What bulb and what resistor???

[pdlethbridge]

While researching the use of bulbs and resitors on the net I found this info at tony's trains that might help someone besides myself.
http://www.tonystrains.com/technews/install-lamps-decoders.htm

[Yampa Bob]

That's a good site.  Here are some more, with useful information when you don't know the amperage rating of the lamp.

In the first link, note the caution about using 12-16 volt bulbs with decoders. 

http://home.neo.rr.com/mrwithdcc/primer.html

http://www.loystoys.com/railline-news/rln_9705_9706.html

http://www.wiringfordcc.com/gorhlite.htm

One of my indispensable testing instruments is a resistor substitution box.  If anyone would like to build one, Kalmbach's book # 12407, "DCC Projects and Applications" has a schematic and instructions.  If you don't want to buy the book, here is a way to get the information online.

First follow this link:

http://kalmbachcatalog.stores.yahoo.net/12407.html

Right click under the title, where it says "View a portion of this product", and download the sample PDF spread.  On the last page is the schematic and instructions. It's very easy to make, using Radio Shack components.

I ordered the book, as it has lots of information on installations. A good companion book for those considering DCC is #12417, "The DCC Guide".

http://kalmbachcatalog.stores.yahoo.net/12417.html

[Yampa Bob]

Paul, I'm glad you started this thread.  I've been experimenting with low voltage lamps and LEDs, very interesting topic.

Upon reading the pages you and I linked to, I found some contradictions.

At Tony's Trains, it says a 10X inrush current can damage the decoder function. Yet in the next section it says a 12-14 Volt 50 ma bulb can be connected direct to the decoder, a 75 ma bulb requires a 22 ohm resistor.

In essence, are they are saying that as long as the inrush doesn't exceed 500 ma, the function won't be damaged?

At the Neo site, it says the 10X inrush probably won't take out the function right away, but over time it can damage it.  The recommendation is to add a 47 ohm resistor, to protect the function while not dimming the light very much.

Bulbs/lamps are cheap, decoders are not.  I would think the prudent thing to do is add at least the 22 ohm suggested by Tonys, or the 47 ohm suggested by Neo.

I tested several 12 volt GOW bulbs salvaged from old locos, and determined that the "cold" resistance was approximately 41 ohms, the "hot" resistance at 415 ohms.  At 12 volts, this calculates to an inrush of 290 ma, and 29 ma fully lit.

The cold resistance is not very accurate but close enough to confirm the 10X inrush. If a 47 ohm resistor is added, the total cold resistance would be about 88 ohms, for an inrush of 136 ma, just slightly over the Digitrax rating of 125 ma.

Granted, the inrush duration may be only a millisecond or microsecond, I don't know the safety factor of a decoder function.

As I understand it, the safest method is to use a 1.5V GOR rated at 15 ma.  The 150 ma inrush would probably be considered a safe value.

I'm hoping Jim will weigh in on the implied 500 ma "safe" value for 12 volt bulbs, and correct any of my calculations.  I know he has a precision high impedance VOM, all I have are cheap digitals.

[pdlethbridge]

They recommended I use a 22ohm resistor for my 12 v bulbs but I went 33 ohm. I figured I could live with a dim bulb

[Yampa Bob]

No, you can't live with me, my wife would object. 

It would help if I better understood the function protection circuitry of the decoder, if there is any.  My original training was in vacuum tubes and basic transistors, now it seems my mind is just a vacuum. I understand FETs and SCRs, but don't ask me to explain them.   

[pdlethbridge]

Dim bulb, I thought you were a dim wit

[Jim Banner]

The function outputs of decoders, including the lighting outputs, are turned on and off by solid state switches.  These are transistors buried inside integrated circuits on the decoder's printed circuit board.  Just like any transistor, there are two concerns when it comes to current - maximum continuous current and maximum peak current. 

Maximum continuous current is the one that decoder manufacturers normally specify.  It is the maximum current that the transistor can pass without internally overheating beyond a safe limit.  It can vary somewhat depending on the air temperature around the decoder and how well the heat escapes from the integrated circuit - adding foam tape that keeps the heat in reduces the current rating; mounting the IC on a metal plate that can carry the heat away increases the current rating.

