Third brake light for my 64

[Yadkin]

Glued into position:

[simplyconnected]

"Most LEDs have low reverse breakdown voltage ratings, so they will also be damaged by an applied reverse voltage above this threshold. The cause of damage is overcurrent resulting from the diode breakdown, not the voltage itself. LEDs driven directly from an AC supply of more than the reverse breakdown voltage may be protected by placing a diode (or another LED) in inverse parallel." - Wikipedia

So, LEDs are strictly DC devices that must be wired with the correct polarity. Yes, one leg is shorter than the other and you will notice on that side, there is a 'flat' at the base. The flat and short leg side must be connected to NEG. I suggested adding a diode to protect your new LEDs against reverse polarity failure. Another consideration is to NOT drive them at full brilliance if you want them to last. Bring the current down slightly and they last forever.

If you series LEDs, it works the same way as series batterys; head to tail, head to tail... In other words, the first LEDs long lead goes to POS., while the first short leg connects to the second LEDs long leg, and so on.

[scumdog]

I bought a length of plastic strip with LEDs in it, about 10" long, I pressed it into a section of aluminium with a cross-section like an E without the middle bit (sorta like a square C!)

That kept things nice and straight and looked like it was meant to be there.

I mounted it under the tail-door of my hot-rod '55 F100 and hooked it to my brake-light circuit.

It's more a 'lo-stop' than hi-stop but it sure is bright and there's no mistake that the brakes are on.

Like Yadkin I did it for increased safety.(The law did not require a hi-stop in such an old vehicle)

[simplyconnected]

There are lots of ways to do this, Tom. Our laws don't require stop lights beyond what came stock. Seat belts are the same story. But hey, I'm all for any measure of safety improvements.

Steve knows what he wants and so far it looks very good. LEDs are striking to see, especially in places where you don't expect. They're like 'eye candy' and sometimes, it's hard to look away. Remember the blue jeweled dots every hot rod had in their tail light lenses? Fascinating.

Solid State components aren't as hardy as the old relay stuff they had back in the day. I keep harping on 'reverse voltage' because it's real and it happens. Cars used to have filter capacitors to suppress radio noise on just about every contact (voltage regulator, gas sending unit, generator, ignition points, etc.). The truth is, car electrical systems have terrible spikes (noise) and just one can ruin solid state components. We think our system is 12-volts but if you measure the coil + terminal, the meter shows over 35-volts to ground with the engine running! Every time an electric motor or solenoid de-energizes, a huge spike is produced and the battery tries to absorb it. Our old cars aren't protected against counter-EMF, but new cars are, and so are LED components you buy. It's all in the design practices. - Dave

[Yadkin]

Dave, that makes more sense now. I basically need to provide "insurance" on positive voltage to protect my investment. I'm heading of the The Shack now to buy a diode, on my way to Ace to pick up a new snow shovel.

Four bulbs in series, the voltage regulator would have to spike at 20 volts to damage the bulbs so I should be OK without a resistor? Or maybe I should just consider an electronic voltage regulator? That would protect my Pertronix as well as a future sound system.

A lot of the Goop "ran" after I applied it and made a bit 'o mess, but it's entirely salvageable. In retrospect I should have used silicone caulk for that step. I just applied a third layer and that should completely seal the rear of the lower slot in the grill so I can pour resin in and near-fill the front.

[Yadkin]

OK now I'm confused. I wired 4 LEDs in series (pos to neg) and energized them with my battery charger. I checked my charger and it's running 9.5V. They all sparked then blew. Each worked fine when I tested then individually with 3V.

[Yadkin]

I was under the impression that all LEDs were 3V. This was reinforced by my testing with a 3V power supply (2 x 1.5V batteries). Apparently "super brights" are 1.85 volts and a max of 2.5 before you fry 'em. At least according to this. (About 2/3 of the way down the page.)

So I should wire 4 sets of 7 bulbs instead of 7 sets of 4.

Car voltage 13.8 to 14.4 so the lights should operate in the 1.97 to 2.06 V range.

