Will an inrush current limiter actually extend the life of a headlight?

(or another question for Dan )

(I'm still figuring out a MOSFET design for a headlight relay system, and it's rather easy (and cheap) to add this feature, but we're getting into feature creep..........)

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

I don't know about extending the life of the bulb; since most blow at turnon, I'd have to say "yes, it could". OTOH, I've also read that bulbs tend to output less light over time, so, IMHO, it's not a bad thing for them to blow every few years.

As for a MOSFET relay, why? Sure, it can be done; but a reasonably sized relay should work just fine. I haven't gone looking for parts like this, so maybe there is a solid state relay that would do this; but it'd be more expensive. A discrete solution would take a number of parts (mostly to protect the fragile MOSFET gate), and for signal inversion. A mechanical relay will eventually wear out, but they do tend to last for thousands of cycles.

One parting thought: if you are making a discrete MOSFET driver, simply by ramping the gate up, instead of a sharp turn on, will get you inrush current limiting by the very nature of the slow turn on.

Theoretically, yes, but high inrush current isn't the primary main factor in headlamp filament life. It is often mistakenly believed to be so, however, because of the "flashbulb" effect (turn on the light and it instantly flashes and dies). In fact, a filament that has essentially reached the end of its life will frequently be finished off (pushed over the edge) by high inrush current, but that doesn't mean that high inrush current caused the filament to be at/near its end of life.

Much more critical to the lifespan of a filament lamp is the operating voltage. Voltage affects filament life exponentially to the power -13 (negative thirteen). Taking a filament with a rated life of 320h at 14v, let's plug in different voltages with that exponent of -13:

10.5V : 13470 hours
11.0V : 7357 hours
11.5V : 4128 hours
12.0V : 2374 hours
12.5V : 1396 hours
12.8V : 1026 hours
13.0V : 839 hours
13.5V : 513 hours
14.0V : 320 hours
14.5V : 202 hours

So that suggests we ought to get to work building a voltage limiter for the headlamps so that they see their rated output voltage of 12.8v and no more, right? Well...that kind of depends how well we want to see. Light output is exponential with voltage change, too, to the power 3.4. So, taking a filament rated 1000 lumens at 12.8v:

10.5V : 510 lumens
11.0V : 597 lumens
11.5V : 695 lumens
12.0V : 803 lumens
12.5V : 923 lumens
12.8V : 1000 lumens
13.0V : 1054 lumens
13.5V : 1198 lumens
14.0V : 1356 lumens
14.5V : 1528 lumens

OK, now what voltage do you want to feed your headlamps? Which is more important: Long bulb life, or maximum possible nighttime seeing distance?

We're all overdriving our headlamps all the time, by the way. Scary but true! The problem is that low beam headlamps' range is geometrically limited because of the need to balance forward lighting with glare control, and the range limit is such that we cannot see far enough to stop within the beam reach from speeds above about 45 mph under optimal conditions, 35-40 mph under typical conditions. OK...now which one's more important, bulb life or seeing distance?
Oh, by all means, build in a soft-start function if you'd like. It will to some degree lengthen effective bulb life. But if you are trying to maximise bulb life without sacrificing seeing performance, pay careful attention to the cleanliness of the output power. Transient spikes eat filaments for breakfast!

_________________
一期一会
Too many people who were born on third base actually believe they've hit a triple.

Yep, a capacitor plus some resistors and you have slow gate turn on.............

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

So we need 3 levels of light.

Low voltage/ long life city lights for when you can see just fine because of street lighting etc (the light is more for others to see you).

High voltage / shorter life low beams for when there's other cars but no real street lights

High voltage /shorter life high beams for when you're alone and no street lights.............

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

This what you propose has actually been done before, most extensively in the UK. See here, Here, and here.

Note if you are thinking of putting together a daytime running light setup, basing it on the headlamps (high or low beams) is generally not the best way to do it, even though such implementations are common. If you want discussion on that matter, let me know. It's too lengthy to post just on spec.

_________________
一期一会
Too many people who were born on third base actually believe they've hit a triple.

Not DRL..........

The idea is a MOSFET relay, but with a failsafe of the normal (lower voltage due to losses) wiring taking over if the MOSFET fails.

The normal wiring could be used for longer headlight life if you want...........

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

Ah, so you'd use the MOSFET to gate higher voltage to regular low/high beam, and use the regular wiring for a low voltage marker bulb? So, the stock wiring is connected to what would be considered the parking lights today, and to the MOSFET driver?

I'm not following on the 'FET comment about "failsafe". I'm not sure on the failure mode of FET's, I'm guessing drain-source short on excessive power dissapation (or alternatively, no conduction whatsoever); but usually the gate blows due to trasients on the gate. At that point, I think the transistor can be stuck on or off, but most likely off, as the gate connection is lost (or is shorted to source, which I think is the normal failure mode).

Not quite......

I've only doodled this out, so there's no picture I can post yet.

