and more importantly, the Vf of the diode will tend to track Vbe if
they're the same temp. Without going through the math, I can tell you that the
current source will give far better results than the series resistor alone,
like the example above. Also, a jellybean SOT-23 PNP transistor and a 1/8W SMT
resistor will cost a lot less than an LM334. The voltage source of the two
diodes and the 1/4W resistor is used by all the diode strings.

The main issue here is the voltage burden of the current source. However, even
if you assume 3.4V per diode, the current source can drive a 10.2V burden as
long as the voltage source is above 11.7V or so. Below that, the transistor
will begin to saturate, and current will go down. That should be acceptable --
people are used to having lights dim a little when the car is starting. The
main reason automotive manufacturers didn't use current sources is cost and
reduced reliability due to circuit complexity. Also, they specify LED Vf, and
generally will not accept delivery of LEDs with excessive Vf (which can be
controlled, to an extent, in manufacture).

Good luck
Chris

M

#### Mike H

Jan 1, 1970
0
[email protected] (CFoley1064) wrote in

....
The main issue here is the voltage burden of the current source.
However, even if you assume 3.4V per diode, the current source can
drive a 10.2V burden as long as the voltage source is above 11.7V or
so. Below that, the transistor will begin to saturate, and current
will go down. That should be acceptable -- people are used to having
lights dim a little when the car is starting. The main reason
automotive manufacturers didn't use current sources is cost and
reduced reliability due to circuit complexity. Also, they specify LED
Vf, and generally will not accept delivery of LEDs with excessive Vf
(which can be controlled, to an extent, in manufacture).

Thanks for the feedback and information. This should give me plenty of
information to work with.

Thanks

R

#### Robert C Monsen

Jan 1, 1970
0
Mike H said:
[email protected] (CFoley1064) wrote in

...

Thanks for the feedback and information. This should give me plenty of
information to work with.

Thanks

One other thing about LEDs in parallel is that they will hog more current as
they get hot, because their TC is negative. Thus, as they get hotter, their
Vf gets smaller, causing even more current to get put through the hot one,
leading to even more current, heat, etc, etc.

To overcome this problem, I'm a fan of the multistage current mirror. You
can regulate the current for all strings quite easily by using your current
source like this:

VCC
+
|
+----+-------+---------+--------+------+---------------+
| | | | | |
| | | | | |
|< |< |< |< |< |<
+-----------------------------------------------------|
| |\ |\ |\ |\ |\ |\
| | | | | | ... |
| | | | | | |
+-----+ V V V V V
| - - - - -
.-----. | | | | |
| | | | | | | |
| | | V V V V V
| \|/ | - - - - -
'-----' | | | | |
+----+-------+---------+--------+------+---------------+
|
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

The box with the arrow is your LM334 current source. If you use it like
this, all the strings in the system will have the same current sourced by
the current source, minus a tiny bit for the biasing.

Its easy to see why this is. Notice that the voltage from emitter to base is
equal on all of the transistors. That means that, according to transistor
theory, the current through them will be equal if they have the same temp
(and Vce, which is a far less important issue.) If you mount them all on the
same chunk of steel, they will stay pretty much the same temperature, so the
current will be pretty much the same.

fluctuates, it still regulates the current properly, since the current sink
isn't dependent on Vcc.

You can also use a resistor as your current source, but you lose the Vcc
independence if you do that. That can apparently be an issue in an
automotive environment, because the voltage can spike quite a bit.

Regards,
Bob Monsen

M

#### Mike H

Jan 1, 1970
0
One other thing about LEDs in parallel is that they will hog more
current as they get hot, because their TC is negative. Thus, as they
get hotter, their Vf gets smaller, causing even more current to get
put through the hot one, leading to even more current, heat, etc, etc.

To overcome this problem, I'm a fan of the multistage current mirror.
You can regulate the current for all strings quite easily by using

Lots of information here and I greatly appreciate it. It gives me more to
think about than I can currently grasp.

