$1 Prius LED Brake Light Repair

wow, you are adventurous, well done!

 

great writeup!

 

LED is becoming a dirty word for me, wonder if I'll rue that we have LED headlights. They're nice, but I've heard horror stories about replacement cost.

 

I don't know much about led bulbs but won't the other ones that you didn't replace soon fail?

 

Thanks!!! I always wondered about the LED's but know (from experience) better than to open up working parts to satisfy my curiosity

The following story is long, too long for many and not meant as critizism toward Mike Carambat!!!!! He had nothing to loose and made a valiant effort to fix a lost cause and still managed to succeed!

That being said, here's some questions, calculations, thoughts, etc.

Judging by your pictures, the dummy load is four 43ohm resistors in series and the current limiting LED's are four 5.1ohm resistors in series. Is that correct?
So roughly the dummy load is drawing 80 milliampere. (From your words, I gather this is the left one and the right one is half that resistorvalue and therefor double the current: even using a relay (e.g. 100mA draw for the coil) for a trailerhitch would already significantly mess with the total of only 300mA or so for a brake light, let alone simply hooking it up in parallel (like most people including me have done) but that's another story...)

The LED's are in a clever arrangement as you say, which is basically two LED's in parallel, series wired with two more sets of parallel diodes.
This does mean more LED's will keep burning when one fails (great they keep working when one goes for safety reasons!!!), but at the expense of the other ones failing faster too.

I am puzzled by the fact you had multiple LED's fail. I would expect two to fail, being two that are wired in parallel: one fails, the other gets more current to handle and goes too. I did see something similar in a 7-LED-domelight in a car recently, but I cannot properly explain it.


Some general information on LED's:

As said, they are Light Emitting Diodes. They emit light when you run a specific current through them and only when going in the right direction (like other diodes, they block current in one direction and let it pass in the other direction). I say "specific" as this current is very important and should be between strict boundaries which are LED type and model specific.
That's why it is very dangerous to attach a battery directly to an LED. It will either not light up or burn out.

The reason it does not light is that an LED needs a minimal voltage to work, e.g. 1.8V for a typical 1990's red LED. If you use 1.5V, it will probably not even light up at all. If you use 2V, it will probably burn out already, that is how tight the limits are.

So although the two D-cells in series worked a treat to find the flawed LED's, it is not common practice to find defective ones this way.

In school I learned that an LED is ALWAYS to be used by a current limiting resistor. In the above example of the 1990's 1.8V LED, it will draw e.g. 10micro ampere (0.00001A) at 1.5V, but 2000 milliampere (2A) at 2 volts for a second (and then nothing because it burned out).

The datasheet of said LED will give you information on what current the LED likes (e.g. 10milliampere) and what voltage it will drop at that current (1.8 volts).

By using a resistor in series, you can limit the current the LED will be getting from e.g. a car (12V).

We know you want 10milliamperes to run through the resistor and the LED. And we know the LED 'eats up' 1.8 volts, so the resistor needs to 'eat' (12V minus 1.8 volts=) 10.2 volts.

There is a very simple and often used formula telling us that Voltage (V) is current (I) times resistor value (R), or V=IxR. Reworking this gives R=V/I. So the resistor value should be 10.2 volts divided by 10 milliamperes is 1020 ohm.
There is a system behind what resistors are available, and we could find a 1020 ohm resistor in a good electronic component store that stores E96 range resistors, or we could get a 1000 ohm resistor in ANY store that caries resistors. 1000 ohm would do quite nicely.

Just a thought experiment: what current will run through the LED when the car starts and produces 14.4V? The higher current might mean the LED now 'eats up' slightly more than 1.8V, but we'll disregard this minute difference. The (hypothetical) LED is rated for 10milliamperes typical current, 5milliamperes minimum current and 20milliamps maximum current, so will work fine when within these limits. Will it work in our example once the car is started?

Answer in white: Resistor now needs to drop 14.4 minus 1.8 is 12.6 volts. We chose 1000 ohm, and reworking the formula gives I(current) = V/R or 12.6/1000 is 12.6 milliamperes. The LED will light brighter but still be well below its maximum currentrating of 20mA, so it is safe to use like this.