This same page says not to connect the series sets in parallel; use a separate resistor for each set. But in this range I don't need resistors, unless there are voltage spikes.

So now my question is, what spike should I anticipate? I've done some reading and some suggest as high as 50.

If I use resistors to reduce the spike voltage, however, I won't light the LEDs under normal voltage.

Even if I could find a fuse small enough (12v 0.1A) it is doubtful it would protect the bulbs.

I'm confused.

[Yadkin]

More reading, looks like I should add this to the circuit.

http://datasheet.octopart.com/PQ05RD21-Sharp-Microelectronics-datasheet-30269.pdf

[simplyconnected]

LEDs are sensitive and they must be operated within good design perameters.

How many milliamps did you put through your LEDs? That is the most important consideration. Start low and slowly increase until you find 0.02a (20 miliamps), then back down to .017 or .018. I usually do this using a potentiometer and the power supply I intend on using. If current flow is more than .018a, I add a resistor at the value of the pot setting.

You have too many 'unknowns' going on:
9-V? Is it regulated? Put a capacitor across the leads and then measure the voltage. You may be pushing 15-Volts at the peaks, so they can AVERAGE 9-volts. The wall warts normally charge batteries, so they can be 'close'. This is different.

A 1N4004 diode tolerates up to 400-volts in the reverse direction. (1N4001 is good for 100-volts, 1N4002=200 and so on). This should be your circuit:


Each branch should be less than .02a DC.

[Yadkin]

My power supply that I intend to use is the brake light circuit. I tested with a car battery charger, so it delivered what the lamps asked for. I don't have a potentiometer on the test circuit. All I have is a cheap volt-ohm meter, which is apparently is not working very well since it measured at 9V, and that would not have blown the 4-bulb circuit.

Each bulb draws .02A, so a set of 7 should draw 0.14A.

Your circuit diagram does not show up.

[simplyconnected]

I was under the impression that all LEDs were 3V. This was reinforced by my testing with a 3V power supply (2 x 1.5V batteries)...

Batteries make wonderful testing power supplies because they are DC with no ripple. They show how your LEDs work with true 3-VDC. This is fine for a short period of time. You really need to get the miliamps under control FIRST. This is where a Volt/Ohm meter (VOM) comes in.

Car voltage 13.8 to 14.4 so the lights should operate in the 1.97 to 2.06 V range.

That's right! So test your LED's amperage using your car while it's turned on. Remember that cigar lighter plug I mentioned? Radio Shack.

...So now my question is, what spike should I anticipate? I've done some reading and some suggest as high as 50.

This is the reason why we tame down the milliamps. First of all, .02a is too bright! If you bring the current down a bit, there is plenty of leeway for momentary spikes. Your LEDs will operate at .150amps but not for long.

If I use resistors to reduce the spike voltage, however, I won't light the LEDs under normal voltage...

That's right as well. Voltage and Current work together. When you regulate your current, you are also regulating the voltage. You don't need a separate voltage regulator if you keep the amps within reason. Separate resistors aren't needed unless your voltage is way too high.

Sometimes these guys throw too many resistors at a circuit. What's the difference whether each LED gets a resistor or each branch, or just one for all the matched branches? No difference. They're still keeping milliamps under control.

Spike suppression is done with a capacitor, which acts like a battery. It quickly fills and slowly gives back any ripples or spikes. The capacitor on your ignition points is a .22-microfarad. Again, you really don't need one. Simply keep your current down to a reasonable operating level. I think the diode will do just that.

Solder a string of seven lights and the diode, and test it on your car; first with the car shut off, then when the engine is running. Tell me what the current is and how bright the LEDs are. - Dave

[Yadkin]

Dave, testing this with the car isn't an option. It's in 1500 pieces at four different locations...

I'll finish this up as much as I can now by wiring up the lamps in 4 sets of 7 in series and then parallel the four sets together. I can test it with my Jeep but that's an entirely different animal.

I know how to test DC amperage, but how do I "bring down" the current? I was under the impression that a circuit would only take the load that it needed (like a 100w bulb uses 100w and 7w bulb in the same socket uses 7w).