I believe that Chris' comment regarding the current limitations of the
LM334 are valid. I did some initial search for another current source
that is as simple as the LM334 but with higher capacity and didn't have
any luck. I'm not sure there would be any advantage with finding one that
is more complex as Chris showed me a design that works with simple
transistors.

In the mean time I've located a LED that has a lower Forward Voltage of
2.2v at 20ma. Thus I should be able to support 5 LEDs in each series.

Question 1 - Is this not a good idea to grow series?
Question 2 - Here is how his circuit looks extended out, keeping the 1k
resistors and 33k resistors. Am I extending this properly?

o---------------------------------------
+12v | | | | | | | |
V .-. V .-. V .-. V .-.
- | | - | | - | | - | |
| | | | | | | | | | | |
| '-' | '-' | '-' | '-'
V | V | V | V |
- | - | - | - |
| |< | |< | |< | |<
-| -| -| -|
| |\ | |\ | |\ | |\
.-. | .-. | .-. | .-. |
| | | | | | | | | | | |
| | V | | V | | V | | V
'-' -\\ '-' -\\ '-' -\\ '-' -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
o---------------------------------------
-12v
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

to the above design in that the current flow through each parallel section
above may vary as each one is being managed within each grouping of 5?

Oh... and re-reading the information from Chris, I wonder if my circuit
above should have looked like this: Rather than that^
o---------------------------------------
+12v | | | | |
V .-. .-. .-. .-.
- | | | | | | | |
| | | | | | | | |
| '-' '-' '-' '-'
V | | | |
- | | | |
| |< |< |< |<
-|---------|---------|---------|
| |\ |\ |\ |\
.-. | | | |
| | | | | |
| | V V V V
'-' -\\ -\\ -\\ -
| | | | |
| V V V V
| -\\ -\\ -\\ -\\
| | | | |
| V V V V
| -\\ -\\ -\\ -\\
| | | | |
| V V V V
| -\\ -\\ -\\ -\\
| | | | |
| V V V V
| -\\ -\\ -\\ -\\
| | | | |
o---------------------------------------
-12v
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

R

#### Robert C Monsen

Jan 1, 1970
0
Mike H said:
Lots of information here and I greatly appreciate it. It gives me more to
think about than I can currently grasp.

I believe that Chris' comment regarding the current limitations of the
LM334 are valid. I did some initial search for another current source
that is as simple as the LM334 but with higher capacity and didn't have
any luck. I'm not sure there would be any advantage with finding one that
is more complex as Chris showed me a design that works with simple
transistors.

In the mean time I've located a LED that has a lower Forward Voltage of
2.2v at 20ma. Thus I should be able to support 5 LEDs in each series.

Question 1 - Is this not a good idea to grow series?
Question 2 - Here is how his circuit looks extended out, keeping the 1k
resistors and 33k resistors. Am I extending this properly?

o---------------------------------------
+12v | | | | | | | |
V .-. V .-. V .-. V .-.
- | | - | | - | | - | |
| | | | | | | | | | | |
| '-' | '-' | '-' | '-'
V | V | V | V |
- | - | - | - |
| |< | |< | |< | |<
-| -| -| -|
| |\ | |\ | |\ | |\
.-. | .-. | .-. | .-. |
| | | | | | | | | | | |
| | V | | V | | V | | V
'-' -\\ '-' -\\ '-' -\\ '-' -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
| V | V | V | V
| -\\ | -\\ | -\\ | -\\
| | | | | | | |
o---------------------------------------
-12v
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

to the above design in that the current flow through each parallel section
above may vary as each one is being managed within each grouping of 5?

No, the one above is not a current mirror. The idea behind a current mirror
is to use the current through one transistor to control the current through
the rest of the transistors. The thing above is simply 4 separate circuits
which don't affect each other (assuming the 12V doesn't sag, which it won't
in a car.)

