Wednesday, November 28, 2012

fixing led string lights

LED lamp strings can be repaired. It can be as simple as replacing one failed bulb but can be more complex. Often it is not easy.

Early preventative maintenance, a dab of grease into each socket when new will help prevent corrosion caused failures. These strings could last much longer if the manufacturers did this for us at minimal cost. But you can do it yourself. Keep these strings running and out of the landfill.

Don't LED lights last forever?

Apparently not, particularly when used outside.

Winter is very damp here in Canada. Water gets into everything, including supposedly sealed outdoor lighting. The water causes the LED leads to corrode and fail. This corrosion failure of LEDs due to water can be completely and easily stopped with a dab of grease, if caught early.

Failed bulbs that you can see should be replaced promptly since they cause the rest of the string to be under more stress and may lead to more failures. I had the kind of failed bulb that prevented the whole string from lighting. Both kinds of failure can occur.

Like most of us, I bought into the LED low power and eco friendly thing a couple of years ago. I threw out the old filament bulb outdoor Christmas lights and bought new GE branded outdoor LED string lights.

Yesterday, while getting ready to put up the Christmas lights, I found that two out of eight strings did not work. Each has 25 LEDs. I spent part of the day exploring why those had failed and putting up the rest. Fortunately I had tried them before putting them up!

(click any pic to enlarge it)

My GE strings are supposed to be repairable. They come with instructions (the full copy, a pdf) and spare LEDs.

The worst part about my two failed strings was that none of the bulbs lit up. Not one or two bulbs that weren't on while the rest worked. The failed bulb could not be seen since it kept all the others from lighting. This meant that each bulb had to be pulled and a working bulb inserted (according to the manufacturer's instructions).

I thought that we had abandoned series wired light strings back in the 60s? Series wiring is back with LED lights!

I can remember my dad back in the day proudly showing off a new parallel wired string, with which one failed bulb would not darken the whole string. He had spent many a time finding a failed bulb in a dead string so he knew the advantage of parallel wiring. From then until now, the advent of LED lighting, most of our filament bulb lighting strings were wired in parallel.

Series wiring is back with LED lights! It turns out that there is a technical reason why LEDs need to be wired in series and this will be discussed below. Since they are wired in series, I need to check every LED bulb if none of the string lights. Any one of the LEDs could be open circuit.

There are 25 LEDs in each string so that is 25 decorative caps to pull off and then 25 LEDs to pull out. Some of the caps come off easily, others do not. I found that a drop of WD40 helped free some of the stubborn caps. I dripped the lubricant onto the joint between the cap and the socket.

Some of the caps broke when I pulled them off, usually a part of the sealing ring was left in the socket. I dug out the debris from the socket and discarded those caps and replaced them with spares from the scrap strings. The caps are keyed and only fit on one way. Why they did this I am not sure.

I admired the first class molded parts that made up the socket assembly including the LED holder, the cap and the base. There was no flash and the parts fit together very exactly and securely with a snap. With ideal plastic that did not weather or change dimensions or properties with UV and temperature, these would be really nice parts. After a couple years outside, some are harder to get apart than others.

Pulling out the LEDs was sometimes problematic. I used my fingernails inserted in a small groove between the LED and socket and just pulled. Here also, a bit of lubrication from the dielectric grease helped with subsequent removals.

Throughout all this pulling on the socket, I tried not to pull on the wires. They seem securely crimped into the terminals but I thought it best not to pull too hard on them. The terminals are a potential point of failure so while the LED was out of the socket I took a look at the joint with the wires and the condition of the terminals before dabbing the grease and re-inserting the bulb. The inner (insert) portion of the socket can be removed and that will be described below. I found it was desirable to remove the insert on those where corrosion was seen in the socket.

It's about RUST

The cause of my failures was common corrosion. Each LED is itself encased in clear plastic but the metal leads that form it's connection with the socket are made of iron or steel and these leads rust if exposed to water and oxygen. The socket weatherproofing can be improved with a dab of dielectric grease to prevent the oxygen and water from reaching the point where two dissimilar metals touch: the LED lead (iron) and the socket contact (some other metal). Why wouldn't the manufacturer do that?

