Add-on LED passenger car strip lighting in conventional mode

Carl - thanks for the info.  The LM2596 is just a DC powersupply, right?  But its not 'buck-boost'?  Presumably using one of the oft-discussed buck-boost modules along with a bridge rectifier would produce nearly constant lighting across most of the input AC voltage range (right?).  I also run conventional for the most part and want to add LED lights to various cars, etc., but I like the idea of having near-constant light output - hence my questions...

The lights appear to go on and off around 8.5 volts which is expected if all of the LEDs are in the circuit as sets of three are wired in series and it takes about 2.8 volts across each one to turn them on.

For a Post War Pulmore motor 8.5 volts may just be medium speed but for a Williams or some other DC motored engines 8.5 volts is about 60 MPH.

By reducing your power supply voltage (buck boost or adjustable regulator) to 3 volts and then shorting out 2 of the 3 LEDs the remaining lights will be at full brightness at around 3 volts AC track voltage.

Pics at home, film at 11.

 

Pete

The main reason I shot the video with the VOM meter and variable voltage was to show the change in light output from the LED's.  To my eye, full intensity is reached between 9-10 VAC. 

 

Per gunrunnerjohn's suggestion, I ordered today the DC-DC buck boost module, as well as the rectifiers and capacitors he identified in the other, currently running LED thread.

 

Full disclosure:  I am no more competent to give advice on LED lighting than I am in giving advice on baking a cake.  However, I can read the directions on the Betty Crocker cake mix box and make an edible cake.  Whatever info you see from me is merely a repetition, accurate I trust, in what gunrunnerjohn has posted.

 

I'll be waiting for the newly ordered components to arrive and then do an installation in some more of the repainted LIONEL 15" cars I have.  I'll also do another video, hopefully a side-by-side, comparison between the AC-DC LM2596 and DC-DC boost buck module installations. 

Norton, thanks for the information. 

Carl - thanks.  I wasn't careful in choosing my descriptors for the LM2596 - its basically a step-down (buck) module that can also convert AC to DC.  The specs on the eBay website say that the input voltage needs to be 2 v higher than whatever the desired output voltage is.

 

Glad to hear that you will be doing a side-by-side comparison with a buck-boost module - I look forward to the video.  BTW, from probably the same OGR thread that you noted, I gather that there isn't a ready-made converter plus buck-boost module available, so that one has to supply a bridge rectifier and maybe a capacitor to smooth out any voltage loss due to dirty track, etc.  Does the LM2596 provide any smoothing on the output side?

I want to switch my passenger cars over to LED lighting, but I'm taking a slightly different approach. I purchased a couple of variable DC LM317 power supplies which will take AC input and can be set to a desired DC output. The plan is to put a power supply in a baggage car and wire the other cars in the train to that car via a through buss. I might even put in batteries. The idea is that only the baggage car will have pickups and this will reduce the problem of "block bridging", plus if I 2-rail the cars it makes changing them over easier since only one car would need modification. The power supply can be found here. I already bought a reel of 12-volt LEDs so I'm almost ready to go. Just have to wait for the parts and get a little block of time.

 

richs09, the thread containing gunrunnerjohn's information and the LM2596 module I used is linked in my previous reply, and, as you mention, there is not to grj's knowledge an AC-DC buck boost module available; hence the addition of the bridge rectifier and capacitor to the DC-DC buck boost module previously mentioned.

 

The LM2596 I used has some capacitance built-in; how much I haven't studied the components to say.  The only thing I did to the modules I used was to add the 22uh choke for DCS operation.

Ok, here is car that uses an LM317 adjusted to 3 volts. In order for this LED strip to light, 2 of the 3 LEDs had to be shorted out. Its a bit tedious but I think for conventional operators this ensures the lights are lit at the first crack of the transformer. A compromise might be to set the regulator at 6 volts and then just short out 1 of the 3 LEDs.

 

 

For comparison here is a car with its full set of LEDs lit. I am using an LM78L09 fixed voltage regulator as I found driving these strips with a full 12V too bright. This car has similar properties to Carl's and doesn't come on till about 8.5 ACV on the track.