The circuit below is a form of current mirror. However, I think you'll get
better results with the one I posted, which also uses fewer parts, and has a
larger voltage compliance (meaning you can possibly use more LEDs per
string.) The one below also has problems with voltage spikes, which my
posted circuit does not have.

M

#### Mike H

Jan 1, 1970
0
....

The circuit below is a form of current mirror. However, I think you'll
get better results with the one I posted, which also uses fewer parts,
and has a larger voltage compliance (meaning you can possibly use more
LEDs per string.) The one below also has problems with voltage spikes,
which my posted circuit does not have.
....

O.k. Sounds like I'm learning some but not quite enough. Taking your
circuit design. Assuming I need 20ma of current through the system. That
eliminates the LM334, Correct? If not, can you demonstrate how it could
be retained?

So if the LM334 is removed, then what goes in it's place to limit the
current?

The circuit below simply puts the number of LEDS in question in place and
adds a blank space for the current control.

The LED's in question have a 2.2v Forward Voltage when run at 20ma.

With this design are we still assuming "any" TO-92 PNP transistor with a
V(ceo) of 60V or greater?

I truly apprecate all the assistance and explinations.

VCC
+
|
+----+-------+---------+--------+------+
| | | | |
| | | | |
|< |< |< |< |<
+-------------------------------------|
| |\ |\ |\ |\ |\
| | | | | |
| | | | | |
+-----+ V V V V
| - \\ - \\ - \\ \\
| | | | |
| | | | |
.---------. V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
| | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
| | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
'----+----' V V V V
| - \\ - \\ - \\ - \\
| | | | |
|---+--------+---------+--------+------+
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

R

#### Robert C Monsen

Jan 1, 1970
0
Mike H said:
...

O.k. Sounds like I'm learning some but not quite enough. Taking your
circuit design. Assuming I need 20ma of current through the system. That
eliminates the LM334, Correct? If not, can you demonstrate how it could
be retained?

So if the LM334 is removed, then what goes in it's place to limit the
current?

The circuit below simply puts the number of LEDS in question in place and
adds a blank space for the current control.

The LED's in question have a 2.2v Forward Voltage when run at 20ma.

With this design are we still assuming "any" TO-92 PNP transistor with a
V(ceo) of 60V or greater?

I truly apprecate all the assistance and explinations.

VCC
+
|
+----+-------+---------+--------+------+
| | | | |
| | | | |
|< |< |< |< |<
+-------------------------------------|
| |\ |\ |\ |\ |\
| | | | | |
| | | | | |
+-----+ V V V V
| - \\ - \\ - \\ \\
| | | | |
| | | | |
.---------. V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
| | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
| | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
'----+----' V V V V
| - \\ - \\ - \\ - \\
| | | | |
|---+--------+---------+--------+------+
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Here is a current source that can replace your LM334, made from a few
discretes.

VCC
+
|
|
| current in
| here = 20mA
| |
.-. |
1k ohms | | |
| | |
'-' |
| |
| |
| |
| |/
+-------| NPN (like 2N3904)
| |>
| |
1N914 V |
- .-.
| | | 33 ohms
| | |
V '-'
1N914 - |
| |
| |
+---------+
|
|
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

You only need one of these to control all of the strings. (it goes in
the box above.) The way it works is that the NPN transistor has a Vbe
of 0.7V. The two diodes have a combined Vf of 1.4V. Thus, there is
0.7V across the 33 ohm resistor, which means 0.7/33 = 21.21mA will go
through the NPN transistor.

You can also use a common voltage regulator, like an LM317. Its
simple:

.------------.
I in In| |Out
---------| LM317 |----+
= 20mA | | |
| | .-.
| | | |
'------------' | | 62 ohms
| |
| |
+----------+
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

The LM317 is designed to keep the voltage at Out equal to Adj + 1.25.
It'll put out whatever current it needs to to do this. Since our 68
ohm resistor is between Out and GND, then we have

Iin = 1.25/62 = 20mA.

For this circuit, you need to ensure that the Iin is greater than

You can get LM317s in TO-92 cases, just like a small transistor.