[simplyconnected]

...Each bulb draws .02A, so a set of 7 should draw 0.14A...

NO! If each bulb passes .02amps in a series, every one of those bulbs in the series is passing the same .02amps.

Don't confuse series circuits with parallel. In a series, the voltage drops at each diode, but not the current.
Those series Christmas Light Sets worked that way, too. Each bulb worked on 2.5-volts, then they strung 50 of them in a series.

...Your circuit diagram does not show up.

Yeah, I asked Ray Clark about that and he couldn't see it either. I think I fixed that. - Dave

[simplyconnected]

...I can test it with my Jeep but that's an entirely different animal.

Huh? Those LEDs don't know whether Jeep or Ford powers them. You can test on your daily driver if you want.

...I know how to test DC amperage, but how do I "bring down" the current? I was under the impression that a circuit would only take the load that it needed (like a 100w bulb uses 100w and 7w bulb in the same socket uses 7w).

Light bulbs have a tiny resistor wire inside. They also have a voltage rating on each one. Certainly, if you plug a 120-V light bulb in a 250-V (European) socket, it will blow.

The same thing works in your case except you are matching current to the voltage. These are Solid State devices that have a definite current rating.

I use my ammeter portion of the VOM; put one prod on the battery and the other prod feeds your LED circuit. Set the dial to read around 200-milliamps, and put the red prod in the mA hole.

This is a long series string of seven lights and a diode. If current is over .018, I add a common Radio Shack potentiometer in series, and turn the dial until the current goes down to where I want it. Again, look at the lights because they may be too bright. Get a pot around 2k-Ohms. A 1K would probably do as well. When you get to the desired current, pull the pot off your circuit and measure the resistance on the pot. That would be the resistor value you would add for that branch. Let's cross that bridge when we get to it. Somehow, I don't think you will get the full .02-amps. I was hoping my order would arrive today but it didn't. I will have a better answer for you tomorrow. - Dave

[Yadkin]

Huh? Those LEDs don't know whether Jeep or Ford powers them. You can test on your daily driver if you want.

Yeah I was referring to the 48 year difference in ages. I figured the Jeep has a cleaner voltage delivery.

Light bulbs have a tiny resistor wire inside. They also have a voltage rating on each one. Certainly, if you plug a 120-V light bulb in a 250-V (European) socket, it will blow.

The same thing works in your case except you are matching current to the voltage. These are Solid State devices that have a definite current rating.

OK I'm not familiar with SS circuits. My electrical education was in 1982: "Electrical engineering for Civil Engineers".

I use my ammeter portion of the VOM; put one prod on the battery and the other prod feeds your LED circuit. Set the dial to read around 200-milliamps, and put the red prod in the mA hole.

Understood.

This is a long series string of seven lights and a diode. If current is over .018, I add a common Radio Shack potentiometer in series, and turn the dial until the current goes down to where I want it. Again, look at the lights because they may be too bright. Get a pot around 2k-Ohms. A 1K would probably do as well. When you get to the desired current, pull the pot off your circuit and measure the resistance on the pot. That would be the resistor value you would add for that branch. Let's cross that bridge when we get to it. Somehow, I don't think you will get the full .02-amps. I was hoping my order would arrive today but it didn't. I will have a better answer for you tomorrow. - Dave

And all this time I've been trying to avoid pot. And my wife has been telling me to **** or get off the pot. Seriously though, that makes more sense now. Off to The Shack in the morning.

NO! If each bulb passes .02amps in a series, every one of those bulbs in the series is passing the same .02amps.

Don't confuse series circuits with parallel. In a series, the voltage drops at each diode, but not the current.
Those series Christmas Light Sets worked that way, too. Each bulb worked on 2.5-volts, then they strung 50 of them in a series.

Wait! Each bulb (resistor) uses .02A of electricity (which we will set to a little less). So the total load is 28 times that.

Yeah, I asked Ray Clark about that and he couldn't see it either. I think I fixed that. - Dave

Got it now. That's what I figured.