I would like to switch 12 power LED's with an Arduino Due (3.3V signal pins). Only 1 LED will be on at a time. I thought I could be cheap and buy only one LED driver (LDB-350LW) and select which LED would be on using transistors. I assumed the Vout [tt]([tt]-)[/tt][/tt] connection on the LED driver would be at ground level, but to my surprise, it is at 12,2V (measured in reference to the ground from the 12V power supply). So I can never control the transistor properly from the Arduino this way.
In the image I drew the first LED in black. The 2nd I drew partially to give you an idea how they would connect. All twelve are the same. Each transistor would be controlled by 1 Arduino pin (+12K resistor).
Since I'm a beginner at electronics I wondered what would be a good alternative:
I was thinking of using relay modules (but then I need to boost the 3.3V control-signal to 5v for most products).
Could mosfets work instead? Or would they face the same issue as the transostors?
The LED cob I use is the LST1-01F06-4070-00 from New Energy (STARBOARD, OSRAM OSLON SQUARE). It has a voltage drop of 3V and uses 0,185A of power.
Can you please post a SITREP (Mil-Speak for SITUATION REPORT)
i.e (or e.g) what exactly have you done so far if anything, other than posting on the forum ?
Do you currently have the parts ?
Have tested anything ?
Have breadboarded the circuit you posted ?
etc.
etc.
n. Where did you get the schematic you posted ?
I have the components and the PCB (made with Kicad) with me.
I tested the components separately, but didn't test the transistor together with the LED and Driver, as I assumed the Vout (+) would be ±3v and Vout (-) would be 0v.
I worked on a different project with the LED drivers before (without a problem), but apparently never measured the voltage of the Vout pins on the LED driver.
When I assembled the minimum amount of components on the PCB for a LED test, I programmed the Arduino with the blink-sketch, but the LED didn’t blink. It stayed on continuously, so I started troubleshooting. I removed and tested components until I found that the LED driver put the supply voltage onto the base of the transistor.
After some searching I understood that it is not possible to switch a current when they can't have a common reference point (common ground in this case I think).
I drew the schematic myself to clarify my question.
Here's a picture of the breadboard setup were I measured the voltages from the LED driver Vout pins.
Remove the transistor from that LED circuit.
The LED should connect directly to the driver.
Try connecting the transistor to the DIM pin.
Collector to the DIM pin, emitter to -supply of the module (not -LED).
~10k resistor between Due PWM pin and base.
Another 10k resistor between +12volt and the DIM pin might be needed, but first try without.
If I connect the Arduino Due signal pin directly to the DIM pin, I can switch the LED without problems. (PWM to dim works perfect too, but I don't need dimming in this case). But in that configuration, I can only use 1 LED dimmer per LED. Since only 1 LED is on each time, I wanted to avoid buying 11 more LED drivers by switching the current of the single LED driver with transistors.
If you think getting the 11 other drivers is the best option, I'll go ahead and buy them.
Your logic makes sense, but you don't want to use transistors. The RDsON is too high.
You need mosfets rated for 1A. This one is cheap and has a low RDs(on)
It is a logic level mosfet and can be driven with a 5V logic signal but I'm not sure if 3.3V will be enough.
Wawa should be able to help you determine what you need to do with it.
It seems Mosfets are the way to go!
Thank you for the link to the component. It brought me to a tiny SparkFun breakout board that uses the same mosfet ( SparkFun MOSFET Power Control Kit - COM-12959 - SparkFun Electronics ).
In the schematic I think I recognise the same problem, namely that the control side and load side of the mosfet don't share a common ground in my project. The (-) and (+) lines between the LED and LED Driver seem to be a separate circuit, since the (-) is not at 0V. I'll try to illustrate it by annotating the Sparkfun schematic ( https://cdn.sparkfun.com/datasheets/Components/General/MOSFET_PowerControlKit_v11.pdf ).
You're using a constant current driver so if the LEDs are connected in parallel the current will be split between the two (if they are both on) and it won't necessarily split equally. If you turn one off, all of the current will flow through the other and it will get brighter.
Each LED should have it's own constant current driver if it needs to be controlled separately).
If multiple LEDs can be controlled together they can be connected in series, up to the voltage limit of the constant-current supply.
and uses 0,185A of power.
Your LED driver is rated at 350mA so it can fry your LED!
I assumed the Vout (-) connection on the LED driver would be at ground level, but to my surprise, it is at 12,2V (measured in reference to the ground from the 12V power supply).
OP says only one led will be on at any given time,
hence the need for the mosfets.
There is no point in having more than one driver.
All the led + terminals can be bussed together to the driver + Out and individual led current can be switched to the driver -Out by the mosfets, one
at a time. The driver will never know there is more than one led.