Note that your voltage compliance (Vc) is going to be less than 12V
(probably more like 11V) so if the forward voltage of your LEDs is Vf
at 20mA, then each string will only support Vc/Vf LEDs. If you put in
more than that, the strings won't light up.

Regards
Bob Monsen

M

#### Mike H

Jan 1, 1970
0
[email protected]_s54:
....
You only need one of these to control all of the strings. (it goes in
the box above.) The way it works is that the NPN transistor has a Vbe
of 0.7V. The two diodes have a combined Vf of 1.4V. Thus, there is
0.7V across the 33 ohm resistor, which means 0.7/33 = 21.21mA will go
through the NPN transistor.

You can also use a common voltage regulator, like an LM317. Its
simple:
....

Thanks again for the assistance. Busy stamping out software bugs in two
different projects started by others and left to me to fix.

I've placed your example into mine and just want to verify the
connections. Does the below look correct?

You also mention a common voltage regulator. If I were to use such a
device do I then get away from my goal of ensuring a specific current to
the leds regardless of voltage?

Thanks

VCC +
|
+--------+-------+---------+--------+------+
| | | | | |
| | | | | |
| |< |< |< |< |<
| +-------------------------------------|
| | |\ |\ |\ |\ |\
| | | | | | |
| | | | | | |
.-. +-----+ V V V V
| | | - \\ - \\ - \\ \\
| | | | | | |
'-' | | | | |
| |/ V V V V
+------| - \\ - \\ - \\ - \\
| |> | | | |
| | V V V V
V | - \\ - \\ - \\ - \\
- .-. | | | |
| | | V V V V
| | | - \\ - \\ - \\ - \\
V '-' | | | |
- | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
+--------|-------+---------+--------+------+
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

R

#### Robert C Monsen

Jan 1, 1970
0
Mike H said:
[email protected]_s54:
...

...

Thanks again for the assistance. Busy stamping out software bugs in two
different projects started by others and left to me to fix.

I hate it when that happens...
I've placed your example into mine and just want to verify the
connections. Does the below look correct?

You also mention a common voltage regulator. If I were to use such a
device do I then get away from my goal of ensuring a specific current to
the leds regardless of voltage?

Thanks

VCC +
|
+--------+-------+---------+--------+------+
| | | | | |
| | | | | |
| |< |< |< |< |<
| +-------------------------------------|
| | |\ |\ |\ |\ |\
| | | | | | |
| | | | | | |
.-. +-----+ V V V V
| | | - \\ - \\ - \\ \\
| | | | | | |
'-' | | | | |
| |/ V V V V
+------| - \\ - \\ - \\ - \\
| |> | | | |
| | V V V V
V | - \\ - \\ - \\ - \\
- .-. | | | |
| | | V V V V
| | | - \\ - \\ - \\ - \\
V '-' | | | |
- | V V V V
| | - \\ - \\ - \\ - \\
| | | | | |
+--------|-------+---------+--------+------+
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

This looks fine, except there are too many LEDs, I think, unless you
are using the standard 1.7V Vf leds, which are not very bright. I
thought you were using those high-brighness white LEDs, which are more
like 3.5V Vf. Just make sure the forward voltage across all the LEDs
is less than about 11V, and you should be fine (with Vcc >= 12V.) You
can add more strings if you need to. I'd say to keep below about 10
strings with this circuit.

Note that the value of the resistor on the bottom left (parallel to
the two series diodes) controls the current through each of the
strings. Less resistance means more current. The relation is current
times the resistance = the voltage drop across a diode, or

R = 0.7/I

where I is the current you'll get through each parallel string.

The voltage regulator circuit I described was really just a constant
current sink; its another way to do what the left bottom side of the
circuit does, except using a single IC + resistor instead of two
resistors, a transistor, and two diodes. It also wastes less current,
cause it doesn't need the bias provided by the resistor + diode pair.

