Beginner getting into electonics - help please

[dbov22]

Hello,
I have spent a little time looking round this forum, but joined 2 minutes ago.

I wanted to make an LED panel so started looking on youtube etc, now different advice keeps getting thrown at me - I am a complete beginner, don't know many of the technical terms but am VERY keen to learn and turn electronics into a hobby - I'm hooked.

Anyway, I used a transformer from an old garden light set and did enough research and worked out i could fit 40x 0.022A LED's onto the circuit without any resistors.

I made the following panel and all the LED's illuminate evenly and brightly, they have been on for 5 hours and no burn outs etc. They also all light up evenly but slightly dimmer using a 3V flat round battery.

By squeekybean at 2012-08-27[/IMG]

By squeekybean at 2012-08-27

By squeekybean at 2012-08-27

This was a prototype to make sure everything worked, please don't judge or slag me off..I am looking for help and advice as to what I should have done in 'best practice'. Some people have told me that this is a terrible way to light LED's and that some will almost instantly burn out (which they haven't) and that some will barely light at all (which is not the case).....confused!

Thank you all in advance

 

[BobK]

Edited to remove mean-spirited response.

It is a bad design. The only reason the LEDs have not burned out is becuase you are overloading the power supply.

LEDs need current limiting, and attepting to draw too much current from the power supply is not an acceptable means.

Also, putting LEDs in parallel will cause them to distribute the current unevely. You probably have a bunch of reasonably matched LEDs from the same manufacturing batch, but you cannot guarantee that unless you test each one before using it in the circuit.


Bob

 

[dbov22]

Yeah, that's the thing, everyone seems to offer different advice - and to be fair, if they're all bright, evenly lit and still working after 5 hours surely that's past the point of instantly burning out?!

 

[dbov22]

I was watching a guy on youtube making LED spotlights, it basically said to divide the transformer Amps (0.8A) by the amps of the LEDs (0.020) and that figure came to 40 - so I put 40 LED's in the curcuit without having to use any resistors. So now, once again I am confused.

 

[Harald Kapp]

because you are overloading the power supply.

what Bob means is that the transformer or power supply in this case is the current limiting element. If you draw enough current from a power supply, it will at some point no longer be able to keep up a stable output voltage. The output voltage will go down (or will be turned off, depending on the design of the power supply). This effect is similar to puttting a resistor in series with the LED(s).

You are using multiple undocumented features:
- inner resistance of the power supply
- matching of the LEDs

If you want a design that consistently works even if the power supply is changed or the LEDs change their characteristic due to aging: use current limiting seriers resistors. Read the sticky post about LEDs in the electronics project section of this forum.

 

[CocaCola]

Yeah, that's the thing, everyone seems to offer different advice - and to be fair,

Some offer poor advice, some horrible and others offer proper advice... And trust me there are a bunch of hacks out there offering poor advice on Youtube...

if they're all bright, evenly lit and still working after 5 hours surely that's past the point of instantly burning out?!

Of course, it's past the point of instantly burning out, doesn't mean it's past the point of burning out in the next second... BTW we are talking transformer failure not LED failure most likely in this case...

You should be using a combination of current limited series strands, wired in parallel for the most efficiency and balance...

 

[KJ6EAD]

I agree with all of the previous responses. Now you have one more opinion to contend with.

 

[CocaCola]

One thing to note, if the output is only 5W @ 12V (as it appears in the photo) you are going to be very limited on the number of LEDs you can use if done properly...

 

[GreenGiant]

if you want to be precise about it try finding another transformer that has a slightly higher current rating (1A or so) or just use fewer LED's
put a 550 ohm (or thereabouts) resistor right in front of each one, this will limit your current to the specified limit (V=I*R... R=V/I... R=12/.022= 545.454545 aka 550+)
using a few less LED's or a slightly higher power transformer will lessen the risk of burning out the transformer and allow for loss due to the resistors

All of the above suggestions are valid

 

[BobK]

5W at 12V = 400ma. Assuming something like 3.6V for the LEDs, you get 20 strands of 3 LEDs, more that the OP is getting by doing it wrong.

Bob

 

[CocaCola]

5W at 12V = 400ma. Assuming something like 3.6V for the LEDs, you get 20 strands of 3 LEDs, more that the OP is getting by doing it wrong.

