Melting Solder on LED Indicators

[Duvn8r]

Hi All,

Hoping someone can impart some electrical wisdom in my direction.

I've been customising a 2000 Cali Special over the past couple of years which lead to me replacing the indicators and taillight with some LED items.
These things have got the housing shape I want but after the install, which included a new LED flasher can, and the creation/installation of the requisite diode bridge thingy to stop them all flashing at once, I rode a few times only to have the rear indicators drop out one at a time.
Seems I've got this problem where the rear indicators are melting at the solder connection of where the leads meet the circuit boards of the LED's.
I thought that it must've been a dodgy solder connection thing (i.e. cheap aftermarket part...) so i resoldered and lo and behold within a ride or two they were gone again.

Front indicators are working fine at this stage and still are and so is the taillight....

Fast forward to me thinking I can fix this, so I make my own using two 10mm LEDs and a resistor on each side (even use the proper formula to calculate...) and I've got these awesome super bright rear indicators now, problem solved...but alas after the next decent ride, one of them is gone again. Not only that, I remove the cover to find that both wires have detached and the leg of one of the LEDS is melted off, like gone,  and embedded in the solder on the end of the wire.

My Question is why am i getting so much heat at the LED-to-wire connections and what can I do about it?
Is it just about adding more resistance?
I think the earth goes back to battery, would it make a difference earthing to the frame?
Or does it need a heatsink?

Would kinda like to take some of the guesswork out of the equation if I can...

Thanks in advance

[wirespokes]

Welcome to the forum Duvn8r.

That's an interesting problem. Since front and rear are the same, the problem is exterior to them.

My guess is a poor ground. Perhaps try running a jumper straight to battery minus as an experiment.

[guzzisteve]

I would certainly check the lead harness from battery area where it plugs into main harness. Should be under L/H sidecover. THEN look under fender where the wiring runs to get to the back. Run all grounds to frame and check the bikes main ground from batt to engine & frame, yours grounds through the battery plate mostly for engine to frame. You ruled out cheap indicators when you made your own.

[Wayne Orwig]

Tell us what resistors and LEDs you are using.
I assume you do not have an obvious short.

[berniebee]

Did you replace the indicator housing/lens with a smaller after market item? Smaller indicators trap the heat generated by the LED lamps which could lead to early failure or melting issues.
 Are the rear indicators dual filament, which means they stay on constantly when riding? Again, it could be heat build up.  (I'm assuming Australia has daytime running light laws.)

I've seen LED over heating issues, but it's usually the surrounding plastic that melts or LED driver boards that get charred or fail. Solder melts at 400F (200C), so something is getting really hot. OTH flames streaming out of the indicators while riding on the highway would make for an interesting sight. 

BTW, what value of resistor did you put in? Is the resistor value written on it, or band colour coded?  (Have you measured the resistance with an ohmmeter?)

[Kiwi_Roy]

It's hard to imagine the rear lamps having time to melt in the short duration of using the flashers, maybe you have a short back there, can you have someone follow you for a while to report anything unusual?
+! on the list of components, picture of any joints.
https://www.thisoldtractor.com/guzzi007/schematics/2002_Stone.gif

[Duvn8r]

Tell us what resistors and LEDs you are using.
I assume you do not have an obvious short.

The LED's are 10mm Orange 2V 20mA
the resistors are 430 Ohm metal film
I used 2 LEDs in series with one 430 Ohm resistor.
No obvious short noticed , I mean I even ran them and the bike constantly for a few minutes after install to see if they would die but they worked perfectly.

[Duvn8r]

thanks everyone for your comments, greatly appreciated.
Just to wrap up responses all in one:

@Kiwi_Roy - could be a short but they were working fine in the first instances, will chase the wiring i think, ta also for the schematic link it's the one i've been using.

@Berniebee - Yes flames out the indicators would look very cool! (i'm thinking piezo/gas indicators, could be a new thing!!!...) They are a full indicator replacement with an anodised aluminium housing, they are not dual filament i.e. they are not constantly on like a running light they are just indicators only.  interestingly there's no charring on the circuit board....

