LM7805

Maybe use a voltage regulator?

CL2N3-G

I would think that maybe one of the current limiting IC's that are available now would be the way to go.   Here is a page that discusses them.

http://www.trainelectronics.com/LED_Articles_2007/LED_104/index.htm

Cool

Those are slick. Didn't know they existed. Thanks for the tip Kevin.

Credit Where Credit is Due...

They are slick but I didn't discover them on my own.  I learned about them from Dr. Geoff Bunza, in his article "Points of Light" MRH Feb. 2012, and again in the Animated Crane article, "Bring a Wrecking Crane to Life", MRH Aug. 2012.

Trying to understand your reply

Isn't that what the OP wants - same brightness?

Don't DC locos need at least 3-1/2 volts to get moving?

Voltage required

Yes.

Rich

Sorry, guys, if my tone is

Perhaps I should take a hiatus. Been dealing with hospitals, ERs, ICUs, etc. almost continually since 12/22. Most of my "posting" time has been in waiting rooms. Linda's mom (84) had a major stroke complicated with GI issues and pneumonia. In the midst of that, my dad (93) was hospitalized with bronchitis. We actually made three hospitals in one day!

Okay, let's sort this out.

If you want the LEDs to come on while the loco is stopped or just starting on DC, you need to provide some energy storage: capacitor or battery.

White LEDs (http://www.mrdccu.com/curriculum/Lighting/LED.htm) need a minimum voltage (about 3.5 volts) to fire. Below that they will be totally dark. The series resistor will limit the current at higher voltages. At about 4 volts, for example, a 750 ohm resistor will allow about 1/2 mA to flow. Commonly available LEDs today have a rated current of 10 mA. They will light dimly at 1/2 mA. Each additional volt above this point will add 3/4 mA. By 6 volts, there will be about 2 mA flowing and the LED will be about 80% brilliance. At 12 volts, that total current will be about 7 mA - VERY BRIGHT!

Why I disagreed with prior posts, is that the voltage regulator or current limiter will only restrict maximum brilliance, which can be done with a resistor. Some locos will move as low as 2 volts.

Hope this helps.

FWIW, the end of my last post got lost in the ether, sorry!

That explains what I was picking up on...

Ahh, I wondered if it was something like that.  It just seemed like you were on edge, not yourself.  Now I understand why.  Best wishes for both parents feeling better soon!

I think I understand what you are saying.  And I agree with a point you made earlier, the current limiting IC really does pass more current than most LED's are wanting.

But, I thought the current limiting IC might be a possible solution for the OP's desire to have constant brightness because it would pass the current that is being delivered at 2 volts, (which may be too few Volts to fire the LED) to a maximum of 20mA, Thus I thought, giving a more consistent brightness over the range of voltage being supplied.

Thanks as always for your input in matters electronic.

Best regards,

The current limiting ICs will

Now that's dedication!

Oh man...

You're too busy to do experiments right now!  I've got time, some CL2 chips, LED's and a DC pack. I'll see if I can run the test before bed or tomorrow for sure!

Thanks Bruce!

Thanks!

Thanks very much folks, you've given me a lot to check out, test and read. 

Just a couple of comments... yes, with a can motor and a quality gearbox a steam engine will move at about 2 volts, unless you have added the diode bridge and 1.5 v lamp for the headlight.  In that case, the first 2 volts go to the bridge and bulb and then the loco will move.  What that looks like on the layout, is that as you crack the throttle, the headlight comes on (at a brightness that doesn't change over the entire speed range) and then the engine begins to move.  It is possible to adjust the throttle to have the loco just sit there with the headlight on and not move at all.  I've found that useful when taking photos, but otherwise it's just a stunt IMHO.  If the LED/resistor combination wired in parallel with the motor will achieve 80% brightness at 2 volts and then pick up the other 20% over the rest of the throttle range, it seems like that would be more than acceptable and not really noticeable during normal operations..  Just changing the resistor size would really then set the apparent brightness baseline (the 80%).  780Ω or higher depending on how bright you want the headlight.

I still need to read a bit more but thanks again!

Dave Vinci

O=='=::

Dave -You can clip the LED

LM334Z

How about this device?

http://www.pollensoftware.com/railroad/

Rich

LM334Z?

That's a neat chip, though it's limit of 10mA makes it seem like it would not suit applications where full illumination was desired very well, would it?

It also takes two components to accomplish what the CL2N3 does with one, though the CL2 is not "programmable".

As for price, checking at Mouser, I see the LM334Z is nearly twice the price of the CL2, but to be fair. they are both less than a  dollar a chip.

Still, as I say, it's a neat chip.

Current

I thought of the chip because in all my experiences, a 1k resistor gave me a current of about 9 ma at 12.5 volts.

I have been reading about LED's in various DCC forums and some are going to at least 2 k because of the super bright LED's that are showing up.

Rich