10-fold LED constant current source Variable.
10-fold LED constant current source Variable.
It's soon the time of lights, Christmas. It's not just for this reason that I developed this circuit. It is also possible to select 10 LEDs for brightness. The current flowing through the LED is very accurate and only depends on the tolerance of the resistors R1, R3, R5, R7, R9, R11, R13, R15, R17, and R19. I used 1% tolerance. In the schematic, the formula for calculating the above is available. resistors and values for 15mA to 25mA. With 12V plug power supply it is also possible to connect up to three white Superhell LED per "channel" in series. The IC used also allow 24V as a supply. In this case, a series connection of six LEDs would also be possible. Maximum 30 or 60 LEDs. If you take colors with a lower flow voltage such as red, the number of LEDs in the series circuit naturally increases.
The circuit is 100% SMD-designed, but everything can still be soldered by hand.
The LEDs are connected via 2-pin connectors. So that not all 10 "channels" have to be occupied.
Have fun replicas
Dieter Aschmann
Diskussion (1 Kommentar)
Knulldulski vor 5 Jahren
why not using a LM3409 of TI? It's a PFET buck controller for high power LED drivers. It can be used for constant current sources with a range from a hand full of mA up to 5A !!! Having such a battery of current sources it will be easy to compare low power LEDs and high power LEDs with the same circuitry.
The output current is adjustable using only a single constant voltage from 0V to 1,24V. This voltage could be derived from a DAC for each channel. Output is absolutely constant current - no pulsed DC!
The IC works with an input voltage range from 6V to 42V. Only reverse voltage protection should be used at the input.
For more information see TI-website.
I used this IC in many projects. It works very efficient. No extra cooling is needed.
Last but not least: LM3409 with all peripherals might equals the cost of one LT3092 @ Reichelt :-)
Regards,
Claus
The output current is adjustable using only a single constant voltage from 0V to 1,24V. This voltage could be derived from a DAC for each channel. Output is absolutely constant current - no pulsed DC!
The IC works with an input voltage range from 6V to 42V. Only reverse voltage protection should be used at the input.
For more information see TI-website.
I used this IC in many projects. It works very efficient. No extra cooling is needed.
Last but not least: LM3409 with all peripherals might equals the cost of one LT3092 @ Reichelt :-)
Regards,
Claus
Antworten
audiofan vor 5 Jahren
Hello Knulldulski,
in its typical application circuit the LM3409 requires 4 resistors, 3 capacitors, a p-channel mosfet, a schottky diode (fast recovery) and a power inductor additionally. Compare this to 2 resistors and a transistor and you probably know the answer.
The LM3904 may be a good solution in automotive mass applications where you can spend multilayer boards with aluminium or copper inlays for better heat spreading, but clearly not for something handwired as a single test fixture in a DIY lab.
klausb
in its typical application circuit the LM3409 requires 4 resistors, 3 capacitors, a p-channel mosfet, a schottky diode (fast recovery) and a power inductor additionally. Compare this to 2 resistors and a transistor and you probably know the answer.
The LM3904 may be a good solution in automotive mass applications where you can spend multilayer boards with aluminium or copper inlays for better heat spreading, but clearly not for something handwired as a single test fixture in a DIY lab.
klausb
Antworten
Mehr anzeigen
1 Kommentar(e)
Updates vom Autor
Dieter Aschmann vor 5 Jahren
Notes: Use of switching power supplies. A voltage of 5V is possible as the lowest value and 24V as the highest value. The pictures were taken with a power supply 15V. If you use a 5V power supply, you can only operate a wise LED.
Good luck in building up
Dieter Aschmann
Kupfer oben (3528kb)
Im Gehäse (3151kb)
Kupfer oben (616kb)
Kupfer unten (859kb)
Target Projekt (43kb)
Material (11kb)
XGerber (63kb)
Dieter Aschmann (10kb)
Dieter Aschmann (40kb)
audiofan vor 5 Jahren
First of all, thank you for that great idea. Why did you choose the LT3092 as constant current source? Reichelt charges 4.15€ for this part.
A simple current source can be made up from a transistor and two resistors. IMHO this application does not really require a precision, low drift (10 ppm) current source. Add a zener diode a potentiometer and an opamp as buffer, and you even get variable current sources. See my design proposal attached. It also improves input polarity protection. But it is not tested so far. It is a proposal and it may require some tweaking.
Care should be taken to the supply voltage used. There should not be more than 5V across C-E of the transistors. This gives a maxium of 0,05W power dissipation at 10mA. I think, 10mA will suffice all requirements. If you stay with 5V supply voltage, you also can test diode polarity.
Finally, you mean white LED and not wise LED above, don't you?
Regards, Klaus
DwayneR vor 5 Jahren
The solution is simple - simply add an emitter-follower to the output of the op-amp. Make sure you take your op-amp feedback from the emitter of the added transistor.
I used exactly the same approach you are using many decades ago - it works well so long as the loading caused by the un-used channels doesn't affect the common base drive voltage.
Michel Grigaut vor 5 Jahren
It is true however that even when no led are present, all the transistors are forward biased. But when you plug a led Vref does not change (nor VBAtt theorically).
Did I messed something?
Regards, Michel
DwayneR vor 5 Jahren
audiofan vor 5 Jahren
Think of following:
The design is for a maximum of 10mA per transistor. It is allowed to assume a beta of 50 for small signal transistors at 10mA IC, I think. So IB will be 0.2 mA in linear operation. Worst case is the not connected situation. Here IB = 2.5V/249ohm = 10mA. So worst case demands roundabout 100mA drive from opamp.
I perfectly agree that most opamps cannot continously deliver 100mA. This was not in my focus at first place. My focus was on the current sources, where a simple transistor for less tha 10 cents does the job of the expensive LT3092 almost as well. The variable voltage reference was just an "add on", which this circuit allows, almost for free.
The obvious solution to this problem is an emitter-follower, i totally agree. But it can be simply jumpers shorting the unused outputs or trasistors with a higher beta (150 and worst case current drops to 30mA) also.
As I wrote: TUN. Use whatever you have in your bench drawer. Simply adjust component values accordingly. You can even use PNP transistors supplying current from +BATT instead, if you also derive Vref from there.
So this is intended as a quick and cheap approach as I wrote:
"But it is not tested so far. It is a proposal and it may require some tweaking."
Thank you DwayneR for pointing this out. I did it from scratch and you proposed the right improvement. It is really encouraging that there are people left, who do not only consume schematics and expect them to work right out of the box.
klausb