If I doubted that the LED leads were ferrous, here is a picture of the rusted ones stuck to a small magnet. You can see a broken off LED lead in the foreground.

I have opened all the bulbs on three of my LED strings so far. One that I scrapped had a high percentage of failed bulbs 8 out of the 25. Another that I did not scrap had only one of these problem bulbs. All of these bulbs failed because the LED leads rusted. Some of the LED leads had fallen off or rusted completely through.

Probably millions of these string lights will end up in the landfill over the next few years because people will give up on them. Hopefully you can use some of these ideas to help you to locate the problems with your own and keep them going for a few more years.

Understanding how LED string lights work

You do not have to cut the LED string apart and unwind it as I did here but it helped my understanding of how the thing was put together and I am going to use this picture to explain what I found.

Parallel wired light strings have two wires down the length of the string and each lamp is wired between the pair of wires. These LED strings have three wires down the length. It is a little difficult to see what is going on unless you do what I did with a string that was damaged beyond repair. I separated the string into two parts by clipping at only one point. I could then unwind the string into it's two sections as revealed in the picture above.

This is a schematic of my LED string. I clipped the wire at the bottom of the canister. The center portion, the string of LEDs, can then be separated from what I've called the extension cord portion. You can see in the schematic, the wires across the very top and bottom connect the plug at one end to the socket at the other. The series string of LEDs is inter-wound, but is separate from, the extension cord except at the very ends where the LED string and the canister join the circuit, across the two hot leads.

So you can see the LEDs are all in a string and that if any of the LEDs fails open, the current flow in the string will be interrupted and the entire string will be dark. Unlike filament bulbs which rarely fail shorted LEDs can fail in this way. No light is produced yet current continues to flow through the LED. It is thus easy to see which LED failed since it will be the dark one.

It is interesting that in the GE instructions, they seem to presume that the LEDs will fail short and you should look for dark ones and replace them. This was not my experience. All of my failures were failed open circuit, due to rust.

I sacrificed two of the canisters to dissection for the cause of this project. You can see that my method improved after I learned to cut through the cover sheath at the joints of the inner cylinder. There are two 2000 ohm 1 watt resistors and a diode wired in series on a three section frame. Kind of a clever thing. No signs of water ingress with either one, unlike with about 10-20% of the lamp sockets. But I had to look to see what was in here.

I'd like to point out the excellent writing of Terry Ritter "LED Christmas Lights and How to Fix Them". Terry has been down this road and he writes about how LED strings work and what he learned with strings made by Philips. These I have from GE seem very similar.

Terry describes that the LEDs are arranged in series strings because they do not tolerate high voltage well. The voltage that comes from a wall plug (120VAC in North America) is high voltage to an LED which normally operates from a few volts - 3.1 volts in the case of my white LEDs. By stacking 25 of them in a string, each LED drops 3.1 volts so the total voltage drop across the LED string is 77.5 volts. The resistors in the canister are used to drop the remainder (120 - 77.5 = 42.5 volts).

Interestingly, if LEDs fail short, they increase the voltage drop on all the other components in the series circuit so the resistors get hotter and all the other LEDs pass more current. So you can see why GE tells us to "replace failed bulbs promptly" because the whole string goes into overdrive mode if LEDs fail shorted. If they fail open, the whole string goes dark and no current flows.

Terry found that the Philips LED string contains a small fuse in the plug end. If the fuse blew, the string would not light. Do I have a fuse? No mention of a fuse in the GE pamphlet. I attempted to find such a fuse and went so far as to destroy the stacking plug end of one of my rejected strings. I think I can say with authority that there is no fuse in the GE unit. Maybe where Terry is, they need to have a fuse? You should eliminate that possibility, that you have a fuse. Sometimes the "fuse" is thermal and permanently cuts the circuit. Something like that would be in the canister, with the resistors.

Terry talks about failed LEDs being dark and easy to spot, so obviously he saw lots of the failed short type. Not here. Terry wrote in 2007 so perhaps the LEDs are different now. He uses LEDs of different colors whereas mine are all white.

Terry also mulls the lack of a blocking diode in the Philips string and suggests that manufacturers add a blocking diode to better protect the LEDs from reverse transient voltages and it seems that GE listened to his advice and added such a reverse blocking diode in the canister.