 

 

Pete

I'll be waiting and watching for updates.  I hope this year to start converting all my locomotives, cabooses, and passenger cars to LED lighting.  I run mostly conventional and sometimes TMCC/Legacy, so I want them to work nice conventionally.  And since I do have some locomotives that will creep at 4-5 volts, I'd like the lights on at that setting.

Matt--looking forward to your reading and seeing your project.  Especially like the idea of eliminating center-rail pick-ups in anticipation of running 2RS.

 

Norton--thanks for the photos and explanation. 

 

Sinclair--If Norton's post above doesn't do what you want; start a thread and ask.

You can also use a couple of diodes in the pickup leads to eliminate the issue of bridging and still have the lighting.  Fooling around with tethers for all the cars doesn't do much for me.

 

I run conventional. I use a 7805 circuit and individual LEDs shown

 

here.

 

It gives steady intensity regardless of throttle setting with power on. Difficult to do with the peel and stick ones,which require 12 volts input.

 

I add 3-5 farads of capacitance on the regulated side. This keeps the cars lighted after power is removed.

 

Here are pics of the cars.

 

Power on, room lights on

 

 

 Power on,room lights off

 

 

Power off 1 minute

 

 

 

 

Power off 2 minutes

 

 

They continue to fade but are still dimly lighted even after 10 minutes,gradually fading.

 

Dale

The uneven lighting is why I like the LED strips.

 

Dale's method is the most efficient. I think a few more LEDs would even up the lighting. I was going to make my own using circuit board and discrete LEDs but its hard to beat the price of the strips.

Now if someone were to offer strips with one resistor per LED it would be easy to build something that lit from 3-5 volts up to 18-22 volts. It will be interesting to see what the new GGD cars come with.

 

Pete

But by using more LEDs, couldn't you even the lighting out?

The issue with doing them all in parallel is you lose a lot of the advantage of low current consumption.  Let's say you use seven LED's and you are running them not at the 20ma max, but say 10ma.  That's 70ma of current, vs. around 20 for the series/parallel of the LED strips.  Also, with track power on the input, you are dissipating a LOT more power in the LM7005 as it has to drop the voltage to 5V from up to 20+ rectified and filtered track voltage.

 

Let's take an example.

 

Say 25 volts rectified power, that equates to 18 volts AC on the track.  You're dropping that to 5V, so you're dissipating 1.4 watts in the regulator.  In addition, you are also dissipating a bit more in the series resistors.  You'll need a heatsink at those levels, that regulator will be as hot as a firecracker!

 

Contract that with the strips I use, running at 20ma for excellent and even illumination.  The same 25 volts rectified power being dropped to around 10 volts at 20ma.  I get great lighting, I'm dissipating 0.3 watts in the regulator, and instead of 70ma/car, I'm using 20ma/car.  What's not to like?

 

It's also a lot more manual labor to diddle with the individual LED's vs. the strips.

 

In summary, I'm losing some of the low current advantage of the strips, I get less even lighting, and I have to worry about my regulator getting hot enough to melt plastic.

 

For conventional operation, if you swing that way, I'd use one of the buck/boost supplies and the LED strips, much more efficient power conversion.

 

John, In the photo I posted above of the 9 volt car (LM78L09 9 volt regulator), it draws a total of 20 ma. I think its still a bit too bright so I added a 51 ohm resistor which drops current draw down to about 9 ma.

I appreciate the value of constant current. I use that technique to balance the light output of 16 high output (700ma) LEDs used create chemical reactions. The question is, at what track voltage does your LEDs light without the use of buck boost or voltage doublers?

 

Pete

Pete, I designed my module for command operation.  However, since the buck/boost modules are available very inexpensively, and offer far better efficiency, I can't see doing it with the linear regulators.

 

It may photographs that way but the lighting is NOT uneven. In the photo the space between the windows gives that impression as well as the varying transparency of the old strips. 5 or 6 LEDs per car, depending on length spreads the light very evenly. The LEDs themselves have a 140 degree view angle. They are about 1/4 inch square or about 1 scale foot. These disperses light better than incandescent bulbs as no reflector is needed.  Mounted on the roof and shining down here are no shadows.

