Dimmable LEDs?

[TomoHawk]

I talk to car people occasionally (over coffee & pie..) about things.  One of them is the Dimmable LED bulb.  I never thought of them as 'dimmable' in the sense of what a filament  bulb can do, so I wanted to ask to get some information in the context of S30 cars.  Are they useful for us?  Do you need special wiring or controls to dim them?  Should we even bother with them?  The only place I would use the dimmable LEDs would e in the gauges, because it's really a PITA to  change them every other year (I use the #55, 7V bulb for the gauges.)

 

What are some advantages and disadvantages?  If you 'Google' it, most of what you get, is for home-use LED bubs, like you use in your hallway or kitchen.

 

I had over the last winter, replaced all the bulbs (except the gauges) on my  '78 with LED bulbs, but I have since changed them back, because it just wasn't as bright as they were before (for the brake  and parking lamps.)

 

thxZ

 

BTW-

 The 4W bulbs MSA sells for the gauges are GE 1816 bulbs (12V, 4.3W)

Edited October 6, 2015 by TomoHawk

[Stanley]

The headlight for my hardhat has dimmable LED's (uses 3- AAA batteries) so they exist, FWIW.

[Captain Obvious]

I believe that all LED's are inherently very dimmable. Assuming you don't overheat it, the more current you push through the die, the brighter it gets. Works great from almost no light emission, all the way up to max brightness where you can't dissipate the heat anymore.

 

But in the world of using LED's for "real world applications" like home lighting and automotive lighting, it's not that simple.

 

The problem is that the LED's you purchase for pre-canned applications (like 120VAC home lighting applications or 12V automotive lighting) are more than just a simple LED. They are multi-component assemblies and unless you know what comprises that "assembly" you cannot guarantee how well they will react to a dimming attempt.

 

Some of the assemblies are simply a couple LED's in series with a current limiting resistor. These should gracefully handle an attempt to dim them. But on the other hand, some of the assemblies are much more complicated than that and attempt to account for wider supply voltages and these might not react favorably to an attempt to dim them.

 

Then there's the WAY they're dimmed... The original resistor potentiometer based dimming scheme won't work if you convert ALL your dimmable bulbs to LED. You'll need to keep at least one incandescent bulb in the system somewhere, or switch over to a PWM based dimmer system.

 

 

[TomoHawk]

Very cool, Brice.  I suppose that in time the LEDs will get 'smarter' and be easily dimmable, but by then, incandescent lighting will have gone the way of the pulse-dial telephone...

 

LEDs will probably end up having 3 pins; one for the anode, one for the cathode, and one for the dimming control.

Edited October 6, 2015 by TomoHawk

[Kurbycar32]

Captain I'm pretty sure that a circuit containing only LED's will effectively be dimmed using a regular potentiometer.  The design of the circuit matters though: if a series of LED's are wired in series, each successive LED will be more dim due to the consumption of the LEDs that came before it.  However if the LED's are wired in parallel the circuit would behave normally with working dim function.  I don't believe an incandescent bulb is required in either case unless we are including a filament based flasher that requires load to function properly.  In the case of a series type circuit I agree that a PWM circuit would be ideal and in the case of a parallel circuit the PWM would also work but might not be necessary.  For the uninitiated a PWM (pulse width modulation) circuit turns the circuit on and off really fast to dim, rather than a consistent voltage drop (via variable resistance) that a potentiometer would do.

[Captain Obvious]

Kurbycar, Yes, LED's can be dimmed just fine using a regular potentiometer, but there's a catch... The catch is that the potentiometer must be of an appropriate value sized for the current going through it, and the stock dimmer pot is not of appropriate value for LED's. Goes like this...

 

The stock dimmer pot is adjustable from 0 to 10 Ohms. When it's cranked to 0 ohms, you get full brightness. No problem.

 

Problem comes into play when you crank to full dim... You're putting a 10 Ohm resistor in series with the LED (or LED's if it's a bulk string) and they could care less. The LED's draw so little current that there's very little voltage dropped across the 10 Ohm pot and the LED's still light up at near full brightness.

 

If you include one incandescent bulb somewhere (or a low value resistor instead) wired in parallel with the LED's, that one incandescent bulb will provide a low impedance path around the LED's and allow significantly more current to pass through the 10 Ohm pot.  The additional current through the pot will increase the voltage drop across it and the LED's will have less voltage to work with.

