Stored Energy Is the Key – How You Use It Makes the Difference

If you have any kind of light in a rail car, LED or incandescent, and the car runs over an unpowered frog or dirty track or a big gap in a rail, the light inside the car may well flicker off then back on. It is incredibly unrealistic. LEDs will turn on or off faster than an incandescent lamp so they will be more noticeable in their reaction to gaps. Please note: this is not about keep-alive for DCC decoders. This is only about car lighting.  While the power is temporarily removed, the LEDs/lamps must get power from somewhere to prevent flickering out. Batteries and capacitors are the typical electrical storage devices. Besides battery power, there are at least four prior approaches to flicker-free car lighting: using a resistor and large capacitor with or without a voltage regulator, and using a super capacitor circuit with or without voltage regulator for lighting. All typically use a small diode bridge connected to the rails to provide a DC voltage. My friend wanted a dead simple, low cost solution. This eliminated the battery and supercap alternatives for both complexity and cost. Non-rechargeable battery solutions were ruled out for long term maintenance and replenishment cost.

Basic Flicker Free Diagram

The problem with the basic circuit focused on the amount of energy you could reasonably store in a capacitor that could fit in a rail car. Bigger was always better, and would even work reasonably well for a single lamp that did not draw much power. The problem was powering passenger cars that might have 3 to 8 LEDs in them, or a caboose that might have 2 LED marker lamps and 2-3 lamps inside. Even limiting the current draw per device, the total current might be above 30 ma. That kind of power draw would have a noticeable flicker for a mere 2200uf 25V capacitor, and larger capacitors or groups of capacitors consumed much space inside the models, or just didn’t fit at all. The use of a regulator allowed a drop in voltage in the keep-alive capacitor, but was itself a load on the keep-alive, ultimately limiting the capacity of this approach. But it was a good idea nonetheless,

A Different Technology Changes the Game

There is another, newer kind of “voltage regulator” that takes a very different approach – a “switching regulator” that actually acts as a DC to DC converter. “Big deal” you say? Well… functionally no, but it can operate with near 96% efficiency, and in this application it does make a difference. Years ago switching regulators were complex and expensive. The advent of small integrated circuits that accomplish nearly all functions changed the game. You can find quite a few on the market. The “Battery Powered Models in HO Scale” article in the November 2014 issue of MRH   http://mrhpub.com/2014-11-nov/land/#83  used two examples.

A new, much smaller (physically) version came to my attention:

Small Cheap Completely Contained DC-DC converter/Voltage Regulator

It can handle up to 28 Volts input (from the track) and has an adjustable output from 0.8 to 20 Volts. In this application we will not be stressing the converter. You can see from the picture it is quite small. The largest component we will use will be a capacitor rated at 25 Volts. If you are running G gauge DCC then you should use a 35 Volt capacitor (but you will have ample room for the slightly larger cap). A small diode bridge less than 1 amp will do well too. I had a 1000uf 25 Volt capacitor and a 2200 35 Volt capacitor on hand for trials. That’s it—3 components! Here they are:

1/2A 400V MINI DIP BRIDGE RECTIFIER  allelectromics.com CAT# FWB-16    .40

3A DC-DC Converter Adjustable Step down Power Supply Module                  .96

http://www.ebay.com/itm/291353891841

CAP ALUM 1000UF 25V 20% RADIAL      Digikey  493-1305-ND                          .51

or

CAP ALUM 2200UF 25V 20% RADIAL      Digikey P5157-ND                                .90

Putting these together is simplicity itself. The capacitor is soldered across the input to the converter – remember to connect plus to plus and minus to minus. Observe correct polarity for the tiny bridge rectifier too. It’s plus and minus connections go to the module input too. Before you connect the completed module to your car lighting, you must measure the output voltage and adjust it to as low as possible for the type of lighting you have. Different color LEDs have different voltage requirements. For my white LEDs I adjusted the output to about 2.8 volts, and used a 47 ohm resistor in series with a group of 4-6 white LEDs.  You can find the tiny trim potentiometer (circular device) just below the largest black component on the module in the picture. Also an advisory to those who do use either module-- the little trim potentiometer is easy to turn, but has NO stops at either end. It would be a good idea to make all adjustments before you connect it to your model lighting. The module itself is tiny, The total volume will be dominated by whatever capacitor you use.

