LED Headlight For MTH Railking SW Switcher

Great  work! It sure feels good when you can figure something out on your own.

FWIW, I use a 220 ohm resistor for the LED with PS/2 boards.  You can do down to around 150 safely, as the PWM output is an effective 6VDC.

Sure it feels good, but it would have felt better if I didn't lunch that CV board. I found one on MTH's website and it will set me back $43 (inc. shipping) for my carelessness. I'm not used to handling live circuit boards and letting the board touch the chassis was just plain dumb. As somebody always says, "Good judgement is the result of experience, which is the result of bad judgement."

When the board comes in I'm going to convert the other unit to LED also now that I've got it all figured out. Meanwhile I'm going to order some more varieties of resistors from Jameco Electronics.

I also want to convert my only Railking EMD SW switcher to LED constant lighting. I have an old circuit design drawn by Don Grabski about 15 years ago. For the non-directional lighting it uses a bridge rectifier to take roller current and turn it into constant DC to the LED. The engine uses QSI circuits for the motor control and so DC is present, so I'm wondering if I can tap into that for the lighting control. For the directional version it does tap into the motor wiring. There's only one big incandescent on the low nose, but no lighting in the cab end. It would be nice to illuminate that end also. My 2mm LEDs are pretty small and can be put almost anywhere.

Time has passed. Trip to Philly over New Year's and finally back in Louisville. What greeted me was the package from MTH with my new CV board for the lighting control and all the electronic components from Jameco and LED/Switch with the parts to add LED headlights to my RailKing SW switcher.

I chose not to add an LED headlamp to the rear A unit of the E8 lash up. I may still do it now that I know how to handle those CV boards without blowing them out.

The new board went in without a hitch, especially since I did it with the power OFF this time. I'm thinking about selling the entire set (Rock Island E-8 A-B-A engine with five standard cars, two car add-on and super-dome car) since it's getting very little use these days and I use the funds for other projects of which I have many.

Next up will be the conversion to directional LEDs for the SW which I will post in detail to replace what appears to be candle lights unless it's going way too fast.

I built the LED circuit for the Railking engine today. I started with the simpler end, the calf unit since it basically was a set of pickup rollers and one incandescent lamp holder. The circuit, as drawn by Don Grabski, that I used was the non-directional circuit that takes AC from the pickup rollers, rectifies it to DC, smooths out the pulses with some electrolytic capacitors, and then controls the voltage to a steady 3 volts to drive the LED. Not having a circuit or breadboard, I just wired the components together with lots of shrink tubing.

I made the LED bracket so the LED was at the height of the headlight lens using some brass stock. I used an LED lamp holder to support the LED, soldered the negative lead to the brass and the positive lead to the circuit coming from the voltage regulator. I didn't have a 270 ohm current limiting resistor so I substituted a 330 ohm. I would maybe dull the LED a bit, but it should work.

Here's the original schematic as drawn by Don. All the components listed were purchased at Jameco Electronics. They're well stocked and prices are good.

Here's the wiring all completed. All the grounds are tied together and attached to the chassis at the same place the original ground was; under the screw holding the headlight bracket. LED is on the right hand side of the picture. Rectifier is the 4-lead device at the top. I ran out of small gauge black wire so green wires are ground wires. The rectifier is a 2 amp unit. The voltage regulator is the same as the schematic as are the capacitors.

Now to the problems:

The circuit works. The LED lights correctly and doesn't change with increased track voltage.

But... the circuit breaker on my MTH Z-750 test transformer kept tripping. The bridge rectifier was getting too hot to touch.

I took the chassis to the layout and powered it with my Z-4000 which has significantly more current capacity. The lamp lit and the rectifier got hot and it pulled this many amps at mid-throttle.

Something's wrong. A 20ma LED shouldn't be drawing over 50 watts. The rectifier got very hot very quickly. It's not shorting out since the LED works and nothing was smoking. 

