Could someone answer a simple question for an even simpler person? Using an LED calc

[nowinaminute]

Hi guys, I hope you can help.

Simply put, I've been buying some 5mm LEDs to use in my radio control cars. Instead of buying ready made sets, I set out to get some of the brightest 5mm units I could find which has resulted in the likes of Cree C503D and Nichias.

I'm just a little confused when it comes to using an LED calculator to determine the correct resistor to use.

The only difficulty I'm facing is that the data sheets for these LEDs dont give set in stone figures. For example it might say "typical" voltage 3.2v but "maximum" voltage 4.0v. And then the current might say "maximum 30ma"

The question is, when I use an LED calculator that asks for the voltage and current, is it just a simple case of entering the both the maxium voltage and current to get the safest max brightness from the LED?

I'm sorry if this is a really dumb question!

 

[DIWdiver]

No, it's not a dumb question. Unfortunately there isn't a simple answer.

The maximum current is specified at some conditions, maybe 25°C, maybe higher. If you are running at other conditions, then the max current really should be different. Higher temperatures mean lower currents, and vice-versa, though exactly how much is rarely easy to figure.

It would be pretty safe to use the max current and MIN voltage in the calculator, since lower LED voltage (or higher supply voltage) calculate to a higher resistor value. In your case you might use the typical voltage instead. If the voltage of your particular LED were a little lower, that would result in slightly higher current, which would not instantly destroy the LED, but just shorten its life. I'm guessing you would say that's okay.

The reason they don't give figures set in stone is that they change from batch to batch, even a little between units within a batch, and a bit with temperature and aging. Also that voltage applies only at the specified current - higher current gives higher voltage.

If you really want a precise, stable and specific current, the only way is to use a current regulation circuit.

You could get pretty close by trial and error. Use the calculator and the typical voltage to make a first approximation, measure, then adjust accordingly.

 

[Katherine Alicia]

the led Voltage is also a bit of a red herring, LED are Current sensitive devices, they have a massive Voltage range though, but will only drop 4 of those volts in your example. but keep the available current at below 30ma and you can happily run a few 100 volts through them without a problem! I`v often used a pair of parralel leds (one reverse biased) as a diac in dimmer circuits, of used them at 300+v DC to raise the cathode voltage for grid bias in valves. so don`t get too hung up on the voltage, as long as it has enough to fire up it`s all about keeping the current regulated.

 

[Lynx_Arc]

There is many things to consider when using 5mm white LEDs in devices as follows:
1)brightness
2)longevity
3)tint
4)ease of replacement

I've used all sorts of 5mm LEDs and there is issues some have over others and some issues are common but I will discuss the first 4 points.
1 & 2)Brightness is not limited by specs totally but also can be inversely related to longevity as some LEDs deteriorate at lower drive currents than other and often substantially quicker at that so in order to make them last you have to adjust the drive level to increase longevity. Cheap no name 5mm LEDs can often be so bad that unless you operate them substantially under the nominal drive current they won't last anywhere near what they are rated at instead of 10,000 hours you may get 500 hours before they start to dim noticably while others like the nichias and crees can be overdriven and even though it reduces longevity they still hold up a lot better over long perios of time. It can come down to what is an acceptable amount of runtime which if chosen low enough you may have to think of how easy it is to replace.
3)Tint is sometimes very important often is changed along the range of drive current of the LED but often LEDs at normal drive currents won't be a "pure" white color I've had nichias look pinkish purple, Crees an bluish color and other tints and these tints can sometimes be irritating to people enough that they seek out LEDs based upon tint first.
4) Ease of replacement is a good thing if you overdrive or use no name LEDs as the LED could fail or grow to dim or you may not like the tint after all the work of installing it.
Another issue of 5mm LED is the beam pattern varies and can greatly sour its use as some beams have variations of colors in them. One nichia model actually has a rectangular purple tint beam... yukk.
Now there is an option vs 5mm LED now.. smd type chip LEDs like 5050 or 3850 or other square/rectangular types. These LEDs can come in 3 or more variation such as cool white, warm white and RGB and perhaps more that I don't know of. The advantage of these LEDs is that the tint can be a lot more pleasing and white, there is no "beam" but rather a flood pattern. The issue is they are surface mounted and have no leads on them so using them could be more challenging. One final thing I thought of about chip LEDs is some versions I think have more output than typical 5mm LEDs and can take more power.
There are several methods I have used with 5mm LEDS.... the calculators online, and using a variable resistor and an ammeter on the same voltage power source and starting with maximum resistance while watching the ammeter adjust it till you reach the desired current draw and then take the variable resistor out and measure the resistance and find a suitable replacement for it. I find if you are using battery power and want maximum output through the life of the batteries. I've purposely had LEDs overdriving a little when batteries were at max voltage so they would be brighter when the batteries were more depleted and still give useful light when totally depleted. If you want stable output you may have to use a high enough voltage source to power a regulating circuit for the LED.