Finally, Terry points out that running LEDs from AC means that the LEDs are only on for a part of the time leading to a flicking effect.

I attached a current meter to one of the wires in the GE LED string and displayed the results on an oscilloscope to show what Terry is describing.

The green oscilloscope trace shows that most of the time (the flat part of the trace), the LED is OFF and for about 1/3 of the time, 60 times a second, the LED string conducts hard, up to about 44 mA maximum and then goes off again. 44mA is a lot of current for an LED of this size which would normally run at between 10-20 mA current if it was continuous or DC current. I suspect that the string would appear much brighter yet use the same cost of electricity if the string was operated from DC rather than AC.

I will look at that possibility in a future article.

Fixing the LED sockets

The main failure mechanism I had to deal with was rust in the sockets (rust is an insulator) and deteriorated LED leads.

The LEDs with deteriorated leads (any sign of rust at all) were discarded and replaced. Fortunately GE had supplied a number of spare bulbs (about six) with each string. Working ones from the scrapped strings provided more spare bulbs and caps.

The sockets were cleaned in the following way. I found that it was possible to separate the sockets into two parts by pushing out the inner section with a small blunt tool. You can lay the socket with its open face flat on a surface and push hard on the center, between the wires, to free it up. Then if you pick the socket up and push, the center portion should just come right out.

This picture is a bit ugly with the socket guts covered with Vaseline, the dielectric grease I had handy. This socket was particularly badly coated with rust and the LED leads had completely disintegrated so it is the one I wanted to show you.

It is interesting that in any galvanic pair of metals, it is one metal that corrodes and the other is fine (more about galvanic corrosion). In the case of these sockets, the LED lead suffers but the contacts are ok except that they are coated with rust. With the socket disassembled it is very easy to get at the top edge of the connector to clean it up. This one has been together for a while potted in Vaseline which seems to be lifting the rust residue which is fine, both the Vaseline and the rust are insulators. The point of the Vaseline is to keep water and oxygen away from the LED lead and the place it touches the conductor in the socket. You can see the bright metal edge on the socket conductor where I have scrubbed it with a small screwdriver to clean up the edge. It is only these top edges that touch the LED leads, you don't need to worry about cleaning up anything else. To re-seat the contacts, just push them back into the socket. They will click into place. The Vaseline helps to seal the joint where the wires enter the socket also.

Finally, I ran a bead of Vaseline around the base of the cap before inserting it into the socket. This improves the lubrication so the cap is hopefully easier to remove next time and the film of Vaseline would help to seal the joint between the cap and the socket, helping to keep out the water.

Hopefully these lights will continue to give good service for years to come.

Your comments are welcome. Thanks for your interest and good luck with your LED light strings.

George Plhak

You might also be interested in my series on
a very bright 1 watt diy led garden light

8 comments:

Chrisman said...

A couple of things...
1. All "Miniature" Christmas lights are wired in series, the difference is that the modern (post 1960-1970 something) ones have a loosely wrapped and lightly insulated shunt inside each bulb that "kicks in" (welds itself across the inputs) when the filament fails, keeping the rest of the string lit. As with the LED strings, it is imperative that dead bulbs be replaced immediately to avoid the rest of the bulbs in that string being overloaded.

I work for a large distributor of light strings, and am currently working on a solution based upon a CL220 current regulator from Supertex, Inc. It will take any input voltage from 5-220V and will drive any number of LEDs in series so long as the total forward voltage does not exceed the supply voltage.

Let me just say that this chip is awesome, the light output is increased, there is no flicker, and even though it has three pins, you only use two, you simply wire it in series just like a fuse, and it automatically limits the current to 20mA no matter the input voltage, number, color or manufacturer of your leds. What could be simpler? You can find more info along with many other LED solutions at:

http://www.supertex.com/feature_LED_general.html
Go to the website and request a free sample!

Did that sound like a commercial? I can assure you that I do not work for them, but after requesting a free sample, getting it in 3 days, and seeing the results, I am very impressed. They also had a rep contact me and he has provided circuit details for our specific application.

Here in CA, and hopefully the rest of the world, we should be seeing LED light strings live up to the promises - they are really bright (even at a measured draw of only 19.6 mA), have no flicker, no reverse voltage, and no stress from surges/ESD (when combined with an MOV, a bypass cap and a filter cap).