Regards,
Bob Monsen

M

#### Mike H

Jan 1, 1970
0
....

This looks fine, except there are too many LEDs, I think, unless you
are using the standard 1.7V Vf leds, which are not very bright. I
thought you were using those high-brighness white LEDs, which are more
like 3.5V Vf. Just make sure the forward voltage across all the LEDs
is less than about 11V, and you should be fine (with Vcc >= 12V.) You
can add more strings if you need to. I'd say to keep below about 10
strings with this circuit. ....

The voltage regulator circuit I described was really just a constant
current sink; its another way to do what the left bottom side of the
circuit does, except using a single IC + resistor instead of two
resistors, a transistor, and two diodes. It also wastes less current,
cause it doesn't need the bias provided by the resistor + diode pair.

Currently I'm looking at a Super-Red LED that has a typical forward
voltage of 2.2v, and a MAX of 2.6v. I thought I wanted white until I
thought about the fact the lens in the assembly is already red. Why waste
energy that will be filtered?

2.6v * 5v = 13v. Woops. I guess I'll be adding another string and
decreasing the count per string.

Regarding the LM317 source you described. Could you describe for me how
it works as a current source as it appears to be a regulator of voltage.

Is the theory that by using such a device, voltage is fixed, thus a
resistor can be again used for current limitation. Thus I would no longer
need to concern myself with the fear of possible voltage issues in an
Automotive electrical system

Would the below be how I would integrate the LM317 source then? (I didn't
change the LED count per string yet assume a reduction to 4)

VCC +
|
+--------+-------+---------+--------+------+
| | | | |
| | | | |
|< |< |< |< |<
+-------------------------------------|
| |\ |\ |\ |\ |\
| | | | | |
| | | | | |
+-----+ V V V V
| - \\ - \\ - \\ \\
| | | | |
| | | | |
+---------+ V V V V
| ____ - \\ - \\ - \\ - \\
| | | | | | |
+--|LM317-+ V V V V
|____| | - \\ - \\ - \\ - \\
| .-. | | | |
| | | V V V V
| | | - \\ - \\ - \\ - \\
| '-' | | | |
| | V V V V
+----+ - \\ - \\ - \\ - \\
| | | | |
+-------+---------+--------+------+
|
===
GND
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

R

#### Robert C Monsen

Jan 1, 1970
0
Mike H said:
pair.

Currently I'm looking at a Super-Red LED that has a typical forward
voltage of 2.2v, and a MAX of 2.6v. I thought I wanted white until I
thought about the fact the lens in the assembly is already red. Why waste
energy that will be filtered?

2.6v * 5v = 13v. Woops. I guess I'll be adding another string and
decreasing the count per string.

Regarding the LM317 source you described. Could you describe for me how
it works as a current source as it appears to be a regulator of voltage.

Is the theory that by using such a device, voltage is fixed, thus a
resistor can be again used for current limitation. Thus I would no longer
need to concern myself with the fear of possible voltage issues in an
Automotive electrical system

No. Consider a device which, given a range of input voltages, can
produce a given voltage. Now, connect that fixed voltage to ground
through a resistor R. By ohms law, the current through the resistor
will be I = V/R. Since V is fixed at about 1.25V for an LM317 with a
grounded "adj" pin, and since the amount of current through the 'adj'
pin is, essentially, zero, then the total current sinked (sunk?) by
this device is V/R. Voila, its a current source! The LM317 can take
voltage differentials up to 30V, so given any voltage from 30V down to
about 3V, the LM317, hooked up as shown through a resistor, will sink
the same amount of current.
Would the below be how I would integrate the LM317 source then? (I didn't
change the LED count per string yet assume a reduction to 4)

Yes, thats it. the resistor R is going to be 1.25/I, where I is the
current you want through each string. So, for example, if you want to
pass 20mA, you use 1.25/0.020 = 62.5 ohms for R. Anything near that
will be fine.