@Wayne Orwig - as per response - The LED's are 10mm Orange 2V 20mA
the resistors are 430 Ohm metal film
I used 2 LEDs in series with one 430 Ohm resistor.
No obvious short noticed , I mean I even ran them and the bike constantly for a few minutes after install to see if they would die but they worked perfectly.

@GuzziSteve and @wirespokes - as above and to your points i think i need to follow the wiring and check earthing

Again, greatly appreciate all of your thoughts and insights.
I think I'm gonna start following the wiring to check the quality of the earthing circuit....and pray to the electron gods.....

[Murray]

Poor ground is unlikely its drawing all the current. what's the value of the resistor per chance 14.6 volts should be used as the voltage value for calculations, I am assuming you are using the resistor to drop the voltage down to stop fuzting the diode.

Do you have a mulit meter that can measure current (you'll need to put it in series to get a current reading). What is the rated current draw or wattage rating of the LED's?  The other thing I don't get is why you need a diode in the circuit to stop all four flashing at once. The switch determins which side flashes.

[Wayne Orwig]

The LED's are 10mm Orange 2V 20mA
the resistors are 430 Ohm metal film
I used 2 LEDs in series with one 430 Ohm resistor.
No obvious short noticed , I mean I even ran them and the bike constantly for a few minutes after install to see if they would die but they worked perfectly.

So, 2V per LED, two in series, so 4V. Say the bike is charging at roughly 14V, that is 10V across the resistor, or about 23 milliamp. (I assume the resistor is in series too of course)
23ma @ 10V means .23 watt. You should have a 1/2 watt resistor. But then even at 1/4 watt or so, there isn't really enough heat there to melt solder. And I assume when the happens the resistor isn't burned up.

BTW, that wouldn't make a lot of light. If it is really bright, you may have something miswired.

[wirespokes]

Yeah, bad ground doesn't make sense - more like the ground is being shorted to a hot lead. Gotta be some sort of short to the rear signals.

[Kiwi_Roy]

So you rolled your own indicator lamps. If ever I do that I put the resistor in the Hot wire that way if there happens to be a short the resistor will limit the current.
Is it possible the resistor is shorting out putting too much current to the less?

[Duvn8r]

Poor ground is unlikely its drawing all the current. what's the value of the resistor per chance 14.6 volts should be used as the voltage value for calculations, I am assuming you are using the resistor to drop the voltage down to stop fuzting the diode.

Do you have a mulit meter that can measure current (you'll need to put it in series to get a current reading). What is the rated current draw or wattage rating of the LED's?  The other thing I don't get is why you need a diode in the circuit to stop all four flashing at once. The switch determins which side flashes.

Hi Murray, yep resistor in series to stop things smoking away, or at least that was the intent. I'm beginning to think I've underdone the resistor value by using the wrong forward voltage i.e. 2V instead of 4 (coz i used 2 globes per side...)

Forward current is 20mA
Forward Voltage is 2V

with regards to the diode to stop the 4-way hazard light effect, the Guzzi's earth the dashlight out thru the other side indicator (incandescant Bulb) to what's selected when you turn, instead of a separate earth lead up front if you get what i mean, so you need a diode set up to stop the current flow thru the dashlight when you switch to LED's coz they have no resistance ...or something like that

[Duvn8r]

So, 2V per LED, two in series, so 4V. Say the bike is charging at roughly 14V, that is 10V across the resistor, or about 23 milliamp. (I assume the resistor is in series too of course)
23ma @ 10V means .23 watt. You should have a 1/2 watt resistor. But then even at 1/4 watt or so, there isn't really enough heat there to melt solder. And I assume when the happens the resistor isn't burned up.

BTW, that wouldn't make a lot of light. If it is really bright, you may have something miswired.

Hi Wayne, yep resistor in series and no it wasn't burned up and yep it was 1/2 watt...I'm thinking more resistance as mentioned above.
Also the LEDs have an intensity rating of 14,000mcd so they are pretty bright, i think normal ones are around 1500-3000mcd...i had a mate complain that my indicators as they were a bit dim so i thought i'd go the hack for brightness!...

[Duvn8r]

So you rolled your own indicator lamps. If ever I do that I put the resistor in the Hot wire that way if there happens to be a short the resistor will limit the current.
Is it possible the resistor is shorting out putting too much current to the less?