If that is a 350mA constant current driver then it is
the wrong driver for a 185mA led.
You can use a 1W 15 ohm bleed resistor in parallel
to bypass half the current.
3V * 0.185 =0.555W
3V/.185A = 16.2 ohms/0.555W
The datasheet says "Don't connect -out to -in".
(Another way of saying don't connect the driver -Out to the driver power GND.)
Do we know if the Vout- is GND? Probably not when it is forbidden to connect it to Vin-. This makes usage of the transistors slightly more complicated.
eliasheuninck:
If I connect the Arduino Due signal pin directly to the DIM pin, I can switch the LED without problems.
(PWM to dim works perfect too, but I don't need dimming in this case).
Elias
PWM-ing the DIM pin with 3.3volt will turn the LED on/off, but only part of the max brightness.
That driver expects 10volt PWM.
eliasheuninck:
The Vf of the LED is 3V
Then why did you buy a buck/boost LED driver. More expensive and lower efficiency.
You CAN't switch -LED that way with a transistor.
+LED is likely connected internally to +12volt, via a current sense resistor, and -LED is hovering at about 9volt with a single COB.
Study a CC LED driver datasheet, like the AL8808 or PT4115.
If you MUST use a 12volt supply, then you NEED a driver for EACH LED.
But you can use cheaper drivers, like the PicoBuck (for three LEDs).
Leo..
That's doable with an op amp amplifier powered with the 12V that you are using to power the driver.
All you need is a gain of 3 non-inverting amplifier to amplify 3.3V PWM to 10V PWM.
That's not an issue. @Wawa,
Are you telling me the driver would be able to tell if each led were switched to the out- pin with a relay ?
I'm not seeing what the issue is with switching it. Is it that it can't be 'hot switched' with the driver output on
or simply that it can't be switched even if the driver control input is OFF ?
Do the leds have to sequence without any deadtime between leds or can the control be turned off for 100mS
to switch the mosfets ?
That driver shouldn't have been used in the first place, so forget about solutions.
You could SHORT each LED in a current controlled string, with a transistor or relay, but that's cheap/dirty.
And not easy, because you're not switching supply or ground, but the output of an inductor.
Have a look at the PT4115 datasheet.
Much easier to use PicoBuck drivers. Or the even cheaper offerings on ebay.
Leo..
A "heretic" option that may be not viable in context of the rest of the circuit:
Power Arduino using a negative voltage regulator and switch LEDs high side (assuming it OUT+ is close to IN+; or simply connect it between OUT+ and GND).
Similar possibly crazier option: connect Arduino between OUT+ and OUT- with a Zener protecting it from overvoltage.
Thanks everyone for all the feedback!
@ raschemmel
If that is a 350mA constant current driver then it is the wrong driver for a 185mA led.
I can’t recall why I choose that LED driver. I started using it in another project 2 years ago and it hasn’t caused a problem yet. I use it with the same type of LED’s as the ones I’m using here. To clarify, the 185mA was the current I measured when the LED was powered by the driver. In the specs of the LED I see “current - test: 350mA”. So I assume that it means it’s indeed sized kind of properly.
PWM-ing the DIM pin with 3.3volt will turn the LED on/off, but only part of the max brightness.
That driver expects 10volt PWM.
If I read it correctly, the LED driver datasheet mentions that the voltage difference between Vin(-) and DIM should be 2 to 10VDC to be detected as a PWM signal, so I should be good with the 3.3V of the Arduino signal pin. I can dim through the whole range nicely if I want.
Thanks for the clarification. I was afraid that would be the case. Thank you for the PicoBuck suggestion!
@raschemmel
I was wondering too about the hot-switching. I should have enough time to turn off the driver, change the relay configuration and turn the driver back on. But since getting individual drivers for the LED's is probably a safer solution, I'll put the relay option aside for now.
@Smajdalf Haha! That’s a creative solution!
Being not very courageous, I’m going to keep it a little safer and keep the magic smoke inside. The suggested PicoBuck seems like the way to go.
If I get 4 of them, I’ll have enough outputs for the LED’s and it’ll be about €35 cheaper than going for 12 LDB-350LW dimmers.
eliasheuninck: @Wawa
If I read it correctly, the LED driver datasheet mentions that the voltage difference between Vin(-) and DIM should be 2 to 10VDC to be detected as a PWM signal, so I should be good with the 3.3V of the Arduino signal pin. I can dim through the whole range nicely if I want.
Seems you're right.
Only some Meanwell dimmers (LDD-1000) have PWM as well as linear current dimming (with a pot/voltage).
Leo..
