LED Tester Peak Finder

Thread Starter

q12x

Joined Sep 25, 2015
909
I am expecting some power resistors at the moment, from aliexpress for another very similar project, also a current driver, but more powerful, around 1A or even more, using a circuit made by mister @Jon Chandler here on AAC forum. I am mentioning this here, because this project right now is not the same thing as the other one. This one is specific targeting LEDs, at low currents (comparative to 1A) - probably up to 100mA or even 50mA. I'm not very sure. He used a LM317 voltage regulator wired to drive the current through a Load. A very nice and good circuit but like I said, I am waiting for some important components to arrive. So in this time.... still related....ill do this tiny and cute and fluffy project.
Another reason I am specifically targeting LED's, is because I have a couple batches from different manufacturers, and also diferent colors that works at different voltages per color, and also different package sizes. Mostly smd, some TroughHole, and some Power LEDs.

- Usually, I use a very simple rule to drive a general LED (I don't care what type or what's his name). Put 10k resistor in series and done. I have the certainty of protection; that I will not blow it up, and I can lower the value of the resistor to how much luminosity I want from my LED, in reason. So I play in the safe side with my simple rule. Nothing risky and proved by time and experience it is a good setup.

But things are a bit more complicated. There are different LEDs, diferent manufacturers, different materials they are made from (impurities in the alloys), different sizes, voltages and currents they work with. My simple 10k rule is overriding all this detailed (crap) stuff and it worked for me so far. But I often thought on a method to actually drive them to their (safe) peak of current and luminosity. I guess is good to say "to their maximum luminosity", but I can make them MoRe luminous but their life shorter, if I am over drive them. For the moment lets say "maximum luminosity" from a safe point of view, when they are not overpowered.

I am thinking a very simple variable current driver made from a GPT(general purpose transistor) ~100ma like bc548 for example, to drive the current directly through a led, without its protection resistor, and gradually rise the current in the led until it's peak luminosity. But I still have the problem of --knowing-- when is the actual (safe) peak limit. One thought is to experiment enough until I get certain, possibly burning up a couple of leds in the process. Gradually writing down their current are drive at, and somehow, their luminosity and heat disipated. I dont have at all any fancy instruments, I am very limited from this point of view. Another aspect, is this question: -Is this an advanced problem to tackle into? Or is a normal thing? I am Not a proffesional electronist, I am a hobbyist in electronics, but an experienced (enough) one. Hehe. I blow up in my life a lot of stuff until I learn better.
I made a circuit based on my original thought with a GPT, voila:
Screenshot_3.jpg
Important components i used: R2=1ohm (or a simple wire). I have 3 ampmeters in the circuit, but I guess in the real circuit I will use 1 ampmeter in line with R2. Pot1=1k.
I also thought on using a LM317 voltage regulator like mister @Jon Chandler suggested for the more powerful current driver, but I think is too much for this LED Tester project here. Correct me if I'm wrong !
My question to you is:
- How do I read a LED peak (safe) limit of current driving? I will be happy to be in the vecinity of that peak, but how do I know it's THE PEAK !? How do I measure a Peak ? And this question is not ONLY for LEDs, but for any other components as well. My intention is to do the exact same thing to a couple of my transistors to check their peak as well, but with the more powerful current driver and Load tester that I already mentioned. Also to some IC's as well, why not. But until then, I want this LED tester Peak Finder solved out. In 1 way or another, orthodox or unortodox methods, I really dont care how, just to get the job done with any means possible (if this problem is not a normal problem usually encountered). I imagine and I guess this is more an industrial problem, or manufacture problem than everyday dude like me or you... Im not completly sure though.
We'll see what you will come with. Success in finding a good solution.
And as always, Thank You ! And happy new year 2022 !
 
Last edited:

SamR

Joined Mar 19, 2019
4,194
This is a bit dated but still excellent information. Notes on LEDs (gizmology.net)

10KΩ? Wow, what voltage are you using? At the usual max for through-hole generic LEDs using 30mA that would be 300V! It all depends on the component specification. I typically use ~20mA or less for generic through hole LEDs. If you are actually trying to use LEDs for illumination, the link I gave covers it pretty well. My standard quick grab for an LED current limiting resistor is 510Ω which @12V gives me ~23mA and @5V gives me ~10mA and I am satisfied with the luminosity as an indicator light. The big issue with series LEDs is their individual voltage drop when driving them with 12V or less. Looking at the Everlight 5 mm Round White LED (T-1 3/4) pdf Microsoft Word - 334-15-T1C1-4WYA.doc (futureelectronics.com) the Absolute MAX current is 30mA! If you want to try driving a few LEDs to failure you will find they will actually fail at different currents individually so that 30mA is the Do Not Exceed target. For longevity back down from there. But once again it depends on the PDF for the component being used.
 