Yep, my bad had the two babies at home both having perceived immediate need crisis distracting me and inadvertently moved a decimal to 40mA in my head...

 

[john monks]

It's not unusual for me to take a contrary view from everybody else but I will.

If you bought all the led's from the same source and have the same part number the differences will not be that much and I would have done the same thing you did and connect them all in parallel because it's too much of a hassle to put in all those resistors. And if you ran this off there AAA batteries in series the internal resistance of the batteries is enough to keep from blowing out any led's as Harold Kapp seems to be saying. In fact many led flashlights (torches) do just that.
But one thing to keep in mind is that if you hook this up to a car battery all your led's will go by-by.
So what I might do is place one resistor in series with the whole works for testing.
I would hate to replace the whole thing.

 

[dbov22]

BTW - I have read that LED sticky and I am getting a much better understanding, but naturally, this does not mean sound and full knowledge, the sticky said that if resistors in a series come up as 100ohms or less then to redesign the circuit...maybe it's easier if I just tell you my plans...

I want the following...

• 1 transformer
  
• One flat LED panel containing 40 LED's (white) Vf 3.4, 0.020A
  
• 3x LED spotlights at the end of 3x 1m long cables - each spotlight will contain 10 LED's (White) as above.

Right - now when I use the LED calculator http://led.linear1.org/led.wiz it tells me to put a 100ohm resistor per series of 3 LEDS then a 470 ohm resistor for the single LED

ARGH man now i'm getting confused even just writing this... do I need to do this calculation based on 70 LED's, or separately for 40 LED's then 3x 10 LEDs, will I end up with the same value?

Can 70 LED's be bright using the transformer I have depending on how it's wired, is there no way I can wire up the LED's so I can have say 6 LED's per resistor...it's these bits i'm struggling to understand..

 

[CocaCola]

And if you ran this off there AAA batteries in series the internal resistance of the batteries is enough to keep from blowing out any led's

No it is not, the internal resistance of an AAA is no where near high enough to protect an LED, they can easily source in excess of 2+ Amps and much more depending on chemistry, I just tested a no name alkaline AAA and it pegged 7A shorted output, more than enough to toast common LEDs...

In fact many led flashlights (torches) do just that.

I have seen button cell and coin cell flashlights 'over voltaging' the LED and exploiting the internal resistance of the battery to avoid failures since those cells internal resistance limits the current output to a few mA max that won't pop the LED...

I have never seen a flashlight run off AAA or larger cells exploit the batteries internal resistance as the limiter... Instead what I have seen in those type of flash lights, is the exploitation of the limited voltage available, knowing that at that specific voltage the LED will only pass a 'safe' current, irregardless of how much the battery can supply...

 

[KJ6EAD]

BTW - I have read that LED sticky and I am getting a much better understanding, but naturally, this does not mean sound and full knowledge, the sticky said that if resistors in a series come up as 100ohms or less then to redesign the circuit...maybe it's easier if I just tell you my plans...

I want the following...

• 1 transformer
  
• One flat LED panel containing 40 LED's (white) Vf 3.4, 0.020A
  
• 3x LED spotlights at the end of 3x 1m long cables - each spotlight will contain 10 LED's (White) as above.

Right - now when I use the LED calculator http://led.linear1.org/led.wiz it tells me to put a 100ohm resistor per series of 3 LEDS then a 470 ohm resistor for the single LED

ARGH man now i'm getting confused even just writing this... do I need to do this calculation based on 70 LED's, or separately for 40 LED's then 3x 10 LEDs, will I end up with the same value?

Can 70 LED's be bright using the transformer I have depending on how it's wired, is there no way I can wire up the LED's so I can have say 6 LED's per resistor...it's these bits i'm struggling to understand..

It's not that complicated. If your voltage is well-regulated at 12V, then using 3 LEDs and a 100Ω resistor (5%) in series will run at 18mA. You can add as many of these units in parallel as you like as long as you stay under the current capability of the power supply. 400mA / 18mA = 22 groups of 3 for a total of 66 LEDs.

It's not a good idea to push the power supply to it's rated limit so if you really need to drive 70 LEDs, you should probably get a power supply capable of ~20W.

 

[dbov22]

I'm getting too bogged down in jargon here - sorry guys...I am trying to understand and take all this in... I will attach a choice of 4 different transformers I have (all from filament bulb sets) and one from my outdoor LED decking lights set.