Any ideas on what's going on? Bad rectifier? If the voltage regulator wasn't working the LED would show it I believe. and one more thing. After putting the body in place, the amount of light from the headlight lens wasn't enough to justify all the work. I may remove the plastic lens and put the LED directly in the hole. That could work better.

I think your common from the track and the circuit ground wire are connected together 

They are. I thought all the grounds would be at the same potential. Where do you connect the circuit grounds to?

Right. The circuit ground comes off the bridge rectifier as DC ground (not AC ground from the track). I'm not sure what the difference is, or why it has this effect on current draw at the transformer. Maybe someone else can explain that. 

If you want to add an LED headlight to a PS/2 locomotive, you just use a 220 ohm series resistor and the LED.  You must NOT have anything tied to the frame ground, that will result in bad things happening!

The reason the rectifier gets hot is that with the grounds connected together, the bottom left diode of the rectifier is shorted and the negative half cycle track power is shorted out by the upper left diode...and it gets hot.

Float the circuit (dc) ground.

That circuit seems overly complex to simply drive an LED.  For one lousy 20ma LED from track power, here's a three component solution that gives you constant intensity over a wide range of track voltage values.

That's a lot simpler. What is a CL2N3? Where's the current limiter? Will the circuit work if I separate the DC ground from the track common and tie all the DC grounds together?

Constant current device, better than a limiter. 

Good choice circuit John.

Probably easiest is as you say, cut the connection from the LED "-" going to the brass bracket which is track/chassis ground.  Then run a wire from the LED "-" to the DC ground which appears to be the green wiring.

The CL2N3 is a 2-terminal component (looks like the 78L05) but provides a fixed 20 mA output current without any external components.  While it is a 25 cent component (or so), it can be difficult to find it in small quantity.  You might end up paying $3-4 shipping. 

Trainman2001 posted:

Will the circuit work if I separate the DC ground from the track common and tie all the DC grounds together?

With what you have now, yes. But you can't have your LED ground tied to the chassis.

Trainman2001 posted:

...A 20ma LED...

BTW, the circuit schematic you posted (270 Ohm resistor) will not supply 20 mA to the LED.   Less than 10 mA.  That you're apparently using a 330 Ohm resistor lowers the current even more.  If you're a fan of those LED calculators note that you're starting from 5V (from the 78L05) and the LED voltage will be around 3.3V.  So for 20 mA of LED current you need a resistor just under 100 Ohms.  If you only have 330 Ohm resistors stash, put 4 of them in parallel to "make" a ~80 Ohm resistor and that will get you darn near 20 mA.

Jeez! I love you guys! How could we survive without a forum like this? I will make the mods and look up that current device. Simple is already better. I do have some other resistors to try. Maybe that's why the light wasn't as bright as it could be. BTW: is there any reason why the resistor calc shows the current limiting resistor on the negative side of the LED? Does it make any difference which side it's on in a series circuit?

Grabski's schematic is 17 years old. Perhaps an newer design needs to be offered, and I think Gunner has done that.

I just ordered 20 of the CL2N3-G LED drivers. That's a great solution since it ignores the input voltage as long as it's over five volts up to 90 V input. On the product literature it seems to show that it has to be DC input.  Is this true? What happens when you drive it with AC?

Trainman2001 posted:

What happens when you drive it with AC?

I believe GRJ's circuit includes a diode to yield half-cycle AC which is then smoothed out by the 220mf capacitor to give you pulsed DC. That should work just fine.

Trainman2001 posted:

BTW: is there any reason why the resistor calc shows the current limiting resistor on the negative side of the LED? Does it make any difference which side it's on in a series circuit?

No. The main thing is that the voltage drops equal the supply voltage. The calculation is not hard:

Vs = supply voltage
Vr = voltage drop through resistor
Vf = LED forward voltage
I = LED max current
R = resistance

R = (Vs - Vf) / I
Vr = I * R
Vs = Vf + Vr

Example: Vs = 12 volts, Vf = 3.3 volts, I = .02 amps (20ma)

R = (12 - 3.3) / .02 = 435 ohms
Vr = .02 * 435 = 8.7 volts
Vs = 3.3 + 8.7 = 12 volts

Trainman2001 posted:

I just ordered 20 of the CL2N3-G LED drivers. That's a great solution since it ignores the input voltage as long as it's over five volts up to 90 V input. On the product literature it seems to show that it has to be DC input.  Is this true? What happens when you drive it with AC?