You see here (assuming you got the same results as me) that the resistor "limits" the current through the LED. Even when the voltage fluctuates.

Back to the Prius example: assuming the LED's 'eat' 3 volts a piece, there is still ~5 volts to be dissipated by the resistors (14.1 is the voltage often seen on my Prius, minus three times 3V (three LED's in series) = 5.1V).
Assuming the four resistors are indeed 5.1ohm, gives us exactly 250mA of current going into the LED array (5.1V/(4x5.1ohm) = 1/4 = 0.25). Since there are two equal paths, each path gets half of this current, so the LED's get 125mA each (slightly off from the value Mike found).
The reason I am doing this calculation is to warn that it is quite likely that the new LED's are going to die a lot faster, as they are being run at a far higher current than their typical currentrating.
I couldn't find a maximum rating, so not sure it will last.

It would be best to change all LED's and adjust the current limiting resistorbank accordingly, but I am not sure that would be legal...

The LED's are rated at 70mA typical, 2.6V typical and 3.3V max.
Maybe I can gets get my point across when doing the calculation in reverse: with six Radio shack LED's in place, what voltage would the light work at with typical voltage and current:
Each path has three 2.6V LED's in series, which is 7.8V. And both paths use 70mA. So the current limiting resistorbank (of four serieswired 5.1ohm resistors equals) 20.4ohm gets 2x70mA=140mA run through it.
V=IxR = 140mA x 20.4ohm = 2.856 volts.
7.8 volts is dropped by the LED's and 2.856 by the resistorbank when using typical values. This can only be achieved by supplying (7.8 + 2.856) = 10.656 volts.
Any higher supply value will ram more current through the LED's. At 14.4V, each LED is already getting (worst case) 160mA (!!!)

 

Totally true.

Provided the voltage is low enough, then yes. But if the voltage gets slightly higher, the current will skyrocket.

Here's a nice graph from this link:



It shows particularly clearly what happens to both a resistor and an LED when you apply a voltage.

The resistor: 1 volt, 10 milli amps. 2 volt, 20 milliamps, 3 volt, 30 milliamps, etc.

The LED: 1 volt, no current, 2 volt, 20 milliamps, 3 volt, no current because it is blown.

 

I dit the same LED echange. It does not look pretty at all. I am using now some epoxy glue to put everithing back. I hope I can still get the correct angle of the srews since the upper most one had to be cut out together with the case.
Bottom line is that the blake LED's work now.

 

My 2004 has a dead brake light array so I need to make the same repair, and I am of the same mindset as the original poster, do this ONLY if you like a somewhat lengthy repair because you like expanding your knowledge - and you are Cheap! ( I am cheap)
I pulled the light assy off the car in the normal way. I started opening up the gap between the outer clear lens and the black body. After a bit I could see that hatch side (vertical edge) would pull away from top to bottom and the top edge and bottom edge as well. All by wiggling with a 1/2" wide thin steel blade ( it is a worn tool so all of the edges are rounded, this is ideal to not create sharp pressure points when prying thin plastic apart. All went well for 3/4 of the perimeter. The outer edge would not give up to prying, so I got the heat gun out and applied heat to the black body. It worked. It took some finesse but I got the clear lens off. Now the assy comes apart for the LED repair.

 

The alternative is close to $150? For a burnt out light bulb?

Tail Lamp Assembly - Toyota (81561-47071) | Toyota Parts

 

I have this problem on my passenger's side tail lamp.
I want to solder new LEDs in there but am not sure which LEDs are the best for the job.
Can you anyone give a little guidance?

The brighter the better of course but I only want to do this job once and not have to get back in there to replace LEDs. Better to put the best & brightest in there now.

Thanks in advance!

 

If you do put in very bright LED's, do try to match the light output to the other side and the remaining ones (if any) in the housing you're replacing them in.

Also, the best way of giving them a long life is to not run them near their maximum: 20mA LED's run at <10mA will pretty much live forever.
I think the Prius LED's are run near their maximum and still last ages, but do fail once in a while. Since LED's are certainly better than 10+ years ago, should make this a once-in-a-lifetime job.

But I realise I haven't actually given you a brand or type to buy...