Can somone double check this for me?

[Andy1845c]

Just want to make sure I know what I am doing before I soldier

I want to wire 80 strings of leds between a 12volt supply and a ground rail like this.
The data sheet for the LEDs can be found here: https://www.electro-tech-online.com/custompdfs/2006/12/tfe30uyc.pdf

Forward Voltage is 2.3 typ. Max current is 70ma.

Is this correct? 12v - 9.2v = 2.8v 2.8v/56ma = 5hm:

Would 56ma be a safe current for continious operation?

Is there any reason my circuit won't work or will shorten the life of the LEDs?

Note: the 3 circles in the diagram are supposed to be LEDs. the applet I drew it in shows them weird until rolled over with the mouse.

 

[dknguyen]

The math is right. I think 56mA is too high though for LEDs. Plus, the forward voltage drop on LEDs varies so you should try and be on the conservative side as a result (unless you individually check and possibly match the LEDs).

EDIT: I checked the datasheet and the recommended operating current is 20-70mA, so 56 should be fine. Be more worried about variances between LEDs. If enough LEDs in a series have too low a voltage drop, you may end up overloading the LEDs.

 

[Sceadwian]

You might want to keep in mind thermal variances in the resitors and the diodes themselves under load and go with a slightly higher current limiting resistor if you can, just to keep the LED's in safe operating conditions.

 

[justDIY]

i'd recommend keeping the leds in the 20-30 ma range if this is for an automotive application. just think of the high temperatures they'll be exposed to ... and would you rather have longevity and reliability or cutting-edge brightness.

the superflux / piranha package can handle ~200mW of dissipation if it is mounted to a double sided printed circuit board, with both sides covered in copper and via's placed around cathode of each LED (the cathode die-cup is also the heatsink in the superflux package)

without special design, aiming for 50-70mW (20-30mA) would be the safe bet.

 

[Hero999]

You also need to consider the nature of your power supply, is it regulated?

If not then calculate the resistor for the absolute maximum power supply voltage.

 

[andy257]

Could you explain your maths for me. I think i am misinterpreting your diagram. Is that supposed to be 4 leds in series, each with 2.3 forward volt drop?

Are we assuming the diode at the bottom also has a forward volt drop of 2.3V? What is the purpose of the diode? I dont know your application or the physical construction of what your making but if its just an (idiot diode) then there are better and cheaper ways.

Cheers

Andy

Just want to make sure I know what I am doing before I soldier

I want to wire 80 strings of leds between a 12volt supply and a ground rail like this.
The data sheet for the LEDs can be found here: https://www.electro-tech-online.com/custompdfs/2007/01/tfe30uyc.pdf

Forward Voltage is 2.3 typ. Max current is 70ma.

Is this correct? 12v - 9.2v = 2.8v 2.8v/56ma = 5hm:

Would 56ma be a safe current for continious operation?

Is there any reason my circuit won't work or will shorten the life of the LEDs?

Note: the 3 circles in the diagram are supposed to be LEDs. the applet I drew it in shows them weird until rolled over with the mouse.

 

[Andy1845c]

Could you explain your maths for me. I think i am misinterpreting your diagram. Is that supposed to be 4 leds in series, each with 2.3 forward volt drop?

Are we assuming the diode at the bottom also has a forward volt drop of 2.3V? What is the purpose of the diode? I dont know your application or the physical construction of what your making but if its just an (idiot diode) then there are better and cheaper ways.

Cheers

Andy

Yeah, its 4 leds in series. I drew the circuit in a sim applet and they show up as round dots until you mouse over them (as I did in the bottom one)

 

[Andy1845c]

Thanks for all the replys guys.

This is for an automotive app, and I was worried about the heat of the sun on it in the summer.

I wanted to regulate the voltage at an even 12v, but I haven't done much research yet to see if I can do it. I think with the car running and at 14. some odd volts it would work, but I want this to work with the car off too.
If I can't regulate it, I will need to use Hero999's method.

I'll have to rig up one of these leds at 30ma and see if it is still bright enough. I do need them to be as bright as possible, but it is more important that they last.

How much would it help to drive them PWM? I assume this can be done with a PIC?

 

[Hero999]

PWM is only useful when dim the LED or use an LED that's most efficient when pulsed at a high current at a low duty cycle, if your LED is most efficient at the normal continious current rating then PWM is a waste of time.

 

[Sceadwian]

Depends on if you have feedback or not, if you can feedback a current signal from the LED array into a PWM module to control the duty cycle it would self limit the average current going to the LED array, giving you a LOT more leeway as far as temperature and voltage restrictions go, so long as you don't violate the absolute PN junction temperature or average power in the diode's.

 

[philba]

uh, actually, your math is in error.

4 * 2.3 = 9.2V
12 - 9.2 = 2.8V (so far so good)
56 mA = .056 A
R = V/I (V = IR rearranged)
R = 2.8/.056 = 50 ohm, nearest standard value is 51
Recalcing with 51, If = 2.8/51 = .0549 A (54.9 mA)
Wattage = V*I = .0549 * 2.8 = .154 Watts

at 5 ohms, you would have been pumping 560 mA though the LEDS and dissipating 1.57 watts in the resistor. its a guess which would blow first.

Now, why not use 5 LEDs in series? Then you would have a Vdrop of .5V, a 9.1 ohm dropping resistor, 55 mA Vf and .028 watts dissipated in the resistor. This way you would have 16 strings rather than 20, .88 A vs 1.1 A total current and heat in the resistors would be .44 vs 3 watts. This would allow you more leeway in your Voltage Regulator design.