Parts & 2 Complete Units – 1000uf and 2200uf Caps

The “Small” 1000uf Keep Alive

There is an even smaller module available, which should also do exactly the same job: You can find the tiny one I used here: http://www.ebay.com/itm/291420624610

The “Tiny” DC-DC Converter

The above picture shows the relative component sizes.

Why Does This Work?

Track voltage is the largest energy source available for car lighting. The capacitor we added is charged to whatever is on the track after DC conversion, whether it is DCC, DC, or AC. The converter lowers the voltage to the minimum workable with incredibly high efficiency (87-96%) so most of the power stored in the capacitor will be used. We set the voltage out of the converter to be the lowest usable (or nearly so). As the capacitor is discharged to light the car, the converter will track the declining voltage and still put out (again with high efficiency) the correct lighting voltage. So the lights remain powered with no flicker. With the 2200uf cap at DCC voltage on my track I have measured about a 3 second keep alive for 35 ma lighting. With a 1000uf cap I get about a 1 second keep alive – all these are approximate.

In the following video you can get an idea of what this can do.

If you are running DC power, you will likely not have 12 volts on your track, 6-9 volts might be more likely. For whatever you set the output voltage of the module, it needs about 3 volts more to be effective. So if we set the output for 2.8 Volts it will need about 5.8 Volts on the track to be effective. The larger the voltage, the longer the keep alive will function. A larger cap for DC operation will also help in spite of the lower voltage margin. The same goes for variable voltage AC operation.

I hope this helps you light up your own cars along your right of way!

Have fun! 

Best Regards,

Geoff Bunza

Nice Application

Nice use of the new micro sized switching regulator. I had seen these on one of the electronics distributors web site earlier. Never connected the dots to realize it could be used this way.

My question is how is this simpler or more compact than the traditional bridge/cap/regulator circuit? I trust you when you say it is but in the picture it doesn't look very much smaller and, assuming one can etch their own boards, only marginally simpler.

Might another approach be to use 2mA LEDs in the first place? Many of the model train LED applications seem too bright and would benefit from a lower brightness LED anyway.

@Alan

Hi Alan,

I don't contend at all that this is smaller than just a cap and a resistor-- it's not. But this will provide more current longer than other approaches. That's it. I freely admit that I like to run LEDs bright inside cars and have a number of them. This will give me some advantage. Your mileage may vary! 

Have fun!

Best regards,

Geoff Bunza

slow soak

Geoff,

I'm being a bit slow.  Your picture shows a circuit board, which I suppose is the module and the capacitor(s) directly attach to it. Is that right?  Also, I see a DIP8 pin bridge and don't see where it fits into the picture.

I've a number of passenger cars I want to 'light up' and this looks like a great approach. Like you, I like the higher intensity led lights.

Thanks again for your work and your willingness to share with us.

Components

Hi Joe,

Both the capacitor and the diode bridge attach to the same place-- the plus and minus +/- input side of the little board (module). Make sure you have the correct polarity for all. The little bridge DIP package is usually marked with a small * and - too. The connection to the wheels/rails go directly to the AC side of the bridge/DIP which are the reamining 2 pins. If you are running with DCC these work incredibly well.

I just got my "tiny" boards in the mail. They are unbelievably small, and I'll post a side by side picture of them together later today/tonight.

Have fun.

Best Regards,

Geoff

Actual measurements would be

Actual measurements would be nice, too... interested in the possibility of putting these in an N Scale caboose!

Actual measurements

Hi,

This should give you a good idea:

You should be able to click on it for a high res picture too.

Looks like the same circuit with fewer connection holes. Like I said, it's tiny. You can find the one I used here: http://www.ebay.com/itm/291420624610

Also an advisory to those who do use either module-- the little trim potentiometer is easy to turn, but has NO stops at either end. It would be a good idea to make all adjustments before you connect it to your model lighting.

Have fun.

Best regards,

Geoff

Another picture

Geoff,

I understand where everything attaches, but maybe another picture or two would make things perfectly clear for those coming into this at a latter date.  If you did a shot 180 degrees to the picture captioned "Parts & 2 Complete Units – 1000uf and 2200uf Caps" then it would be obvious where the bridge attaches as it is hidden mostly by the capacitor.  Also if you were to take a picture with a completed unit flipped over to see the components of the DC-DC convertor that might make things clearer as well.  When I first looked at the pictures I wondered what you did with the DC-DC convertor then it dawned on me you just mounted the cap and bridge to the non-component side of the convertor.