W

#### Watson A.Name \Watt Sun - the Dark Remover\

Jan 1, 1970
0
Robert C Monsen wrote:

[snip]

The above schem assumes that all five PNP transistors have the same Vbe
drop, which would be a valid assumption if they were all on the same
chip in the same package. But they're usually not, and they are not
thermally coupled. So there could be a wide variation in the current of
each string of LEDs.

R

#### Robert C Monsen

Jan 1, 1970
0
Watson A.Name "Watt Sun - the Dark Remover" said:
Robert C Monsen wrote:

[snip]

The above schem assumes that all five PNP transistors have the same Vbe
drop, which would be a valid assumption if they were all on the same
chip in the same package. But they're usually not, and they are not
thermally coupled. So there could be a wide variation in the current of
each string of LEDs.

How much do you think? What percentage?

Regards,
Bob Monsen

R

#### Rich Grise

Jan 1, 1970
0
Robert C Monsen said:
Watson A.Name "Watt Sun - the Dark Remover" said:
Robert C Monsen wrote:

[snip]
VCC +
|
+--------+-------+---------+--------+------+
| | | | |
| | | | |
|< |< |< |< |<
+-------------------------------------|
| |\ |\ |\ |\ |\
| | | | | |
| | | | | |
+-----+ V V V V
| - \\ - \\ - \\ \\
| | | | |
| | | | |
+---------+ V V V V
| ____ - \\ - \\ - \\ - \\
| | | | | | |
+--|LM317-+ V V V V
|____| | - \\ - \\ - \\ - \\
| .-. | | | |
| | | V V V V
| | | - \\ - \\ - \\ - \\
| '-' | | | |
| | V V V V
+----+ - \\ - \\ - \\ - \\
| | | | |
+-------+---------+--------+------+
|
===
GND

The above schem assumes that all five PNP transistors have the same Vbe
drop, which would be a valid assumption if they were all on the same
chip in the same package. But they're usually not, and they are not
thermally coupled. So there could be a wide variation in the current of
each string of LEDs.

How much do you think? What percentage?

Regards,
Bob Monsen
Off the top of my head, a 2N2907 has a beta from 100-300, so I guess
300%.

But that's a very quick off-the-cuff WAG - but I've never really understood
the theory behind a current mirror, which is what this looks like.
Personally,
I would have done just "ordinary" current sources, like this:

+V
|
----o----
| |
[D] |
| [R]
[D] |
| / E
------B|
| \ C
[R] |
| [LED]
| |
----o----
|
-V

But the other circuit is probably more efficient.

Cheers!
Rich

R

#### Robert C Monsen

Jan 1, 1970
0
Rich Grise said:
Robert C Monsen said:
"Watson A.Name "Watt Sun - the Dark Remover""
wrote in message news:[email protected]
Robert C Monsen wrote:

[snip]

VCC +
|
+--------+-------+---------+--------+------+
| | | | |
| | | | |
|< |< |< |< |<
+-------------------------------------|
| |\ |\ |\ |\ |\
| | | | | |
| | | | | |
+-----+ V V V V
| - \\ - \\ - \\ \\
| | | | |
| | | | |
+---------+ V V V V
| ____ - \\ - \\ - \\ - \\
| | | | | | |
+--|LM317-+ V V V V
|____| | - \\ - \\ - \\ - \\
| .-. | | | |
| | | V V V V
| | | - \\ - \\ - \\ - \\
| '-' | | | |
| | V V V V
+----+ - \\ - \\ - \\ - \\
| | | | |
+-------+---------+--------+------+
|
===
GND

The above schem assumes that all five PNP transistors have the
same
Vbe
drop, which would be a valid assumption if they were all on the same
chip in the same package. But they're usually not, and they are not
thermally coupled. So there could be a wide variation in the current of
each string of LEDs.

How much do you think? What percentage?

Regards,
Bob Monsen
Off the top of my head, a 2N2907 has a beta from 100-300, so I guess
300%.