 

As for labor,it takes me about 10 minutes per strip to make up and hard wire. Getting the cars apart takes more time for me,especially the aluminum ones where the floor slides out. Sometmes they stick.  The individual LEds have come down in price,I think they are about 10 cents each or less with free shipping from China.

 

As far as the reel LEDs,for conventional operation,there is no easy way around making them constant intensity. They are wired in series of 3 for 12 volts input. The first rung of a transformer is 5 volts. You could use a voltage doubler but that would give almost 50 volts DC at maximum throttle. I use 35 volt capacitors.

 

I do not use a heat sink on the regulator and have had no problems. 5 LEDs would use 100ma and at 18 volts in, that would need to shed  a maximum of 1.3 watts. Normally I run the train at 14 volts so the regulator would need to shed less than 1 watt. I have bench tested the cars at 20 volts in from a Z4000 for hours with no ill effects.  I use the regulator back as a connection with the bridge and cap soldered to it and sometimes solder on an extra solid wire to spread the heat. As long as the regulator is in the open there is no problem. On cars with seating it extends into the washroom and does not show. The lady passengers appreciate the heat from the regulator when the use the washroom and apply their makeup.  These LEDs have the same intensity as original light bulbs. The circuit would use much less if I ran the cars dimmer and pulled 20ma per car but I am a toy train guy.

 

I use button type 1 Farad supercaps on the regulated end. All this mounts on the strip between the LEDs and sticks to the roof. Nothing would be visible unless it ran on a shelf layout.

 

Here are some better pics of another car, believe me, the light is very even when viewing. 6 LEDs will do even a 21 inch car evenly

 

 

 

Dale H

Suggestion for powering LEDs for overhead lighting.

 

Use a simple 5 volt zener diode and a resistor to regulate the voltage from the track. Then add a nine volt battery for a "keep alive" voltage along with a series diode. The battery is only needed when the track voltage is removed (or temporarily interrupted). Also add a diode in series to protect the zener from a reverse voltage. An on/off switch may be used to disconnect the battery when the car is not in use. The battery and on/off switch could be placed in a baggage car and wired through to each passenger car. Miniature "disconnects" are available from Miniatronics and other suppliers.

 

Note: Each LED requires a separate series connected current limiting resistor if not already provided with the LED.

 

 

I know you guys (Dale and John) have been going back and forth on this for a while, but I'm glad for it.  It's wonderful for guys like me with very little understanding on electricity to see the pros and cons for both setups.  Thanks!

 

But I have one question.  How did Lionel light the LEDs in the VL BB?  My understanding is all the lights are LEDs, and it is nice and bright in conventional as well as command.  Has anyone looked?

Looks like just a constant voltage headlamp,etc.

 

DYI is here

 

Might need some small ones to fit in the markers. Theres is driven off provisions made in the boards most likely.

 

John and I defiantly have our own way of doing things. His plug and play circuit is great for command users who don't want to solder. I like DYI and have tried many methods on an experimental basis. I settled on the 7805 because I can leave the lights on when power is removed for a while. The fading is a nice effect. If you don't want that I recommend the CL2-n3 chip. Whatever method works for you is the right method. 

 

Dale H

Sinclair, Like Dale says John's circuit is fine for command control but doesn't lend it itself for conventional operators unless you like to run really fast. Constant voltage circuits require more current to drive a circuit while constant current requires a higher applied voltage to power the same circuit. Conventional operators work at lower voltages than command operators.

You can still use a series resistor with a constant voltage circuit to reduce brightness and current draw. Depending on which LED you use you can likely lower the brightness and still have it as bright or brighter than incandescents. The LEDs used in these strips seem to be very bright at rated voltage and current so there is room to adjust down. Discrete LEDs have be selected based on lumen output. Most of the older ones are not as bright. Places like Digikey will have the actual specification for light output. Many eBay vendors do not offer these specs so it can be a gamble.

 

Pete

here are the mentioned LEDs with the specs.35,000 mcd 20ma for warm white

 

link

 

My old ones I did a while back were 20k so they keep improving them. They offer them in larger quantities with free shipping 10 cents each about.