And I'm stumped what does KAOS stand for?  I figure its Keep Alive something.

@Jim re: Component Placement

Hi Jim,

Good enough... here is another shot:

These capacitors are Surface mount devices. However if you wanted, you could simply bend the two pins up and slip them out of their plastic frame very easily. I chose not to do that. Instead, I "tin" the leads in place with a very thin coat of solder, then "tin" the pads on the non-component side, hold the 2 together, and apply heat to flow the solder together. It makes for a very compact assembly, as you can see in the photos. You do not need to use SMD caps at all, in fact you can separate the cap by location if you are trying to hide it in another part of the car.  In using the other "tiny" module perhaps for N scale, separating the cap by wires may be desirable in some instances. Both sizes of caps that I used can be soldered to the larger board as shown.

The diode bridge was soldered to the top (component) side of the board/module. I did not place the leads into the holes, just soldered them onto the top pads, as shown. Both the cap and the diode bridge connect to the same points electrically.

These actually work remarkably well, and will perform better than simpler resistor/cap combinations, and at a radically lower cost point than every supercap circuit that I know of.  If this needs more clarification, please ask. I hope this makes it more clear. If not, please ask again.

Oh, and KAOS = = Keep Alive On Steroids, harkens back to both my MRH article on Battery Powered Models in Nov 2014 MRH, and to the old Get Smart TV series that I enjoyed in a previous lifetime (between building models of course).

Have fun!

Best regards,

Geoff

Thanks

What a quick response! I see from the pictures and your additional description exactly where the additional components are installed.

I'll have to go check out your article I don't read the magazine cover to cover so must have missed that article.

And now I'll never forget what KAOS stands for, well at least in this instance.

Hi Geoff Great idea for

Hi Geoff

Great idea for flicker free car lighting.  Thanks for posting.

Couple of Questions:

I have a passenger train with four passenger cars (small layout) and would like to add car lighting, using 3 to 4 SMDs in each car.  Would you suggest installing the flicker free lighting circuit in each car or could one circuit be used for all four cars?  The later would require connecting wires between cars, but I do not plan to uncouple the cars and always run them as a four car consist.

I would like to control (on/off) the lights with a function decoder (like the TCS 1002 FL2 decoder).  I assume the decoder would go between the rail power and the bridge rectifier.  Is that correct?  This would also require connecting wires between cars and one flicker free circuit.  Do you see any issues with this arrangement?

Cheers

Rick

@Rick re: Questions

Hi Rick,

No, if I were doing the equivalent, I would add it one per car for a couple of reasons: 1. The interconnect from car to car (I assume you would not permanently connect them) is about as complicated to install & maintain as simply putting them in each car -- it's the hassle factor more than anything here for me and 2. The current demand for all the cars would obviously be effectively higher, requiring a physically bigger capacitor and would likely provide less keep alive time for the lights and 3. Personally I have a bias to like to be able to separate/switch  the cars.

Great question! First you would use the function at full 12 Volts/Full DCC power (direct connection to the decoder Blue wire). Depending on the decoder you are using the function could put out somewhere between 200 and 500 ma max. I don't think this would be a problem for either the decoder or the converter, but good design considerations would say to put a current limiting resistor -- something like a 22-56 ohm small resistor in series with the blue wire to the converter (Blue decoder wire to resistor, other end of resistor to plus (+) side of converter board. Also, connect the decoder function wire directly to the minus (-) of the converter board. Yes this does mean that you don't need the small bridge rectifier at all in this configuration. The capacitor still gets attached across the +  & - inputs to the converter board, as before. Realize too, that if you turn off the function controlling the lights, they will likely stay on for a second or two depending on how many you use and the total current draw.

Because of current limitations in the decoder function, I would not recommend powering the set of cars' lighting with one decoder function. Remember in this configuration all the current is coming from the decoder function line, not directly from the track.

If you had your heart set on using one decoder to control all the lights in all cars, you might consider using the decoder function to turn on and off a small relay, controlling the track power to the bridge in the original circuit I showed, and then running two wires carrying the switched track power to the rest of the cars, each car equipped with their own keep alive circuit. This would get around the likely, current limitation of the decoder.

'Hope this is of some help. Let us know how you progress. Have Fun! 

Best regards,

Geoff

Hey Geoff

You sound like your talking underwater.