 

[nowinaminute]

No, it's not a dumb question. Unfortunately there isn't a simple answer.

The maximum current is specified at some conditions, maybe 25°C, maybe higher. If you are running at other conditions, then the max current really should be different. Higher temperatures mean lower currents, and vice-versa, though exactly how much is rarely easy to figure.

It would be pretty safe to use the max current and MIN voltage in the calculator, since lower LED voltage (or higher supply voltage) calculate to a higher resistor value. In your case you might use the typical voltage instead. If the voltage of your particular LED were a little lower, that would result in slightly higher current, which would not instantly destroy the LED, but just shorten its life. I'm guessing you would say that's okay.

The reason they don't give figures set in stone is that they change from batch to batch, even a little between units within a batch, and a bit with temperature and aging. Also that voltage applies only at the specified current - higher current gives higher voltage.

If you really want a precise, stable and specific current, the only way is to use a current regulation circuit.

You could get pretty close by trial and error. Use the calculator and the typical voltage to make a first approximation, measure, then adjust accordingly.

Hi and thanks for taking the time to answer.

I wish I was more electronically minded so I could understand a little better. I think the part that confuses me is that when I read tests of LEDs in forums they always seem to be tested based on current and the resulting voltage drop is more of a coincidental thing. That's why it confused me that these calculators ask for both figures. Is that because for a simple setup with a resistor, the voltage is used as a means to dictate the current? Again, I apologise if I come across as being incredibly ignorant.

I don't understand if there would be a significant difference in brightness between entering 3.4v or 4.0v into the calculator, or does the fact I'm also entering the current mean it wouldn't make much of a difference?

In terms of reduced lifespan, probably not an issue for my use. The actual amount of time I would any given vehicle in the dark is going to amount to maybe 5-10 hours per month so something like 500 hours wouldn't be an issue although obviously if it was as low as 50 or so it might be a bit of an issue.

I have one vehicle with the Nichia in and I entered the maximum 3.7v and 80ma into the calculator to choose the resistors and they have worked ok for maybe 10 hours so far but now I'm worried that the figures I entered are too much.

 

[nowinaminute]

the led Voltage is also a bit of a red herring, LED are Current sensitive devices, they have a massive Voltage range though, but will only drop 4 of those volts in your example. but keep the available current at below 30ma and you can happily run a few 100 volts through them without a problem! I`v often used a pair of parralel leds (one reverse biased) as a diac in dimmer circuits, of used them at 300+v DC to raise the cathode voltage for grid bias in valves. so don`t get too hung up on the voltage, as long as it has enough to fire up it`s all about keeping the current regulated.

Hi thanks for the answer.

This is all pretty confusing for me lol. I see it said that they work on current and not voltage but then I see people say the current draw will increase exponentially if you go over a certain voltage? :duh2:

Is it a bit different in my situation because I'm using a simple resistor to regulate the current? The calculators I use do not ask for the current of the power supply so is the voltage being used as a means to dictate the current?

I'm sorry to come across as so ignorant.