I also agree with the suggestion to lower the If to something like 30 mA. however, a lot depends on your usage duty cycle. If you were planning on a 50% duty cycle, you could safely use 56 mA since the average dissipation would be 1/2. one point - plan on a reasonable worst case scenario.

 

[audioguru]

I think your math is wrong. 2.8V/56mA is 50 ohms, not 5 ohms. With 5 ohms everything will smoke.

Who says your LEDs are 2.3V? They could be less than 1.71V or as high as 2.91V.

The LEDs are rated to have a forward voltage of 1.71V to 2.91V at 70mA. What if some of yours are 1.71V and will be less at lower current?

The battery voltage might be 14.0V when charging. If the LEDs are actually 1.71V then they total 6.84V. The current in a 50 ohm current-limiting resistor is (14.0V-6.84V)/50 ohms= 143mA and everything will smoke.
Try 56mA, then the resistor will be 128 ohms. Its power dissipation will be 0.4W so a 1/2W resistor will be very hot.

What if the battery is not charging and is 12.0v. If the LEDs are their max forward voltage of 2.91V then they total 11.64V. The current with a 128 ohm current-limiting resistor is (12.0V-11.64V)/128 ohms= 2.9mA. Very dim.

You can't use a 12V regulator because they need a minimum input voltage of 14V.
You can't use a 9V regulator with 3 LEDs in series because 3 x 2,91V= 8.73V which is too close to 9V.
You should have bought LEDs that have a narrower voltage range.

 

[HarveyH42]

I have the same LEDs, had one blow up on me...

Anyway, I hooked up an LM317T as a constant current regulator (almost 65mA) and ran 2 for a month or so 24/7. Got some LM317L (TO-92 package, 100 mA) about a month ago, and did an upgrade. Got 2 strings of 5, each with it's own regulator, running with a 12 volt wallwart. The 10 are plenty bright enough for me to find my way around the spare bedroom I use for my electronics lab. They seem very stable, and only slightly warm on the bottom of the breadboard. Regulators run cool, no heatsink (no tabs anyway...).

 

[justDIY]

You can't use a 9V regulator with 3 LEDs in series because 3 x 2,91V= 8.73V which is too close to 9V.

wait a min ... why can't he use a 9v regulator

the only problem I see is dropping 14v down to 9v, which is a lot of wasted power.

Harvey's lm317 idea is a good easy solution, and is going to be as (in)efficient as any other linear method would be. going with something more modern and cheaper than the 317 would yield less waste in dropout voltages, so the led string's forward voltage could be closer to Vin (Vout = Vin - Vdo)

We're not talkin luxeons here, there is no way those amber superflux will draw 2.9 volts.

 

[audioguru]

We're not talkin luxeons here, there is no way those amber superflux will draw 2.9 volts.

It lists a max voltage of 2.91V right on the datasheet. You can buy them (10,000 or more?) binned with the forward voltage tested and grouped into narrower voltage bands.

Flashlight manufacturers probably buy all the lower voltage ones and the leftover high voltage ones are sold on E-Bay.

 

[philba]

If they really do have such a huge variability, the first thing he should do is test a bunch of the leds to see what the Vf really is. My guess is they will be closer to the 2.3 value - manufacturers often CYA in the datasheets. Also, it is likely that what ever he has, assuming bin select rejects and/or from the same mfg run, will all be in a narrower range than the data sheet says but only testing will tell.

On the subject of power supplies - remember, he wants to light 80 of them. Any individual series solution will wind up being pretty pricey.

 

[Andy1845c]

Well, first off thank you philba and audioguru for fixing my math It would have been a bummer to wire up a bunch and have them all blow out in milliseconds.....

I never gave the lage range of forward voltage a thought when I ordered the LEDs. But half the reason I am attempting this project, to learn these things. I only ordered a few, so if they don't work, no big loss, I'll find other uses for them. But I kinda would like to use these. I like the color, brightness and the price.

3 in series would work, i'd just have to change my design a little. And that would be okay too. The more I think about it, regulation is a must. I need it to be reliable and able to withstand any voltage spikes that may come with jumpstarting or somthing like that. I'm gonna dig up Harvey's LM317 Constant Current topic and see what I can learn about that method.

philba- A stupid question I'm sure, but how do I go about measuring Vf? Is it possible to do with a battery? I tried with a 6v latern battery (kinda old, 5.79v with no load). It measured 5.70v with one of these LEDs running though a 100hm: resistor. Across the LED v is 3.67. So can I say that this LED had a forward voltage of 2.03v and is running at 37ma? When calculating current would I use the voltage across the battery while it is running the load?

 

[philba]

It's pretty straight forward. You need to put a resistor and the led in series and apply voltage. Did you put your dmm probes on either lead of the LED? If you measured 3.67 there then that is the Vf. seems high.

 

[Andy1845c]

It's pretty straight forward. You need to put a resistor and the led in series and apply voltage. Did you put your dmm probes on either lead of the LED? If you measured 3.67 there then that is the Vf. seems high.

Nope, I did it backwards from the battery to one side of the LED and I think I measured the drop across the 100hm: resistor opps. Your method gives me 2.003v.

A quick check of a few more yealds: 1.991v, 2.006v, 1.999v, 2.029v, 2.010v.
Is this close enough as to not cause me a huge problem in my series circuit?
How close is the Vf on high quality LEDs from the same batch?

 

[philba]

that's a 38 mV spread or around 1.8% (+/- 0.9%) I would measure a few more but I think that is just fine - I'd target 2.1V just to be conservative. You will be fine. I'd be surprised to see a non-bin selected batch being much closer than that.

By the way, you should probably measure the Vf at close to the targeted If. It does change with the current.