Bernd

@Bernd

Hi Bernd,

I have no idea what happened to the audio! I converted the video for upload and some gremlin got in there!

I think I'm going to have to re-do it... UGH! ! !

Best regards,

Geoff

Led wiring

Geoff,

I have already ordered some of the small DC-DC converter and am planning to use the circuit in HO passenger cars.  However I am a little confused on how to wire the LEDs.  In your video, there appears to be no resistors on the LEDs, but in the text you indicated that you used a 47 ohm resistor in series with 4-6 LEDs.  My original plan was to use a bridge rectifier, a resistor to reduce the light intensity and then a 12v strip of LEDs ( most likely 4 sections or 12 LEDs.  Would this be too much current for the circuit?  Would setting the Converter to 12 volts or somewhat less reduce the ON time?  Would a better solution be to use individual LEDs, say 6 to 8 3528 surface mounts and if so what would be the proper resistor?  Would they be connected in series or parallel?  Does the converter limit the current or does it have a current max ? 

As you can see, I'm not sure I understand how the circuit works and would appreciate you input to help me use the circuit which seems ideal for my application.

Bill Feairheller

@Bill re:Led wiring

Hi Bill,

Let me reply to your questions:

No, the little converters are rated at 3 Amps, although I would likely never recommend using them at current levels over 1.5 Amps without a little heat sink attached. But both my and your applications do not approach these current levels.

Yes, it would reduce the keep-alive time dramatically. In fact, if you want to power LED strips (I use these too by the way) with 12 Volts, I wouldn't bother with the converter board as it would yield a very marginal, or no advantage. Just the bridge and a cap would likely do nearly as well.

Short answer, yes, but bear with me through a longer answer: This approach works because it takes what is presumably the larger track voltage (and power), stores it up and then with extremely high efficiency converts the stored power to the lower power used for the LED lighting. If the "output" power requirement is nearly the same as that stored in the capacitor (like for lighting a 12 Volt LED strip with a bunch of LEDs) it will still "work" but the keep-alive duration will be short. If you reduced the power consumption of the LED strip with an added resistor in series with the entire strip, it would help extend the keep-alive time. This is an issue of power (in this case Volts X Amps), the bigger the power difference between what's stored in the capacitor and what is drawn by the LEDs, the longer the keep-alive time will correspondingly be.

I have found that taking whatever LEDs you want to use (in your case the 6 to 8 3528 surface mounts) putting them in parallel, and adjusting the converter voltage quite low (roughly 2.8 Volts for white LEDs) with one small 10-56 ohm resistor in series to the group of LEDs does a good job. I give a range for the resistor because people differ about how bright they like to light their models. I like brightly lit models but others like much dimmer lighting -- that's fine. Even I find the 3528 LEDs too bright for most anything but headlights (and I do enjoy bright headlights!). They are great LEDs to use and relatively easy to solder. I have used them from time to time in passenger cars with good results, but I have to admit that 0603 LEDs would probably work well for most people too (and they are easier to hide!).

The converter I used has a max specification of 3 Amps -way more than what is needed for LED lighting in a single car. Even if it has some current regulation, if you tried to draw that much current out of the module continuously I think you would melt it down before long! I don't think this is of concern here.

I'm hoping this answers your questions and makes things a bit clearer. Please ask again if I haven't succeeded. A little experimentation before final installation in your models will help a great deal.

Have fun! 

Best regards,

Geoff Bunza

Thanks Geoff - another question

Geoff,

Thank you for your quick and informative response.  That helps a lot.  However, I don't understand what is limiting the current into the LEDs.   Usually when we use a circuit with a LED, we use a series resistor (as they are current driven devices)  that is based on the voltages of the source, firing voltage of the LED and the current through the device.  In this circuit with the DC-DC converter there is no series resistor.  What is keeping the current to the LED at a safe level?  Also the video and your response shows the LEDs connected in parallel without a resistor for each which for most applications is a no-no.  How is the parallel wiring of the LEDs without the separate resistors OK in this case?

Thanks again for taking the time and effort to help us understand things that are new to us.  ( If you have a question about Physical Chemistry, I may be able to answer that!)