 

[nowinaminute]

There is many things to consider when using 5mm white LEDs in devices as follows:
1)brightness
2)longevity
3)tint
4)ease of replacement

I've used all sorts of 5mm LEDs and there is issues some have over others and some issues are common but I will discuss the first 4 points.
1 & 2)Brightness is not limited by specs totally but also can be inversely related to longevity as some LEDs deteriorate at lower drive currents than other and often substantially quicker at that so in order to make them last you have to adjust the drive level to increase longevity. Cheap no name 5mm LEDs can often be so bad that unless you operate them substantially under the nominal drive current they won't last anywhere near what they are rated at instead of 10,000 hours you may get 500 hours before they start to dim noticably while others like the nichias and crees can be overdriven and even though it reduces longevity they still hold up a lot better over long perios of time. It can come down to what is an acceptable amount of runtime which if chosen low enough you may have to think of how easy it is to replace.
3)Tint is sometimes very important often is changed along the range of drive current of the LED but often LEDs at normal drive currents won't be a "pure" white color I've had nichias look pinkish purple, Crees an bluish color and other tints and these tints can sometimes be irritating to people enough that they seek out LEDs based upon tint first.
4) Ease of replacement is a good thing if you overdrive or use no name LEDs as the LED could fail or grow to dim or you may not like the tint after all the work of installing it.
Another issue of 5mm LED is the beam pattern varies and can greatly sour its use as some beams have variations of colors in them. One nichia model actually has a rectangular purple tint beam... yukk.
Now there is an option vs 5mm LED now.. smd type chip LEDs like 5050 or 3850 or other square/rectangular types. These LEDs can come in 3 or more variation such as cool white, warm white and RGB and perhaps more that I don't know of. The advantage of these LEDs is that the tint can be a lot more pleasing and white, there is no "beam" but rather a flood pattern. The issue is they are surface mounted and have no leads on them so using them could be more challenging. One final thing I thought of about chip LEDs is some versions I think have more output than typical 5mm LEDs and can take more power.
There are several methods I have used with 5mm LEDS.... the calculators online, and using a variable resistor and an ammeter on the same voltage power source and starting with maximum resistance while watching the ammeter adjust it till you reach the desired current draw and then take the variable resistor out and measure the resistance and find a suitable replacement for it. I find if you are using battery power and want maximum output through the life of the batteries. I've purposely had LEDs overdriving a little when batteries were at max voltage so they would be brighter when the batteries were more depleted and still give useful light when totally depleted. If you want stable output you may have to use a high enough voltage source to power a regulating circuit for the LED.

Thanks for taking the time to answer.

Longevity isn't a huge issue due to the application, I might take a particular RC out in the dark maybe a few hours a week if that so even to build up 500 hours would take a long time. But on the other hand, I don't to be wreckless with them either. I suppose you could say I just want to get the potential out of the lights so I see the advantage over lower performing and/or cheaper units. Not much point having an LED with 4x the output if I don't get a performance that's in the ballpark. It's not even that I have any kind of figure that I need for output, just a case of "this car has holes for 5mm LEDs, let's get the brightest ones I can"

The tint isn't so much of an issue for my use, I can well appreciate how unpleasant it can be to have tints etc but for my use it's basically to just get some illumination when driving toy cars around at night so it's not too critical. It's more of a problem when I find out the lights have different temperatures AFTER I've soldered them together, doh! I thought they did something called binning so that when you purchased them, they would be fairly close together but I've had some quite obvious discrepancies. Maybe it's a bit different when you're only buying a dozen at a time from RS!

Ease of replacement isn't a problem, most of the fixtures I'm working with are just moulded to accept 5mm LEDs, they just click into place with some pressure. Quite easy to replace.

I have noticed the surface mount lights and that they come in various sizes and some of them offer much more brightness than 5mm LEDs but like I say, I'm usually working with fixtures that have been designed for the humble 5mm style units. I might look into them if I ever try to make some custom fixtures though. I noticed you can buy them on strips too which might make things easier but I wonder if you can actually search out strips that have specific LEDs on them like you can if buying them standalone?