 

Dale

Like I said, for the strips in conventional mode, you can use buck/boost power supplies with a bridge rectifier and optional extra filter cap.  Since they already have a filter cap, most applications don't require anything but the bridge rectifier.

 

10pcs DC-DC Step Down/up Buck Boost Power Supply Module, $23.88 Shipped

 

These will give you full set intensity from the get-go.  If you want additional hold-up time, just add more capacitance to the output.

 

 

This is way more complicated, not to mention a real power hog!  You need to be dropping all the voltage from the track to the 5V of the Zener somewhere, that turns into heat.  A Zener diode is a very poor way to power LED's from track voltage.

Buyer beware on Dale's source.  They shorted me 50 on my order of 100 and won't respond to my requests.

BTW, I'm wiring up my passenger cars using Dale's CL2 circuit for constant current with conventional operation.   My breadboard test with 6 LEDs draws about 200ma at about 15 volts in.  I don't know if it's good or bad, but it looks nice with constant brightness.

The CL2 is rated at 20 or 25ma, depending on the model.  What are you using for 200ma?

 

Drawing 200ma is almost like incandescent lighting!   Typically, I see about .3 to .5 amps for passenger cars before LED upgrades.

John, I drew 200 ma with  6 LEDs in parallel, each driven by its own CL2.  The 200 ma is what I measured on the input from transformer to bridge rectifier.  I may have measured wrong, as I was using a new DMM.  Should I be pulling 120 ma then?

Carl,

  I like the results you achieved where the evidence of interior lighting is barely noticeable during daylight running.  Far more realistic than the "don't stare directly into the LED's" tanning bed illumination. 

 

I have some of the cheapo Lionel streamliners that I equipped with the 12V strip LED's and nothing else.  I never plan on running them under a command environment so they rarely see over 12V AC anyways.  I seem to recall they came to life at the slightest crack of the throttle.   It'll be interesting to see how long they hold up under AC abuse. 

 

 

Bruce

I'm not sure how you get 200ma with 6 standard CL2 chips, you're right that it should be around 120.

 

FWIW, I typically draw 20-25ma for a whole car lit with the LED strips.  My lighting controller maxes out at 45ma current, which lights the cars way to bright for my liking anyway.  I'd expect the lighting to be like sunlight with 200 or even 120ma of current to the LED's.

 

Here's a brief bench test of one of my cars with the lighting module.  I ran it on 18V DC so as not to have AC ripple cloud the results, the lighting is exactly the same if I run it on AC.  I can just measure DC current to 1%.

 

There is a lighted room shot, a dark room shot, and the current draw to produce that lighting.

 

IMO, there is plenty of light in these cars, and as you can see, insignificant power is consumed to do the job.

 

 

For my purposes, I'm very pleased with the strip LED's.  Since I've finished up the GM&O cars, I'm looking forward to trying the DC-DC buck boost circuit on basically the same PW style cars, but with NP North Coast Limited livery.  The parts should be here in 10 days or so.

 

battery powered, easy,  no problems with color, brightness, flickering etc.

How does battery power control the color?  The color of my lighting is the same as if they were powered with a battery, so I don't know what you mean there.

 

I have no problems with brightness, flickering, etc. and I never have to replace batteries or remember to turn them off.  

I see no schematic...

Were you using the DMM's "AC current" function?  What transformer is feeding your bridge?

 

Without descending into techno-babble, some DMM's take short-cuts on how they measure AC voltages and currents and can give deceptive readings.

 

If practical, measure the DC current (using the DMM's "DC current" function) AFTER the bridge rectifier and before your CL2 circuits.  I suspect you'll measure 120 mA or very close to it.

I took the liberty of including your attached schematic below.  Perhaps you attached the wrong schematic?  The following is not a constant lighting circuit; the brightness will vary with track voltage.  Also the capacitor voltage (20V) is under-rated for O-gauge command voltage operation.  I realize this thread is about low-voltage conventional so maybe your transformers can't reach command voltage levels.  But higher voltage capacitors (e.g., 35V, 50V) are about the same price so irrespective of control circuitry used, if modifying an O-gauge AC-powered car for DC LED operation, I'd use a higher voltage capacitor.