I will also attach a pic of what I want to do...can someone please be very kind and post a quick drawing of what the 20 LED panel should be wired up like (and I'll make x2) and what the 10 panel should be wired up like (and I'll make 3) then i'll join them all up to one cable going to one transformer.

I could use 2 transformers if that will ultimately be the best thing...

By squeekybean at 2012-08-28[/IMG]

By squeekybean at 2012-08-28

By squeekybean at 2012-08-28

By squeekybean at 2012-08-28

By squeekybean at 2012-08-27

 

[CocaCola]

I could use 2 transformers if that will ultimately be the best thing...

Those all like like unregulated AC output, at least that is what I would guess without actually testing...

IMO for an LED application you would be best sourcing a new regulated DC wall wart... Something like a 12VDC @ 750mA to 1A or a 9VDC @ 1A... I'm not saying you can't do it with those transformers, just that it would not be my first choice, especially if they are indeed unregulated AC output...

For SIMPLICITY I would use a 9VDC @ 1A transformer and wire the LEDs like bellow for 9V... This is not the most 'efficient' but it's one of the most straight forward in that you only need one resistor value and the series pairs in the parallel bundles are all the same...

Remember there is an infinite number of ways to wire them, if you must use one of your existing transformers that could be done as well...

If you want efficiency and simplicity an 18VDC @ 500mA, this would allow 5 LEDs in each series with 56 Ohms resistors, like below for 18V...

 

[john monks]

Not so fast CocoCola. You are right that if dbov22 used three AAA batteries to run his LED's my way he would be overstressing his LED's.
Taking your figure of 7amps for a shorted condition the internal resistance of the batteries would be about .643ohms. Then you include the internal resistance of the LED's you add about 0.03125 ohms. Then you drop the normal forward voltage of about 3.4 volts you get about 41mA per led.
This will shorten the LED's life but won't kill them right away. Just to make sure I wasn't making a silly mistake I just took my 9 LED flashlight apart. It has 3 AAA cells connected in series and 9 LED's in parallel and no resistor. It's a cheap flashlight so maybe it won't last very long but this kind of arangement is typical.

But I agree with you whole heartly that this is not ideal.

 

[CocaCola]

You are right that if dbov22 used three AAA batteries to run his LED's my way he would be overstressing his LED's.

Well that is what you suggested isn't it? "And if you ran this off there AAA batteries in series"

Taking your figure of 7amps for a shorted condition the internal resistance of the batteries would be about .643ohms.

That was a no-name generic battery, that in no way represents what a name brand or other other AAA battery chemistry will pump out... It's really a roll of the dice...

Then you include the internal resistance of the LED's you add about 0.03125 ohms. Then you drop the normal forward voltage of about 3.4 volts you get about 41mA per led.

Can you show the math you used to get those values, as I can't seem to work any numbers and come to the same conclusion?

I just took my 9 LED flashlight apart. It has 3 AAA cells connected in series and 9 LED's in parallel and no resistor. It's a cheap flashlight so maybe it won't last very long but this kind of arangement is typical.

That is voltage limiting as I said previous, it's not current limiting dependent on the batteries internal resistance... The LEDs don't immediately fail because at 4.5V they pass a current bellow a popping threshold, it has nothing to do with the batteries internal voltage limiting the current...

 

[john monks]

Be happy to.
First I took your short circuit figure of 7 amps and divide the 4.5volt battery voltage by it and got 0.674 ohms.
Then I took the typical forward current for the LEDs of 0.02 amps and divided that into a textbook approximation forward impedance factor of 0.025 volts and got 1.25 ohms.
Then I divided 1.25 ohms by 40 to find the approximate impedance of all 40 LEDs and got 0.03125 ohms.
I added the battery impedance to the total LED impedance and got 0.674 ohms.
I subtracted the normal LED forward voltage from the battery voltage, 4.5-3.4 and got 1.1 volts.
I divided this 1.1volts by the 0.674ohm impedance and got 1.62 amps.
I divided the 1.62amps by the 40 LEDs and got 41ma per LED.

There is a certain amount of approximation but I believe my final current to be about right.
What may throw most people is the 0.025volts divided by the forward current. This is an approximation found in textbooks for silicon diodes but I found to be slightly higher in germanium diodes. But that is all I had and I believe this to be close.