LED tracffic sign

[amindzo]

yes, i can buy it but i want to make mine.

 

[4pyros]

amindzo; Google "Joule Thief" this is what you need to get every bit out of your battery. Andy

 

[amindzo]

i want to make a smart batter charger with PIC microcontroller and i determine it ina new thread in pic microcontoller forum but before that i want to be sure about my LED array and what they need to complete as a good one.

 

[Electroenthusiast]

I would say to use a power-source with pretty high current rating(about 500mA or 1A).
And do connect all the LED's in parallel, for it to glow. It's as simple as that.

 

[ronv]

Then you need to decide how bright you want it and what viewing angle you need. That will dictate your circuit. Not the other way around.

 

[audioguru]

I have some Philips Luxeon SuperFlux red LEDs in the same 4 pins package. I made two night-lights with them. One night-light has 36 LEDs each running at "only" 53mA and they are extremely hot. The second night-light has 16 LEDs also running at 53mA each and it is not too hot.
The LEDs are mounted fairly close together and the cases are well ventilated with many holes. These LEDs are rated at 70mA like the ones you posted. LED traffic lights are brighter than my night-lights.

 

[blueroomelectronics]

That 19 LED sign is not 19 small 5mm LEDs but 19 clusters of LEDs. A single LED would not be very visible as a road sign.
**broken link removed**

 

[Mosaic]

Hmm...Why not use the PIC to drive the constant current supply so u don't need a dropping resistor?

How about:
1) Use PIC PWM as a charge pump for a largish Capacitor running off a MOSFET driver to derive a new DC supply of your LED voltage (1.5V?) .
2) You want the final Voltage drive to the LEDs to be just under their rated V to prevent overloading em. So use like a 1.6V tgt for 1.7V LEDs etc.
3) This requires NO dropping resistor....hence no power wastage and less cost.
4) In order for the PIC to manage the V to the LEDs u must employ a small resistor inline with the current flow on the 'sink' side.
5) Thus allowing the ADC on the PIC to measure the voltage drop on the Resistor .
6) Use V=IR to set the PIC ADC control voltage. If your MOSFET is driving a bank of 20 LEDs pulling 350mA, a .25 Ohm resistor will dissipate .03W and cause a voltage drop of .0875 V. This small Voltatge drop is the tgt of your PIC ADC control. in 10 bit ADC terms and a 5V ref it will result in .0875/5*1024 = 17.92 binary value.
For simplicity let's go with 18 as the ADc tgt value. That voltage drop tgt is then .0879 V with a discrete control range of 1 bit = .0049V.
A variation of 1 bit ADC control on the overall current flow is then 1/18th of it = 5.5% = about 20mA or around 1mA per LED. This will cause a small variation in brightness but this will not be really significant given the rate of control adjustment by the PIC. Once the control adjustment cycle exceeds 50Hz u won't see any brightness change. Plus the capacitor will 'smooth' the pulsing out very well.

Your PIC code just needs to monitor the ADC value and continually adjust it's PWM pulsing to closed loop control on the target adc binary 18 value. Thus the capacitor value isn't very critical. But must be in the ballpark for the PIC to control. I'd suggest using a non LED load at first to ensure good control....perhaps a resistor that draws about 350mA at the tgt voltage.

Note that the LED load represents the R part of the RC combo charge pump.

I'd suggest a resistor of perhaps 22 to 100ohms on the gate of the MOSFET to manage in rush current to the Capacitor on startup. Also keep your PWM rate between 50 to 100Hz to reduce switching power losses to under .1%. use a decent sized logic MOSFET with a low Rds. Include the Rds as part of the resistance calculation in determining the ADC binary control value. eg. An Rds of .01 ohm makes your (precision) ADC resistor value go to .25 +.01 = .26 ohm. Use thick PCB traces for the power lines or you'll have to include their resistance in the calc as well. These are non obvious sources of error in your ADC sampling.

Interestingly if u use a MOSFET with a relatively HIGH Rds ( .25ohm) (cheaper) u can eliminate the .25ohm ADC resistor and just use the RDs. Thus u sample the ADC Voltage drop across the MOSFET Drain/source pins. Simpler design with a cheaper part. Of course the .25ohm is just an example, if u get one with .3 ohm just do the calcs to determine the ADC tgt value for that Rds.