Hi Kiwi Roy,

Yep resistor on hot wire..as mentioned I think i might've under done it resistance wise....so I'm still scratching my and and it still doesn't explain why the bought ones melted.....maybe its the LED flasher can??

[Duvn8r]

Thanks all for your input, it's gonna be a while before I can get in to try any fixes or experiments.
happy for any other opinions or a fresh set of eyes over the calculations.
Worst comes to worst I can go back to hand signals!!

[Kiwi_Roy]

You are just overloading the LEDs, if you assume 14 Volts the current works out at 23 milliamps.
It could also be that the specs for your lamps are BS, how about increasing the resistors to 560 Ohms or even the next preferred value, 1/4 Watt should be ok.
I don't think it will make a great deal of difference to the perceived brightness.
There's not much heat in the circuit but perhaps add a bit of a heatsink where the resistor meets the LED lead.
Can you show us a picture of where it's melting so we can better understand the problem.

[acguzzi]

for a resistor to dissipate its rated power often requires temperatures above 200C depending on what type of resistor it is, as already pointed out 2X power rating is a minimum in practice, we are usually 3X to 4X. One end of the resistor is being cooled by the supply wiring the other is connected to a diode which is a heat source not a sink so the temperature is going to rise, it's not a lot of heat but it has to go somewhere or the temperature increases. You have to allow the heat to get out, you can put a heatsink in there, which could be just copper wire, or improve airflow, maybe both. Commercial LED lights are usually PCB mounted and the copper cladding spreads the heat and keeps things acceptable. You need to get the heat out somehow.

[Wayne Orwig]

Hi Wayne, yep resistor in series and no it wasn't burned up and yep it was 1/2 watt...I'm thinking more resistance as mentioned above.

Nah.

A 430 ohm resistor can not practically pass enough power (say about 23ma, or about 1/3 watt TOTAL at 14V), to melt solder.

Something else is going in if the LEDs and resistor are in series.
Is the resistor yellow, orange, brown?

[oldbike54]

 This is odd . Have converted the Bassa to LED flashers , it only required the correct flasher  unit .

 Dusty

[StuCorpe]

with regards to the diode to stop the 4-way hazard light effect, the Guzzi's earth the dashlight out thru the other side indicator (incandescant Bulb) to what's selected when you turn, instead of a separate earth lead up front if you get what i mean, so you need a diode set up to stop the current flow thru the dashlight when you switch to LED's coz they have no resistance ...or something like that


I had the same problem on my 03 EV when I went all LED.  I took the dash off and added a resistor and led from each side of the turn signal indicator light socket to ground.  Drilled two holes and inserted them on each side of existing indicator socket.  Solved the problem instantly and now have separate left turn/right turn indicators.  Kiwi Roy did not believe there is enough current flow through the indicator lamp to make 4 way flashers out of the system but it definitely will.

[Kiwi_Roy]

  Kiwi Roy did not believe there is enough current flow through the indicator lamp to make 4 way flashers out of the system but it definitely will.

That's strange I have been saying all you need is a couple of small diodes and a single re-wired idiot light to chassis for at least 15 years.

Depending on how they are configured LEDs in series or LEDs in parallel some lamps will operate on a surprisingly small current.
A Griso for instance passes 40 microamps through the lamp filaments to check if they are ok, this is there even with the bike off and key out, enough to light some LED lamps.
All it takes is a 1k resistor across the lamps to drop the Voltage below the forward bias level.

This is odd . Have converted the Bassa to LED flashers , it only required the correct flasher  unit .

 Dusty

I believe that the original poster is using bare LEDs, not 12 Volt lamps with an in-built resistor.

Typically you would have to modify the idiot light wiring to prevent cross talk from Left to Right.

 
 

[acguzzi]

a 1/4W wirewound resistor rated at 200C reaches 200C at 1/4W dissipation with the lead temperature held to 25C, now imagine what happens if the lead is not cooled, the connection will also rise to 200C driving the resistor temperature higher, above the melting point of solder. Even a small amount of heat will cause high temperature if there is no cooling, the resistor is dissipating 250mW and the LED is dissipating 40mW if there is nowhere for the heat to go then the solder will melt.