But that's a very quick off-the-cuff WAG - but I've never really understood
the theory behind a current mirror, which is what this looks like.
Personally,
I would have done just "ordinary" current sources, like this:

+V
|
----o----
| |
[D] |
| [R]
[D] |
| / E
------B|
| \ C
[R] |
| [LED]
| |
----o----
|
-V

But the other circuit is probably more efficient.

Cheers!
Rich

Rich:

Current mirrors just depend on the fact that, for equal transistors,
if the Vbe is equal, the current is equal. Its as simple as that. So,
setting a group of transistors with the same Vbe means they'll allow
the same current through them. Use the first one (the one with the
base-collector connection) to set the current, and everybody else is
following along. Its a neat trick.

I'm guessing that for the OPs application, the diference in current
between transistors from the same batch will be less than about %5. If
its more, and he cares, he can use a small resistor between the
positive rail and the emitter, since he can ensure that resistors are
%5 much more easily. However, that decreases the voltage range the
source will generate, meaning less LEDs he can support. Since the
worst that can happen is that one of the LED strings will be slightly
brighter than the others, I'd say that its not an issue.

Regards,
Bob Monsen

M

#### Mike H

Jan 1, 1970
0
I'm working on a circut design to replace the LED's that came in a
parking-light assembly on an automobile. The manufacturers design
used some HP designed circuit that simply used current limiting
resistors to drive the leds. For whatever reason something shorted out
and things became warm enough to take the isulation off the power
connection at the circut board, melting the board itself.

Ok. To review.

The goal here is to use a current source. Initially the thought was to
use the LM334 as a current source to drive each string of LED's. Then the
idea became a mirrored Current source. But the LM334 current limits put
it outside the realm of driving LED's that need 20mA to reach their
brightness.

A constant current source is prefered due to the variations that can occur
within the automobile charging system, and the delicate nature of LED's

Information from CFoley1064 and Robert Monsen has been very helpful. So
now it's time to look at the preliminary design and look for issues.

Below is diagram using a LM317 voltage regulator in combination with a
resistor to ensure a fixed current source regardless of changes in
automobile voltage output.

Here are the planned devices to use:

12 Volts (as high as 14 v is possible, as low as 11)
1-EA LM317 (http://www.national.com/pf/LM/LM317.html)
1-EA 62ohm Resistor .1 watt 1% tolerance
6-EA 2N2905A PNP Transistors (Has Vceo of 60)
20-EA RL5-R3545 Super-Red LED Foward Voltage of
2.2 typical, 2.6 max (http://www.superbrightleds.com/leds/r2_specs.htm)

When wired as shown below, the goal is to achieve a current level of 20mA
across all of the LED's in a fairly even manner to achieve near equal
brightness across all of them. The theory being that changing the
resistor value can impact the amount of current flowing as it will impact
the setpoint of the voltage regulator.

Questions:
1) Are the devices listed above appropriate for the goals?
2) What is a good source for the above components. (LED's I know where to
go)
3) Any other comments or changes to the circut below recommended?

Thanks

VCC +12
|
+--------+-------+---------+--------+------+-------+
| | | | | |
| | | | | |
|< |< |< |< |< |<
+-+-------+---------+--------+------+-------+-|
| |\ |\ |\ |\ |\ |\
| | | | | | |
| | | | | | |
+-----+ V V V V V
| - \\ - \\ - \\ \\ - \\
+----------+ | | | | |
| ____ | | | | |
| | | V V V V V
+--|LM317---+ - \\ - \\ - \\ - \\ -\\
|____| | | | | | |
| .-. V V V V V
| | | - \\ - \\ - \\ - \\ -\\
| | | | | | | |
| '-' V V V V V
| | - \\ - \\ - \\ - \\ -\\
+------+ | | | | |
| | | | | |
+---------+------+---------+--------+------+-------+
|
|
===
GND -12
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Replies
2
Views
764
Replies
2
Views
575
Replies
20
Views
2K
Replies
5
Views
922
Replies
17
Views
1K