 

You probably still won't get full brightness range, but at least you'll get more range than you would without the bulb. If you really want full brightness range, you need to switch to a higher pot value (probably on the order of a couple hundred Ohms) or switch to a PWM type dimmer.

 

http://www.classiczcars.com/topic/52446-pulse-width-modulated-pwm-dimmer-control-upgrade/

[Captain Obvious]

And one other thought about leaving an incandescent in the system. I suspect the people who have replaced their dash gauge bulbs with LED lighting have done just that... Either by accident, or on purpose, I suspect they have left a bulb or two (or more) as incandescents. Depending on year and options, some possible culprits include:

 

HAZARD switch, cigarette lighter, HVAC faceplate, A/T shifter bezel, original radio.

 

If you got a 260 with an auto trans, you could leave FIVE incandescents in the system and still have LED's in all your gauges. 

[Kurbycar32]

As usual Captain that was super detailed, well done.  Assuming the stock rheostat/potentiometer is in use i gather that you are using a higher draw device (incandescent/resistor) to change the scale of the voltage drop from the useless 10 ohm drop.  Thats a cool trick; assuming an all LED system was in place you would need I'm guessing around a 400ohm potentiometer to effectively dim the circuit which is exponentially more than the stock unit can handle.  Ill have to try that out

[Captain Obvious]

The problem is the huge disparity of currents between LED's and incandescents. If you make the pot large enough that the LED's will get completely dim, then your incandescents will all go completely dark in the first ten percent of rotation travel of the pot. And if you make the pot small enough that you get reasonably good control over your incandescents, your LED's won't ever get completely dark.

 

I think the best solution is to use all one technology or the other and not try to mix the two. And if you DO have to mix the two, a PWM dimmer would do a much better job of handling the two differing technologies because it's a duty cycle based control instead of a resistance based current limit.

 

I think that if you do switch everything over to LED's, a 400 Ohm pot would be a good place to start. You could use traditional fixed value resistors first though just to test the brightness and then once you're confident on the upper end, buy the pot. But be warned... I expect you're going to run into the same problem with the pot shaft being too short and will need some sort of extend like what I did for the PWM module I made.

[TomoHawk]

I tried out this concept today.

 

I installed a BA9 LED bulb (the 5 SMD type) in the heater control panel, and it really did dim  with the other regular bulbs that were in the gauges.  I adjusted it from fully-off to fully-on using the stock rheostat.  I will get a bunch of white LED bulbs and start putting them in the gauges and indicators.  

 

Would this LED bulb work for gauge illumination?  It's 1Watt.

Edited October 10, 2015 by TomoHawk

[Captain Obvious]

Cool. So far. so good. I suspect that as you increase the number of LEDs in the system (and reduce the number of incandescents) your "range of dimming" will decrease. You'll always still be able to get full bright, but the "dimmest" lower end will move up.

 

And with that in mind, I think I would concentrate on LEDs that are dimmer rather than brighter. Everyone selling LEDs seems to advertise how BRIGHT they are and all the customers seem to think that the brighter, the better.

 

I've done a little bench testing with a few LED bulbs in spare gauges and I've come to the conclusion that you don't need much. Now, I also know that how much brightness you want from the dash is different for everyone, but I'm not looking for way brighter than stock. I want white, not necessarily a whole lot brighter. YMMV.

[TomoHawk]

Good,  thanks for the note. 

 

I'll keep checking things as I put in LED bulbs.    Keep in mind that some of the lights don't need dimming, like the turn signal, brake, high beam etc. indicators, so you could just use the single-LED bulb for that.  You only really need those 7:  speedometer (2), tachometer(2), oil(1), fuel(1), clock(1) and heater(1) so if you can leave the rest as regular bulbs, things should be fine (I think.)   You just can't get those rare little  things that go into the hazard and defroster switches, so maybe whatever LED you can get into those will be fine at whatever brightness they get.  Personally,  I don't need to see every little mark or speck of dust on the gauge face; just the position of the needle (except the car tachometer of course.)  I've been getting the super-bright ones so far, but just a few.  Maybe I'll look for the warm-white kind and try those.

Edited October 11, 2015 by TomoHawk