So far (only about one week of continuous use) there is no visible signs of failure. And the price at Mouser is only $1.43 for single units.
http://www.mouser.com/search/refine.aspx?Ntk=P_MarCom&Ntt=109749218&utm_source=supplier&utm_medium=featured+spotlight&utm_campaign=109749218

Chris

Heather said...

Thanks for the info - I read mr ritters article too and still not sure if I can fix my squirrel-chewed lights . The squirrel has chewed thru the wire directly at a bulb on each of my stings. I could not splice the light since it was at the light, only I had to cut out the light. I soldered the wires together w no luck. My brother suggested removing a second light, saying sometimes circuits are every other one so I did that but it to did not fix the light. So one half of each string will not light up. Am I understanding from your article that if a single bulb is removed the voltage is too high for remaining bulbs n they will burn themselves out? So I have too much juice running thru my string now? If so could I perhaps use this string as my stock and replace the chewed lights (entire lights w wire ) and splice them into the remaining damaged cords?

George Plhak said...

Thanks for your interest Heather. It sounds like probably the best thing is what you suggest, to use the damaged strings as your repair stock.

led Lights In Uk said...

orshowc 122Thank you very much for this information. Actually I live near sea sied and I do face these problems onces in a month so now I can fix them by my self

Jignesh Dharia said...

I WANT TO KNOW HOW TO MAKE FILAMENT RICE BULB ROPE LIGHT.I WANT INFORMATION FOR WHAT TYPE OF MACHINES AND FROM WHERE I GET THAT MACHINE. ALSO NEED INFORMATION FOR THE SOURCES OF BASIC RAW MATERIALS.

Joe said...

I shuddered at you suggestion that LED Christmas lights could run off DC. Unfortunately, some already do, and spew Radio Interference (RFI) "Hash" across the radio spectrum!

The culprit is the cheaply made, non-filtered switching power supply used to convert AC to DC. You usually find it in a small box attached to the plug. They don't have to be so poorly designed- but being made in China, where they don't care about regulations, and few pennies are saved by leaving out good design requirements.

All these cheap, unfiltered, unshielded boxed switched mode power supplies (SMPS) and digital electronics are destroying our Radio spectrum by increasing the noise floor and wiping out full sections of the radio frequencies.

Even AM Broadcast radio stations are now having to increase transmit power or move to FM because that band is becoming useless. I wont even get into the cost (financial and opportunity) to the full shortwave and UHF/VHF spectrum!

Thank goodness all theses LED Christmas lights have not gone to a SMPS. Their long wire leads would act as an antennae for all the radio interference they would illegally be "transmitting" into the ether and into a radio near you. Imagine- with all the lights hung in all the neighbourhoods- the effect it would have!

I prey that day never comes.

Gary said...

This is an informative article, because I was curious about whats inside an LED light string.

There seems to be some errors though.

When calculating the voltage acros the resistors, he used an input voltage of 120 V RMS, it should be volts peak, 120 V RMS x 1.414 = 169.7 V Peak.

I calculated the following.

The peak current is:

I Peak = [(V IN RMS) x √2 - (V LED) x 25 - V Diode]/R

I Peak=[120 x √2 - 3.1 x 25 - 0.7]/4000

I Peak=[169.7 - 77.5 - 0.7]/4000

I Peak=0.023 Amps

23mA is way lower than his measured current of 44mA.

If you take his current measurment of 44mA peak x 4000 ohms = 176 V peak.
Now add the voltage across the LEDs: 176V + 77.5V = 253.5 V peak, which is not possible with 169V peak input.

I think there is an error in his current measurement.

George Plhak said...

Thank you Gary.
You are correct about the peak voltage being higher than the RMS and that I should have used that value. My point was to illustrate how the voltage was divided by the string. I didn't actually measure the 3.1 volts per LED either, that was an estimate so that could be incorrect also.
On the current scope shot, I was trying to illustrate that the current flowed for only a short fraction of the cycle and that DC drive might give better results, less flickering, more brightness and longer bulb life. I will do some more work on this. Thanks for correcting my math and for your interest.
George