Ian0

Joined Aug 7, 2020
5,147
There is no maximum peak current. The limit is entirely thermal. Most datasheets will show a maximum current vs. temperature graph, and that will be the maximum current that will, on average, give the LED the lifetime quoted in the datasheet.
The life will halve for every 10°C increase in temperature.
The temperature of the LED will depend on the ambient temperature and the thermal resistance between the LED chip and ambient - a big difference between a plastic-encapsulated through-hole LED on a fibreglass board and a surface mount LED on an aluminium substrate board.
Just stick to what the manufacturer says in the datasheet - his lab probably spent quite some time working out accelerated lifetimes and measuring temperatures.
 

Thread Starter

q12x

Joined Sep 25, 2015
909
Thank you mister @SamR for the reading ! I read both articles. And Very interesting things the author is covering and very well written and very well exemplified. I know many things from the lectures like 80% I want to brag, haha, but some details are novel to me. Which is good. I see there, it is a current regulator already. That is good ! And that resistor can be changed into a potentiometer.
1641202488505.png
And from the pdf 5mm LED datasheet, this -reversed zener- is new to me:
1641202678449.png
All the leds build until now had those reversed zener diodes inside them? or this is a new concept for the super bright LEDs? that are from 2010 I guess. Or is something that behaves like 2 reversed zener diodes, inside the LED? I never payed attention to this detail before.
And also thank you mister @Ian0 I agree with everything you say, especially on the current over temperature dependency, but...
" Just stick to what the manufacturer says in the datasheet " - you said
Because all of my LEDs are from ebay, made in great China we all love, there is no datasheet I could obtain. I didnt find any datasheet for the smd LEDs I have and trust me, I got 2 batches at different times from different sellers, of the suppose same type of leds, and both were different, even the color was the same, and the current was somewhat the same (I didnt measure it to the bone), but the intensity was different because the structure of the LED was visible different. One was point luminous and another batch was difuse luminous, so, I have 2 distinct types, same color, about same current and voltages. Thats for the smd. And the same for TH (Through Hole), I have GaAs phosphor 3mm and 5mm (they are called super bright LED officially on ebay) and the old GaN that mister @SamR is reffering.
The pdf that mister @SamR give me to read, covers only 1 type of 5mm super bright LED model that I have (a few of them) and most probably 1 batch of smd versions as well, not the same package I know, but the characteristics, I bet they are in the same circle of properties. Like I am saying, I have quite a ZOO of LED types here. 1 pdf does not cover all. And I am not stupid, this LED tester I am thinking, will be a great thing to have and to use. Its not a flying fart of an idea. At least for me. I hope you get my point and most of all, believe me on what Im saying about many types of LEDs that I have, at least in my drawers (not in the world). Today we live in a very experimental world of LEDs ! They changed structurally and electrically immensely, in a couple of years, and in a good way.
 

SamR

Joined Mar 19, 2019
4,194
I remember an LED tester from long ago and here is a similar one also using a TO-92 LM317 as a current regulator that is similar to my cloudy memory. I don't remember the regulator but the 500Ω pot and 47Ω resistor will allow you to ramp the current up and down with an ammeter in series.
As to other LED PDFs, I simply entered Everlight into my datasheet finder as the manufacturer and found quite a bit. EVERLIGHT Datasheet catalog
As to the Zeners, I've seen that circuit but can't remember the purpose of the Zeners. Try it and see...
 

Thread Starter

q12x

Joined Sep 25, 2015
909
To mister @SamR
" 10KΩ? Wow, what voltage are you using? At the usual max for through-hole generic LEDs using 30mA that would be 300V! "

I made an experiment page specifically for you, showing the real 10k safe rule in action.
I saw you got intrigued on my 10k way. But seeing is believing. I know exactly what you calculated there, but things are weird now with the new LED technologies. Well, not that new.
I used the extremes and the optimal settings of powering those LEDs. Also, in reality, the brightness is blinding compared to the camera image.
q20220103 - leds intensity copy 1.jpg
Ive also simulated your recommended circuit. In this circuit I put 2 ampermeter but in reality I will use 1 for the led.
Is working nice, I like it. The potentiometer is linked to gnd to get the full range of its value.
1641213081591.png
 
Last edited:

Audioguru again

Joined Oct 21, 2019
4,688
You want to test cheap garbage LEDs to destruction. Some of the LEDs might survive your current test because they will not all be identical. Garbage no-name-brand LEDs have many poor ones and maybe a few good ones.

Your very simple current tester set to zero then slowly turned up will do nothing until the slider reaches 0.6V then the transistor begins to turn on. If the transistor is a weak one then it and the LED will survive when the pot is turned up fairly high but if a strong transistor is used then it and the LED will be destroyed. A BC548 has a minimum to maximum current gain of 7.3 times and you do not know the current gain of the transistor you are using.

Almost all LEDs I use are name-brand with a detailed datasheet and the LEDs are almost identical. One time I bought a bag of no-name-brand blue LEDs and half of them did not work and the other half all had different amounts of brightness. Never again.
 

Audioguru again

Joined Oct 21, 2019
4,688
I forgot to describe the "zener diodes" in an LED. They represent the absolute maximum reverse voltage spec of about 5V where an LED has avalanche breakdown like a zener diode.
 