Bill Feairheller   Modeling the Reading Company in the 1950's

@Bill re:LED resistor

Hi Bill,

Actually there is a resistor in all the cases, including the video. They are either so small (I was using 1/8 watt resistors) or in one case I was holding three resistors tied together in my fingers. For example, if you look very closely at the board with the 6 LEDs, look in the lower left corner of the board and you will barely make out the one limiting resistor for all the 6 LEDs. I was using about a 47 ohm resistor.

It is the combination of a lower than usual (in my case 2.8 Volt) voltage setting from the converter (which also regulates a stable output voltage) and the limiting resistor which will determine the current through the LEDs and hence the brightness. It is possible but not generally recommended to set the regulated voltage to the LEDs at precisely the right setting to have a corresponding current through the LEDs alone. However, the converter does not respond quickly to quick changes at its input and the resistor is also in part protecting the LEDs, besides giving you finer control to set the brightness you would like. Lowering the voltage from the converter also limits the amount of power dissipated in the resistor, which means more power is available for the keep alive to the LEDs.

I didn't dwell on the LED resistor because I assumed people already experienced the need for such a resistor in the general case. I make the mistake (more than once) in my blogs that I can be more general and less precise than when I pen an article. Entirely appropriate questions like yours remind me I need to slow down constructing these entries and pay more attention. In this, I am in your debt ! Thanks.

'Hopefully, I am converging on making this more clear. If not, please come back with more questions.

Bill, when I start messing with battery chemistry, I'll keep you in mind! 

Have fun !

Best regards,

Geoff

Thanks again Geoff - I'm anxious to try it

Geoff.

Thanks again for all the information.  I'm beginning to understand and am anxious to get the parts and try the circuit.  The DC-DC convertors are on a slow boat from China,  so along about some time next month I'll be able to give it a go.  Your animation is great and I really appreciate all the effort you put into the designs and the write ups.  Please keep it up and don't give up on us that are slow to understand. 

I did some teaching in a local junior college and realized that things that were obvious to me,  were not always to the students.  So I understand what you are saying.

Thanks again!!!

Bill

Do you or would you sell

Hi Geoff,

  I am very interested in many of your ideas especially "The Seaplane Scene" and "Harry's TV" don't think I have the ability to do on my own. Would you consider building them for me. I am building a HO layout and they would make it very special. You can reach me at 631-903-0850

                                     Thank You very much, John Pitti

Step-Up Low Voltage Lighting Circuit

I like your neat new circuit and will use it on caboose markers and passenger car lighting on my DCC system.

I have a friend who is running a G train slowly around the Christmas tree.  The operating voltage at this more reasonable speed is generally around 3 volts (DC not DCC).  This would not be high enough to operate your neat flicker-free circuit.  It seems to be designed more for DCC systems.

Do you have any lighting circuits that could draw low-voltage track power and step-up or raise the voltage enough to then operate your flicker-free circuit?

@rdj6737 re: 3 Volt DC Operation

Hi rdj6737  (name?),

The DCC version of this works because it takes and stores power at a high level and delivers it at a lower power level (volts and current) to the LEDs. You could use a DC-DC step up converter to raise the voltage from a low DC source (3 Volts) and store it at a high level in much the same manner. Such converters will likely trade off current in the conversion, so "powering up" the capacitors would likely take more time at startup and might display noticeable "dips" if repeatedly crossing dirty track or power gaps. Nonetheless it will be better than using track power directly.

A DC-DC step up converter like:  http://www.ebay.com/itm/381374576465 
looks like it could do the job. You would need to place this just after the bridge rectifier in the diagram.
'Hope this helps. Have fun! 

Best regards,

Geoff Bunza

Re: 3 Volt DC Operation

Geoff,

Thank you for your reply with the eBay listing for the Step-up Convertor. My friend, Brian, and I will try this step up convertor for our Christmas train G passenger cars.

Reynold De Jager

Sorry for missing my name earlier, and for being so late with my thanks.  I got distracted and forgot about it until I re-searched for your great electronic articles.  It was the animated cranes on YouTube that lead me here first and now again!

Capacitors and me

Hello Geoff

I have soldered and breadboarded the parts. I have them set to input 12V and output 4V and use white LEDs so they should get 3..3.2V. Using several LEDs and a 100 Ohm resistor Or 1 LED and a 68 Ohm resistor as soon as i remove power the LEDs go off... whether i use a 1000 or a 2200 cap.

Are my values out of range or any other problems that stand out from a textual description?

Thanks for providing these great projects and helping us along.

Befuddled, ILH