Bulbs and DCC - so what if I want to use a bulb instead of an LED?

[David]

Ok, I know that LEDs last forever, use far less amps, and generate no meaningful amount of heat. But what if I want to use bulbs, specifically the ones that come with a model?

• How would I go about identifying the voltage? Assuming a 14V DCC system, I assume bad things will happen to 12V bulbs.
• Are there any considerations when wiring bulbs to a decoder? Unlike LEDs my understanding is that there is no +/- side to be worried about.
• Are resistors useful in any way to protect bulbs? Does it matter which terminal they are attached to?

For a real world example I have a Kato 115-1000. This is an old tooling - truck mounted rapido, 3 pole, bulbs. I wired the motor car with a DZ125, and now I want to wire a cab car with an FL12. While I've got LEDs, it seems that the original bulbs would fit a lot better then 3mm LEDs unless I cut down the casing. I'm also on a bit of a streak wishing that Digitrax decoders implimented PWM for on/off dimming to LEDs (or is there a way to configure this?).

So my plan is to take the bulbs from the original board here, and solder them to the decoder. I assume the 2 little bits on the light board are just to prevent current from flowing in the "wrong" direction, so that only one light goes at a time. I don't see why I would need these for DCC. The board would also go, I'll likely just add some wire to solder the decoder straight to the pickup strips instead of using those pins.

[CaptOblivious]

In general, bulbs are quite tolerant. A 12V bulb on a 14V signal isn't going to cause problems, for example. You don't need resistors, and they are, as you observe, unpolarized.

But there is one very, very important consideration when using a bulb with a decoder: The inrush current. The filament of a cold bulb presents, as it turns out, very little resistance. The filament of a warm bulb presents a considerable resistance. Thus, when a bulb is first turned on, there is a very large initial current (the "inrush" current) through the bulb—about 10 x the rated current, e.g. a 50mA bulb will have an inrush current of 500mA (!!!).

Thus, you must select a decoder that is capable of handling the inrush currents of your bulbs. In particular, the Kato FL12 will not tolerate any inrush current. :( :( So that's a problem for you right there. On the other hand, a decoder like the TCS FL4 won't have any problem with your bulbs (I use them in Kato cab cars that have bulbs all the time), but it's bigger and you might not be able to make it fit easily. But that's the price you pay…

(And the two little orange and blue bits are indeed diodes that ensure the right bulb lights up depending on the direction of travel. You can discard them.)

[David]

I've been poking around the Digitrax manual trying to see if software based dimming is supported (outside of rule 17 - I want the lights to smoothly go on and off instead of being instant). It's certainly possible using simple PWM - the tsumami decoder does it with SMD LEDs and it looks fabulous. And at the very least the motor decoders already have the timing circuits.

Anyway, I came across this in Digitrax's general decoder manual (the decoders only have 4 page pamphlets).

For regular 12 to 16 Volt lamps that draw more than
50 mA when lit, we recommend that you put a 22 to 33 ohm 1/4 watt resistor
in series with the lamp leads. This will ensure that the lamp "start-up currents"
(up to 10 times normal current draw) do not overload the outputs.

Does this make sense for protecting the FL12?

[CaptOblivious]

A resistor of that size is only going to half the inrush current; but a bigger resistor is going to dim the bulb perhaps too much. The FL12, designed for LEDs, does not tolerate very much current draw at all—about 125mA as I recall (don't have a source handy). Which means to tone down the inrush current will require something bigger.

Calculations:
normal current draw of bulb @ 12V = 50mA
Suppose:
Inrush current of bulb @12V = 500mA

So,
Cold resistance of bulb = 24ohms
Warm resistance of bulb = 240ohms

We want to limit inrush current to <125mA @ 12V
Ohm's law: R = V/I
so
R = 12/0.125 = 96

So you need a greater than 96ohm - 24ohm = 72ohm resistor to effectively limit inrush current. That's assuming that both bulbs never come on  at the same time, though! Which might be false! Although I think the FL12 does not permit such to happen, you should verify carefully, on pain of a burned-out decoder.

With a 72ohm resistor, once warm, the voltage to the bulb will be (bulbs are voltage-driven devices; brightness is proportional to voltage, not current) a bit over 9V, so about 2/3's brightness (I leave the verification of the calculation as an exercise for the reader).