I know what you mean about not having beams so to speak, that's another factor when it comes to RC use. For example, the Nichia I've used have a fairly wide 70 degree beam which means they are bright but the light is spread out quite far and at a farily low intensity. The Cree ones I have are only 15 degrees but the intensity inside the beam is much stronger and perhaps better for my uses in some ways. I try to combine some of both depending on the vehicle and how many lights can be fitted.

 

[Lynx_Arc]

As far as longevity the nichias can take overdriving better than most and I'm thinking the crees are able to handle it good also but the no names overdriving them you may not get even 100 hours before they dim to half or burn out. No name LEDs can vary widely on tints while the name brand tend to be a lot closer to each other in tint. The cree LEDs I have are mostly a bluish tint the nichias I have the GS and I think the one previously the GS is purple and I believe has 2 dies in it. rectangular beam pattern the other I think DS is a previouis version one die and more bluish and about half as bright. The 2 dig GS can take twice the current than most LEDs. One thing that sometimes helps 5mm LEDs is mounting them with as short of leads as possible as the heat is better and quicker dissapated off the die which goes down the legs. It won't be a huge improvement though as the leads are rather tiny so not a huge amount of heat will be wicked.

 

[DIWdiver]

One of the reasons you're having trouble is that you are trying to do the really simple thing (which is exactly what I would be doing), and not using a proper current regulating LED driver, but just using a simple resistor.

If you're using a proper driver, the LED voltage hardly matters, except that you have to make sure you have enough voltage to drive the thing. Any excess voltage is taken care of by the driver, and you don't have to worry about it.

Say you have a 5V source and your LED drops 3V at the desired 30 mA. The 'excess' voltage is 2V (5-3=2). The resistor, taking the place of the driver, has to deal with this voltage. To calculate the resistor we need, we use Ohm's law: V=IR, where V is voltage, I is current, and R is resistance. We can re-arrange this into R = V/I. 2V/30mA = 2/0.030 = 66.7 ohms. This is exactly what the online calculators do for you. If you can solve (5-3)/.03 and get 66.7, then you don't really need the calculator.

Say the batch has really crappy consistency, and another LED drops only 2.8V at 30 mA. What happens if you use the same resistor? Rearrange Ohm's law to I=V/R and you get (5-2.8)/66.7 = 0.033, or 33 mA.

Say the third one is 3.2V. Then you get only 27 mA.

Does all this really matter? Probably not. The difference between 30 mA and 33mA is only noticable if you carefully compare the two side by side. The difference between 27 mA and 33 mA is more noticeable, but you'd still have to be paying attention.

Where this gets to be a problem is when there is very little excess voltage. Then a small change in voltage in either the supply or the LED can cause a very significant change in current. Many of the flashlights built, bought, sold, and used by CPF members have exactly this problem. That's why we have constant current drivers. It allows us to have good control over current even when the voltages are not precisely known (like LEDs) or can change significantly (like batteries).

 

[JustAnOldFashionedLEDGuy]

Probably complicating a ham sandwich here nowinaminute, at least for your application.


RC Car -- I assume LiPo or similar? What is your pack voltage? Is that going to be your voltage source?


When LEDs are manufactured, they have a distribution, that is why typical is 3.2, max is 4. Look at the BIN code a few pages down on the Cree Data sheet. Bins start at 2.8-3.0 and go up to 3.8-4. Most of the LEDs manufactured will be 3 - 3.4V at typical drive current and ambient temp of 25C (makes some assumptions about how the LEDs are mounted).


RC Car, so you don't care about life too much. The LEDs have a maximum 30mA, but that is to meet their life spec. I would pick 25mA as my design point, and use 3.2V as the LED voltage knowing it will go down a bit as the LED gets hotter.

Let's say you are using a nominal 2x cell, 8.4V max, 7.4V nominal pack. I would use 8.4V for the design.