Anyway the MOSFET is essentially in parallel with old wiring. If MOSFET isn't turned on, the old wiring will light the bulb at lower voltage. If MOSFET is turned on, it would be a higher-voltage source to light the bulb. MOSFET would be turned on by old wiring with a possible 'and' condition of you wanting/not wanting the higher voltage. If MOSFET fails in 'off' state, old wiring is lighting the bulb. If MOSFET fails in 'on' disconnect the separate fuse for MOSFET and then you still have old wiring. Components (other than fuses) are hidden behind bucket, and each side has separate systems.

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

Ah, I see now. I still think a relay is best as it would have less components; but these days Rdson of FET's is getting quite low.

I'd use a PFET, should be able to get to 0.01ohms or so, less is better. TO-220 probably, or bigger--might be able a larger package rated for more amps. I'd run 100k from source to gate, and 10k from gate out of "unit". A 15ish volt zener across won't hurt either, although a TVS might be better. Maybe the bigger FET's have built in gate protection, haven't looked.

At the dash, I'd use a single pole double throw switch. Tie line to ground for extra bright, and tie to 12V when wanting "regular" lighting.

I'd also check the "other" wiring at the light, the ground wires. Make sure they are up to par also (larger and low voltage drop to the battery).

NFET (don't like PFET because they're on unless you turn them off)

TO220

.0059 ohms 50 amps 25 volts about $0.80

power to drain
source to lightbulb
old light circuit wiring to gate thru (TBD) resistor (can 'and' an on-off switch here)
gate grounded thru 10k-15k resistor
capacitor (size TBD) on gate for turn on delay (inrush current limiter).


Might have more components, but they're smaller and cheaper than a good relay............


BTW, I'm not an EE, only know enough to be dangerous

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

Sorry if I intrude too much--I am an EE, 'tho some days I only know enough to be dangerous.

Problem with your circuit is, what will turn on the NFET? Most power FET's require a gate drive of a minimum of 4V to just start to turn on; 10V is a good "on" voltage, and often more is used, up to the (typical) 20V limitation. PFET's require the gate voltage to be *below* the source voltage; NFET's require the gate voltage to be *above* the source voltage. [And for other readers, PFET source goes to more positive voltage in the circuitt, NFET source goes to more negative voltage--they flip drain/source connections.]

And that's the rub: you're going to have to put the gate of your NFET to 25 to 30V to turn it on (relative to the car chassis, or battery neg). You could achieve this with a small battery pack, wired in series with the cars 12V line, and supplying this voltage to the NFET gate.

Usually, when gating loads with FET's, NFET's are chosen due to better parameters--but they are placed into the negative lead of the load. The positive lead is left connected to the power source. A nice advantage is the "active on" nature of the NFET makes it easy to turn the NFET off and on, independant of the power supply used to power the load.

As you can guess, however, it'd be "nicer" to power by gating the positive lead to the load. This is easier to do with a PFET, as the NFET requires a signal *above* the power supply. Plus, if the supply voltage is too high, the gate must be protected, as the highest gate voltage I've seen is 20V.

Also: MOSFET's are notorious for ESD and/or transient damage. The gates are the weak spot. Hence my suggestion for both a zener (to clamp over voltage) and a series gate resistor (limits currents and also works with the zener; also works off the gate capacitance and whatever additional capacitance you might be adding).

Sorry, once I get going on this stuff, I have a hard time stopping...

Without the old circuit in parallel (this might be a problem to be figured out later)

Power NFET as relay

Initial conditions

Drain 14v
Gate 0v
Source 0v (then headlight, then ground)

Turn it on (gate > source)
Drain 14v
Gate 13v (from old wiring)
Source 0v

Current starts to flow and you get
Drain 14v
Gate 14v
Source 14v

Depending on design of NFET, you can keep current flow until voltage is removed from gate (even though drain and source are at same voltage).

Back to parallel failsafe mode: this might not work now that the source side will have voltage just about equal to the drain side and I may not be able to turn it on............ hmmmmmmm.... depends on voltage difference and the voltage difference the NFET needs.


I know some PFETS can be biased so that 0 volts is on, and any volts above that turns them off.


Or put the NFET on the ground side of circuit...............

_________________
Ed
64 Valiant 225 / 904 / 42:1 manual steering / 9" drum brakes

Ah, depletion mode 'FET's. I don't think I've seen discrete power NFET's like you're describing--I'd have to look, but the market for such items are driven by switch mode PS (SMPS), and those want the NFET driven by logic signals--no negative voltages. I don't think you will find the device that you are describing, not in the power levels required.

You still have the issue of, the signal that you want to turn on the lights is the same signal that will be "boosted" by the FET--meaning, once it turns on, it would tend to "latch" the wire at 14V. In this configuration, that is, a FET booster, you must control the gate seperate from stock wiring.

It's an interesting idea, but you'll either have to run a seperate wire to control this, or use the stock wiring to control this and lose the failsafe configuration.

Placing the NFET in the ground side is do-able, the stock wiring would just put the signal to the gate, and you'd run the pos side of the light to the battery. I'd still protect the snot out of the FET though.

If I get a chance, I can try to draw up some schematic ideas for you.

It is. They are simple, inexpensive and reliable, and this business with FETs in parallel with the old wiring means a bunch more potential failure points.

_________________
一期一会
Too many people who were born on third base actually believe they've hit a triple.