If you were to pulse the transistor switch on and off, you could handle larger currents during the on time because of the extra cooling during the off time.  But there is a limit to how far you can take this.  At some point, you reach a current that no matter how short its duration, it will cause localized overheating inside the transistor.  One term for this rating is absolute maximum peak current.  Exceed this current by a little bit and a little bit of the transistor melts and becomes useless.  It may not alter the operation of the circuit in any obvious way, at least not the first time.  But that damage is cumulative.  After many repeats, the transistor fails.  And we are left scratching our heads and wondering why.

Using Bob's excellent calculations, the inrush current would indeed be about 500 milliamps.  This would be the absolute maximum peak current for the switching transistors deep inside the decoder.  Whether you need 22 ohms or 47 ohms to limit the inrush current, it all depends on what IC the decoder manufacturer has used.  We are not usually privy to that information but I believe we can trust the decoder manufacturer's recommendations on what resistor to use.  Or do as PD did - use a slightly higher rating.

Now Bob, about that vacuum tube decoder ...

[pdlethbridge]

A z scale vacuum decoder would Just fit in a G scale box car. In O scale, On30 it would be 30". In Z scale, what would the track gauge be for G scale boxcar? Oh, those trucks would be tiny and the boxcar would need training wheels. How many z scale locos would it take to pull a G scale box car? As many as would fit inside.

[Yampa Bob]

Not even the "Mad Scientist" could cram a pair of 50L6 output tubes into an HO tender. 

Jim, if you're still "listening", refresh me on the average current requirements for standard LEDs.  The "expert" at the local store had no idea. Since the recommended resistor is 680 ohms, that comes out to about 18 ma at 12 volts.

Also, there seems to be an abundance of 12 volt lamps available, but limited 1.5 volt types.  Perhaps most modelers just feel safer using the 12 volt, but with the higher power consumption they often require a 1 watt or higher resistor.  Due to space constraints, I would rather use a 1/4 or 1/2 watt.

[Jim Banner]

LEDs will typically handle up to about 50 milliamps but typical white LEDs give excellent results at 10 to 20 milliamps.  Over the years, the required resistance has slowly gone up as LEDs become more and more efficient.  I usually use 1000 ohm resistors in series with headlight LEDs that light up acrylic light pipes or fiber optics and 1500 to 2200 ohms in series with LEDs that form headlights directly.

The resistance required also affects the wattage rating of the resistor and therefore its physical size.  As a rule of thumb, you should use a resistor with about double the wattage rating of the power that it has to dissipate.  This will generally keep it cool enough that it will not melt plastic.  Lets do a few calculations of the wattage required for different LED currents.  Suppose we are using a Bachmann E-Z Command that puts 19 volts peak on the rails (like mine does.)  Then the voltage available for operating a headlight is about 17 volts (2 volts are lost in the bridge rectifier plus the transistor switch.)  A white LED takes about 3 volts to light it up, so that leaves about 14 volts across the resistor.  Suppose we selected a 680 ohm resistor.  Then the current through the resistor (and through the LED) would be 14 volts/680 ohms = 20.5 milliamps.  The power dissipated by the resistor would be 20.5 milliamps x 14 volts = 288 milliwatts.  For that we would have to use at least a 1/2 watt resistor.

Now suppose we decided that the headlight was bright enough with a 1500 ohm resistor.  The current would be 14/1500 =  9.3 milliamps and the power would be 9.3 millamps x 14 volts = 130 milliwatss.  Now we could safely use a 1/4 watt resistor.  If that were still too large physically, we might use a 1/8 watt resistor and epoxy it to the locomotive's metal frame.  Or we might use two 1/8 watt, 750 ohm resistors in series to effectively give us a long, skinny 1/4 watt, 1500 ohm resistor.

Now about those 50L6 tubes.  The filaments alone would need 50 volts x 150 milliamps =  7.5 watts per tube.  We had better make that tender out of titanium.  But wait.  There is also a matter of maybe 100 milliampts of plate current and perhaps 25 milliamps of screen current, both at 100 volts or so.  Let's call it another 12.5 watts per tube.  A total for the pair of 40 watts.  Lets cancel the titanium and make the tender out of pure tungsten.  It's going to light up light a light bulb anyway.  Transistors - ya gotta love em.     

[pdlethbridge]

Just what we need , a glow in the dark tender!