Resistor = (8.4 - 3.2) / (0.025) = 208 ohms, but call it 220 ohms for a bit of added safety margin. Given the short run times for an RC car even over it's life, you could probably go 180 ohms and not have any issues. This gives you 30mA. Use at least a 1/4 watt resistor.

As the battery drains from say 8.4 to about 6.8 (likely where you cut off), the current will drop from 30mA approx to 20mA. That sounds like a lot, but our eye response follows a 0.5x power low response approximately, so that will only look about 20% darker.

You should be able to extrapolate this to other voltages.

As noted above, you can get different LED beam patterns. Some of the LEDs targeted towards signage have oval patterns. That may be interesting.

 

[nowinaminute]

Probably complicating a ham sandwich here nowinaminute, at least for your application.


RC Car -- I assume LiPo or similar? What is your pack voltage? Is that going to be your voltage source?


When LEDs are manufactured, they have a distribution, that is why typical is 3.2, max is 4. Look at the BIN code a few pages down on the Cree Data sheet. Bins start at 2.8-3.0 and go up to 3.8-4. Most of the LEDs manufactured will be 3 - 3.4V at typical drive current and ambient temp of 25C (makes some assumptions about how the LEDs are mounted).


RC Car, so you don't care about life too much. The LEDs have a maximum 30mA, but that is to meet their life spec. I would pick 25mA as my design point, and use 3.2V as the LED voltage knowing it will go down a bit as the LED gets hotter.

Let's say you are using a nominal 2x cell, 8.4V max, 7.4V nominal pack. I would use 8.4V for the design.

Resistor = (8.4 - 3.2) / (0.025) = 208 ohms, but call it 220 ohms for a bit of added safety margin. Given the short run times for an RC car even over it's life, you could probably go 180 ohms and not have any issues. This gives you 30mA. Use at least a 1/4 watt resistor.

As the battery drains from say 8.4 to about 6.8 (likely where you cut off), the current will drop from 30mA approx to 20mA. That sounds like a lot, but our eye response follows a 0.5x power low response approximately, so that will only look about 20% darker.

You should be able to extrapolate this to other voltages.

As noted above, you can get different LED beam patterns. Some of the LEDs targeted towards signage have oval patterns. That may be interesting.

Hi there, thanks for the reply.

If only I had a brain!

The supply voltage comes from a spare channel on the radio reciever. This is dependent on the voltage the speed controller supplies to the receiver, regulated at a fixed voltage usually between 5-6v. This is the same supply that's used to power the radio reciever and the servo.

In the olden days, the radio and servos were powered by 4 seperate AA batteries but newer systems have something called a BEC (battery eliminator circuit) which is just a voltage regulator built into the speed controller. The stable BEC voltage is passed to the reciever which then makes it available to anything that is plugged in to the reciever.

I think the thing that threw me off a bit is that when I went to purchase the LEDs on the RS website, they always put the maximum voltage in the specs, even in the product title sometimes, it was only when I read the data sheets that I realised the typical voltage drop was lower.

It almost seems like RS are trying to use those voltage values as a sales point which confused me because at the time, I didn't know that they worked based on current.

So when you talk about different batches having different voltages, that is to say different units will have a different voltage drop per the same current draw?

I apologise for what must seem like a severe lack of intelligence lol

I'm still a little confused truth be told.

If I put 5v supply and 3.7v/80ma into a calculator for example, how does that differ to putting 3.2v/80ma. Does entering the current itself not ensure the maximum current is not surpassed or does it simply mean it wont be surpassed if the voltage truly is 3.7v? Is this down to the fact that the resistor way has to control voltage AND current?

I currently have some Nichias set up with the resistor value i got from entering 3.7v and 80ma (both the "max" specs) but now I'm worried I will blow them or burn them out in short order.

Again, I'm sorry if I come across as a bit dense. I'll read through the thread a few times and see if it makes a bit more sense.