[Wayne Orwig]

a 1/4W wirewound resistor rated at 200C reaches 200C at 1/4W dissipation with the lead temperature held to 25C, now imagine what happens if the lead is not cooled, the connection will also rise to 200C driving the resistor temperature higher, above the melting point of solder. Even a small amount of heat will cause high temperature if there is no cooling, the resistor is dissipating 250mW and the LED is dissipating 40mW if there is nowhere for the heat to go then the solder will melt.

The heat will exit on the wire. I have designed items like this a thousand times, well, maybe a few hundred. If you were close to 1 watt, you could heat the resistor AND wires enough. But with less then a third of a watt, nah.

[barrypenn]

Totally lobbing a guess here. Could something on the LED boards, or the boards themselves be conductive, allowing more current through than the solder points are capable of ? Perhaps use a VOM set to DC amps or mA inline of the positive wire,  and see if there's parasitic leakage through, or at least excessive amp draw when the circuit is active ?

[acguzzi]

My profession is designing electronics for vehicles and like many here I am old and have done this for 50 years, and while it is correct that the heat will exit from the wire I point out that only one end of the resistor is connected to a wire and the other is connected to the LED, in order to be able to dissipate its rated power without exceeding its maximum temperature (which could be 150C or 300C depending on the resistor) a resistor needs to be cooled at both ends (at 25C, at higher ambient it will get hotter). In this case it is not cooled at both ends. Embarrassingly a board I had designed recently melted a 1W resistor off the board, it was dissipating about 250mW, because the thermal path was insufficient to cool the resistor, it was cooled by convection and through other components that were also heat sources, much like this case. If a component over-dissipates the solder can melt even at 25mW.

[Wayne Orwig]

Being curious, I ran a quick test. I did not have a 430 ohm, but I had 470 and 390 ohm , 1/4 watt. I used the 390 ohm, to pass more current and generate more heat.
I cut the resistor leads short, and tacked some 30 awg wires close to the resistors.
I applied 12V for a long time, no heat at all that I could feel. 13V felt a little warm after a while. Sitting at 14V for a long while, and it was too painful to squeeze the resistor, but no issues. Then 16V for a long time. Then 17V for just a few minutes. While cranking it towards 18V the solder finally melted. (it may have melted at 17V if I had waited longer, but I was well above what I wanted to check so I didn't care any longer).
So somewhere around 17V @ 390 ohm, or 3/4 watt on a 1/4 watt resistor, the solder melted. That is old lead/tin solder BTW, not higher melting point ROHS.
About what I expected. And 14V at 390 ohm, or 1/2 watt in that 1/4 watt resistor, was not an issue.
I'm sure the original LED setup had wires heavier than 30 awg that will carry away the heat, so I'm still confused as to what is going on.

[Kiwi_Roy]

, so I'm still confused as to what is going on.

Me too, especially because the lamp is only powered about 50% of the time as it flashes On and Off.
However solving it is probably a lost cause, i'm not going to waste any more time on it, the original poster needs to use some proper lamps with proper lamp holders not something from a hobby shop.

[wirespokes]

The big mystery as I see it is why are the front ones working, yet the identically constructed rears failing? What's different about the rear installation?

[Duvn8r]

Hi All,
I think I've solved it!
Sorry for the f@#k around and thank you for all of your input, big call out to Wayne for doing a test to try and replicate the issue.
Also I actually thought Jaycar was a good place for aftermarket bits and not a "hobby shop"...

Anyway...

So I woke up this morning at 5 am to a brainwave, this is not a HEAT or electrical issue, this is a VIBRATION issue!!!
So as I've been customising the bike I've made and attached a carbon fibre guard which bolts solid to the rear frame to which, in my infinite wisdom, the indicators are mounted with no vibration isolation.....

This explains why the front ones are fine, why the first lot of aftermarket rears failed in the same way, why there's no burn marks, why the LED legs have snapped(not melted) clean off and so on......
So now it'll be a strip out, rewire and remount to try and isolate the area a tad from the ground running vibes.
I'll try throw some pics in so you can see what I mean.

thanks again for all of your help