[Tenorikuma]

Wouldn't a capacitor be able to soak up some of the initial inrush? (I'm a total newbie still trying to figure out electronics.)

[CaptOblivious]

Wouldn't a capacitor be able to soak up some of the initial inrush? (I'm a total newbie still trying to figure out electronics.)

Only if it was previously charged. But of course, it can't be charged before power up, which is precisely when the problem occurs. So, no :(

[inobu]

Perhaps a CDR?

Inobu

[CaptOblivious]

Perhaps a CDR?

Inobu

Perhaps? Intuitively, it seems to me that this wouldn't work, since they, as I understand them, limit current by adjusting voltage; but bulbs have the problem not of varying voltage but of varying resistance. But, what do I know? I'd like to see some tests run, because if your intuition is right, that would be a most excellent solution to a vexing problem.

[Tenorikuma]

Would inrush current limiters do the trick?

http://www.ametherm.com/inrush-current/

[CaptOblivious]

Would inrush current limiters do the trick?

http://www.ametherm.com/inrush-current/

With a name like that, I would think so! Shame they are so big, though. I had no idea there were purpose-built devices!

[The_Ghan]

Hey Cap'n,

I have the solution to David's problem: just use LEDs instead of bulbs.  Oh, wait ..... 

David, since you're going to be doing work such as soldering and the like, you may as well just replace those bulbs with LEDs.  To answer a couple of other questions you asked:
 1. The "two little bits" are diodes, which allow the current to pass in one direction only.  Thus, as the current is reversed, only one light will come on.  LED stands for Light Emitting Diode.  They also only allow current to pass in one direction.  By wiring them in opposite directions only one will come on as the current is reversed.  So yes, you don't need those little bits;
 2.  If 3mm bulbs won't fit purchase 1.8mm ones!
 
Just out of curiousity, what do you hope to gain by using the original bulbs?

Cheers,

The_Ghan

[David]

I was hoping to keep the smooth on/off quality, as well as take advantage of the fact that the bulbs are on flexible wire and are the correct size to easily fit into the little enclosure used (the enclosure keeps the forward and reverse light seperated).

Technically it should be possible for LEDs to be given smooth on/off operation (instead of instant on/off) using software driven PWM, and my F59PHI w/tsunami decoder does impliment this effect. Unfortunately the ability to understand the flexibility of software (and hardware built with that in mind) is not the strong point of most model railroaders in the US - witness how there are a dozen different SKUs for the same $100 tsunami decoder. The only difference is which sound scheme has been loaded onto the flash memory. Bad for consumers, as it makes it harder to track down the correct SKU, really bad for shops who have to carry a large number of expensive SKUs, each one with a limited market, instead of one SKU that covers the whole market.

[CaptOblivious]

I was hoping to keep the smooth on/off quality, as well as take advantage of the fact that the bulbs are on flexible wire and are the correct size to easily fit into the little enclosure used (the enclosure keeps the forward and reverse light seperated).

Which is precisely why I left the bulbs used in the cab cars on my Super Hitachi (has basically the same lightboard). Mostly, it was the flexibility of the leads. The bulbs on my model fit into a plastic housing, and getting LEDs on stiff leads to cooperate was most likely not going to happen. But I used a TCS FL4 for that install, which basically has zero issues with in-rush current. So, that's the other solution: Keep the bulbs, but use a more robust decoder.

Technically it should be possible for LEDs to be given smooth on/off operation (instead of instant on/off) using software driven PWM, and my F59PHI w/tsunami decoder does impliment this effect. Unfortunately the ability to understand the flexibility of software (and hardware built with that in mind) is not the strong point of most model railroaders in the US - witness how there are a dozen different SKUs for the same $100 tsunami decoder. The only difference is which sound scheme has been loaded onto the flash memory. Bad for consumers, as it makes it harder to track down the correct SKU, really bad for shops who have to carry a large number of expensive SKUs, each one with a limited market, instead of one SKU that covers the whole market.

I generally strongly dislike the inflexibility of most decoders available in the US. I find that inevitably something I want to be able to do, something well within the capabilities of the decoder's hardware, is impossible (usually the word used by tech support) because of the extremely rigid system of CVs. Blah.