Seems like the biggest source of confusion though is that the calculators ask for a specific voltage figure but in reality there isn't one. I suppose that fact that it's a simpler way of doing it compounds factors too compared to a current regulator. Almost like chasing a moving target.

 

[RetroTechie]

The question is, when I use an LED calculator that asks for the voltage and current, is it just a simple case of entering the both the maxium voltage and current to get the safest max brightness from the LED?

There IS no such thing as "the safest max brightness" !! :duh2: In essence it works like this:

• You 'push' some amount of current through it
• That current causes a voltage drop over the LED
• Voltage x current = electric power (W) 'consumed' by the LED
• Some % of that power (say, up to ~30% (?) for specific state-of-the-art LED types, usually less) is converted into light
• Remainder (so somewhere between ~70% and 100%) of that electric power is turned into heat
• That heat flow (again, measured in watts) multiplied with the "junction -> case thermal resistance" (Kelvin/W, also in datasheet) causes a number of degrees K (or C, if you want) temperature rise of the LED's light-producing area, with respect to the outside of its casing
• Cooling & mounting method determines how well that same heat flow is removed from the LED's housing (and thus how hot that housing becomes)
• Together with that temperature rise mentioned earlier, that determines how hot the LED's light-producing material gets
• That temperature (both averages & peaks) determines how fast the LED ages, and how its light output diminishes over time

In theory, that means you could calculate all of this exactly if all factors are known. In practice, you can't because there are always some things that are variable and/or unknown. For example cooling + mounting methods are rather INexact science to say the least... Not to mention LED aging, bond wires, electromigration, thermal cycling, etc etc.

Now there is an option vs 5mm LED now.. smd type chip LEDs like 5050 or 3850 or other square/rectangular types.

Nice tints is not why smd LEDs are used. There exist some high-CRI, neutral white & other nice tint options in 5mm package. It's mostly other reasons:
* No leads, and thus no need for drilling holes in a pcb.
* Thermals are much better for typical smd packages vs. 5mm LED. Both within the package, and for package -> environment heat transfer. Read: for similar size, you can push much more power through those smd parts.
* For same power that means smaller parts, more efficient fabrication & thus lower cost. Or more bang/buck.
For these reasons, 5mm have been pushed into a niche. And are only used where in a specific application they are easier to use (like, for hobbyists with a soldering iron ), or where the power is so low that thermal advantages don't matter.

This is all pretty confusing for me lol.

It is to us, too. A # of posts in, and you still haven't told us what exact exact LED types you're using, how much light you need from them (or for what kind of use), how they are mounted / cooled (or how you plan to do that), and what you're using as a power source. Having such info makes it easier to do the math, give some 'guesstimates', tips & tricks etc.

 

[peter yetman]

Such a useful thread.
Thank you all.
P

 

[nowinaminute]

It is to us, too. A # of posts in, and you still haven't told us what exact exact LED types you're using, how much light you need from them (or for what kind of use), how they are mounted / cooled (or how you plan to do that), and what you're using as a power source. Having such info makes it easier to do the math, give some 'guesstimates', tips & tricks etc.

Nichia NSDW570GS-K1-B-P9-P11 & Cree C503D-WAN-CCbEb151

I don't have a specific light requirement, I just want to get the benefit of using LEDs that are known for being bright vs the Chinese generic units all over ebay or the ones that come with pre-wired LED light kits. My only aim was to simply get as much of the potential from them as any other typical setup and counting on the fact that the units are inherently brighter to achieve more light. The more light the better of course.

There will be mounted in a simple plastic enclosures, no active of passive cooling although they be exposed to the air somewhat.

Power source as mentioned will be a relatively stable voltage which can vary from 5-6v depending on the vehicle. 5v for the one I'm currently working on for example.

 

[nowinaminute]

I read this on another web page, it seems to suggest that the setup will allow the LED to run at it's correct forward voltage even if the forward voltage isn't known. Is that correct?

http://ch00ftech.com/2012/09/03/a-l...0ftech-and-why-you-cant-run-leds-in-parallel/