The diode and cap create the filtered DC required for the CL2. 

I actually knew Ohm's law at one point early in my life. While, the calculations are not easily at hand, I still remember that current is constant in a series circuit and varies in a parallel one, and voltage behaves oppositely. Just as a reminder, when putting in a diode, the side with the line is the output side?

Again, this information is terrific!

One more (final?) question. Is the bridge rectifier wrecked since it was operating in a shorted condition? I have more of them if I have to replace it.

Yes, the band on the diode is the cathode, that will be where you tap off the positive voltage. 

Trainman2001 posted:

Is the bridge rectifier wrecked since it was operating in a shorted condition?

Hard to say. You could test it with a meter to see if it is putting out the proper DC voltage. If you've removed the short to AC ground, it shouldn't run hot anymore.

Well, using all the great input and some more trouble shooting, I not only got the Calf's LED to work correctly pulling basically no more amps than the tracks do with nothing running (about .9 amps), I also got the LED to work in the COW using the directional LED lighting scheme also drawn by Grabski. This one takes it's input from the DC running to the motors through a diode, the voltage regulator and the LED with resistor.

I rearranged all the ground wires based on the recommendations and it lit, but then the regulator started smoking. I'm not an electronic maven, but it seems to me that with model trains smoking should be limited to the smoke unit which this simple calf unit does not have. The voltage regulator was getting very hot. 

I had chosen the pin out array by looking it up on the web and thought I had it correctly, but then I noticed that on Grabski's schematic he shows the pin out and I had it reversed. I cut the burned out regulator out of the circuit and replaced it with another, correctly installed, and Voila!, everything works!

Notice that I lost the brass bracket. Instead, I drilled out and removed the plastic lens in the loco and drilled out the opening to accept a 3mm LED from inside. The LED is direction and has a lens in front so I figured it would direct a strong beam... and it does! I cleaned up the wiring using some small cable ties and reinstalled the body after pushing the LED into the opening. No glue was needed. Unfortunately, there are no motors or motor drivers in this dummy unit so I was unable to put a directional circuit here. So the calf's light is on all the time.

And here's the Calf on the layout. The light is brilliant and being warm white matches the color of a real engine's from that era.

For the Cow, I had to determine the polarity of the motor leads when the engine was moving forward (long nose first). I put it on the test track and got the unit moving slowly enough that I could get a volt meter on the motor leads and find out the polarity. After building the circuit I temporarily affixed the leads to the motor terminals and tested it. When it worked, I soldered the leads, used some liquid electrical tape on the one hot lead that was exposed at big capacitor, and then cable-tied it all together. I had already drilled out the headlight housing to accept the LED.

I put it on the track, but unfortunately, this light is directional and only turns on when the engine is in forward. This meant I had to try and photograph it while it was moving. This was the best shot I got...a blurry mess...but it shows a very bright and directional head light. The light does not vary from crawl to full speed.

I've ordered from Arrow Electronics that LED drivers. These seem to be terrific little devices that negate the need for resistors to limit current AND the need to worry about supply voltage as long as it's 5 or more up to 90 volts. LEDs could then easily be driven from the AC accessory outputs from the transformer, or any type of AC converter without worrying about a collection of matching current limiting resistors (although I have a ton of these and just ordered more). The LED drivers were cheap ($0.34 each) so I bought 20 other for a little more than $6.00.

Again... thanks all for all the great input.

Yes! Way to go! It looks great!

Good job, Myles.  Funny how those electronic components don't like to be wired backwards.  You an make the calf light directional by running a small tether between the two engines. whatever you decide, they look great.