Ian0

Joined Aug 7, 2020
5,147
Many years ago, I used the base-emitter of transistors as zener diodes. Experienced engineers said then those transistors will never work properly as transistors.
I found that out to my cost. When I learned that the E-B junction would zener at about 6V I expected that, like all zeners, it could do that indefinitely. The transistor's hfe reduces slowly over time. I always remembered the reverse diode after that.
 

dl324

Joined Mar 30, 2015
14,468
But I often thought on a method to actually drive them to their (safe) peak of current and luminosity.
If you have standard (or high efficiency) LEDs for which you don't know the manufacturer part number, it should be safe to operate them at a continuous current of 20mA (though some could be as low as 10mA). If you're going to multiplex them, you'd operate them at a higher peak current, but you need a part number to determine the maximum current allowed and the duty cycle for currents above 20mA.

Newer LEDs seem to have a maximum peak current of 100mA or so. From Kingbright:
1641238599447.png
1641238574748.png

But some older (70's/80's) LEDs have peak currents of 1A:
1641238646831.png
1641238662717.png

This doesn't apply to high power types of LEDs.

When I'm using LEDs as indicators, I choose to operate them at a few mA because I just need to know if they're on or off; sometimes lower depending on what resistors I have handy.
 

Audioguru again

Joined Oct 21, 2019
4,688
Look at the extremely small time that each high current pulse turns them on. The HP LED has a maximum of 1 millionth of a second on time with only 300 of those narrow pulses per second.
 

Ian0

Joined Aug 7, 2020
5,147
Those 1A pulse types were often infra-red for TV remote controls. The huge current but very short pulse is easier for the photodiode to detect than a 10uS pulse at 100mA.
 

Jon Chandler

Joined Jun 12, 2008
192
Sorry I didn't see this earlier.

The LM317 constant current circuit will work great for LED testing over a wide range.

It's really impossible to determine the 'peak' operating current without a datasheet. Modern 3mm and 5mm LEDs can be very bright on just a few milliamps – as bright or brighter than old-generation LEDs at 20mA. Most 3mm and 5mm LEDs are rated for 20mA, but 10mA will usually be safe and be plenty bright.

For LEDs intended for illumination, you can make some guesses about maximum current. For a bead-type LED like the one shown, they are usually rated for 1 watt, 3 watts or even 5 watts. Try one at 1 watt with a heatsink and see if it survives. If it's not bright enough for you, try it at 3 watts and then 5 and see what happens.

Remember, heat kills LEDs. Use a big enough heatsink to keep them (sort of) cool. If they get too hot, the brightness will (permanently) fall off quickly and they will ultimately fail.

SmartSelect_20220103-151334_Edge.jpg
 

Thread Starter

q12x

Joined Sep 25, 2015
909
Hello, and thanks for the multiple responses so far.
It seems I absolutely failed to explain properly my true intention.
Mister @dl324 quote my earlier statement: "But I often thought on a method to actually drive them to their (safe) peak of current and luminosity." and you built the entire response over this very fast pulse peak that is beyond safe continuous limits. What I failed to mention was the continuous part. So my statement should be:
"But I often thought on a method to actually drive them to their (safe) continuous (not pulsed) peak of current and luminosity."
Why I used the "peak" word is because for me is a peak, since I got used to drive every led UNDER powered significantly, by using large protection resistors like my 10k rule, and slowly going down the resistance if I wanted a brighter LEDs. But I always gambled how much lower I can get, and how bright they can be, until I burned some, but never could point out what is the optimum level. I felt the need to have some sort of instrument to measure that limit or "peak" where is safe to continuously work, without burning or degrading because of it's internal heat dissipation.
From the data you present there, what im after is that 25 to 30 mA of forward current for the newer leds and the old (70's/80's) LEDs at 50mA.
Its not that I dont believe the datasheet... It's not that I dont know these values... but I dont believe my collection of components that I gathered and buy. So I want a way to be able to actually see it with my eyes, because, seeing is believing. Haha. Also, this LED tester is a test ground, and I will apply this knowledge to the more powerful (1A) current driver later, for more complicated, sophisticated or just too sensitive components and even entire circuits and IC's, why not. Well, in reason. But Mostly! for components alone. That's my plan.
I suppose I should find another word instead of "peak" because it is automatically telling you it's a "pulse" and not what I really mean.
Maximum safe forward current I guess is more close to my intention. But is a mouth full, haha. I stand corrected then.

To mister @Audioguru again , thank you for the zener explanation! I'll have to test this zener thing somehow... It remains for future experiments to do. I didn't pay attention to this detail before.

Thank you very much for the corrections and explanations so far.
I am telling you, this is a great project...
 
Last edited:

Audioguru again

Joined Oct 21, 2019
4,688
If you exceed the maximum allowed current of an LED, resistor or transistor then a few will fail soon and others will fail a little later.
If you test the maximum allowed reverse voltage rating of an LED then it will probably draw a massive current and not work properly anymore.
Without a datasheet then you are just guessing about maximum allowed ratings of cheap parts.
 
Top