Train sucked all the amps out of the transformer!

As GRJ says, a bridge rectifier solves the direction change problem.  However for conventional DC operation, I'd wonder if there's enough track voltage at slower speeds to drive a 12V DC LED roll. Lower voltage operation is discussed in the 300 LED Rope thread though focused on AC operation so some techniques won't apply to you.  You might want to start a separate DC thread.

You might want to use some series diodes strings for the motor to raise the track voltage enough for the LED's to run with that circuit.  Obviously, a voltage doubler doesn't work here as it requires AC to function.

Railride, was there a problem that caused you to add the choke?

The choke prevents the current-regulating circuit from degrading a DCS signal.

 

While the Superliners I'm working on will be run conventionally at first (I'm pulling the train with a pair of Williams Genesis units, seeing that they're the only easily available ones in the current Amtrak scheme), I'll eventually upgrade them to command operation in a mixed TMCC/DCS environment. Adding the choke now means I won't have to reopen the cars and add it later once the locos are upgraded (however long that takes), and it won't hurt conventional operation in the meantime.

 

---PCJ

FWIW, I add the chokes to all my LED lighting projects, I do have DCS and I'd like not to have an issue when I run them.

There is a pic of the choke I use in this thread

 

http://www.jcstudiosinc.com/Bl...=487&categoryId=

 

I do not use DCS but I made this for person who did. No signal issues were reported.

 

Dale H

Thanks RR & GRJ.  I know the purpose of a choke, but was wondering if this circuit has been causing problems without a choke.

Lighted passenger cars with constant intensity lights have been known to cause issues with DCS.  This is just cheap insurance that there will be no problem.  The thing is, the problem is cumulative, so as you add offending circuits, the DCS signal gets weaker.  So, you don't necessarily know if you're screwing it up until it gets bad.

 

From my understanding,anything with a capacitor can be an issue. This includes PS1 CV circuits for headlamps,and lighted cars with CV circuits.

 

Dale H

It would depend on where the cap is, and the value.  Anything that attenuates the signal between the center and outside rails can be a problem.

Without the cap, you'll find this.

 

 

With the cap, but no load, you'll end up with DC.  As the load increases, you're likely to see a waveform something like this.

 

GRJ, yes, that's a good graphic.

 

I'd like to get closer to 12 volts (max) on the LEDs.  The 68 ohms has them dimmer than I want.  Does increasing the resistance brighten or dim the LEDs? 

 

For the station platform lighting, I think I'll skip the regulator

Higher resistance, lower current, dimmer LED's.   Try a 47 ohm resistor, if that's still dim, try a 33.  I'd be using 1/2w resistors.

 

I'd like to say I got that waveform with my digital 'scope, but I just downloaded them from the web.

yes I know that's the general rule with Ohm's law. However, since a control circuit is involved I didn't know if the results could be opposite.

It's true in this case.

I must be getting old.  If I'm only inputting 10.1 VAC, I shouldn't expect too much output.  As the graph notes, when current is drawn, the wave form changes, which means there should be greater DC voltage drop than one would expect from Ohm's Law.

Depends on the current, the size of the cap, and the capacity of the transformer.  If you have a really huge cap, and a really powerful transformer, even drawing power from the load, the output voltage would be very close to the peak voltage of the rectified output.

 

The problem with using unregulated supplies is you can't be certain what voltage you'll get under varying input voltage conditions.

 

A one foot length of the LED's would probably be 18 of the LED's if these are the 300 LED 5 meter strips.  This number would require 120ma to light at full intensity.  The easy way to do this task is use the LM317T regulator in a constant current configuration to supply 120ma to the strip.  You'll get full intensity and the voltage input won't affect the supply current.  You need about a volt and a half of headroom from the input of the regulator to the output.

 

Truthfully, that strip running at full intensity is going to be pretty bright, I rarely run LED's at full current in modeling applications.

I notice you use the term "constant current configuration" but that the 317 is billed as a voltage regulator.  While voltage affects current if the resistance remains the same, I suspect that the term you used implies something else.

And, I use the term correctly.   The LM317 can be wired in a constant current configuration, and it's perfect for this kind of application.  Note the resistor value sets the current, pick your poison.  I've included the configuration below, and if you visit the LM317 Current Calculator you can quickly compute the resistor you need for any desired current.

 

 

Thanks, GRJ, that link is excellent.  The schematic you included in the just-above post is the same as you posted a few days ago, but with a diode vice a full-wave bridge.  Correct.

 

Now let me ponder a moment.  If the 317 were to be set up as a voltage regulator rather than a current regulator, then it could be set so it didn't matter how many LEDs you have in a circuit.  Correct?  This would permit one to assemble & put on the shelf a bunch of them for use with any application using these 12-volt strings, regardless of length. 

You can indeed use it as a voltage regulator, AAMOF, for an even easier method, just buy the LM7812 regulator, it's a fixed 12V regulator.  It eliminates the resistor, just put DC in the input and get 12VDC out the other end.

Wish I had been aware of the LM7812 before I ordered 20 of the 317's.  I want to convert all of my cabeese also, to eliminate flicker

What can I say, you just mentioned the fixed voltage idea.   I like the constant current approach as I can vary the intensity with a single resistor.  If I use a 12V regulator, I have full intensity which usually isn't what I want for model lighting.

How is a cv working regardless of led number?  Are these designed to work on 12 v and have groups of leds in series and parallel?  G

The reels have three 3V LED's in series, and a 150 ohm resistor in series with that.  The design voltage for full intensity is 12VDC.  If you have a larger string, each 3 LED section runs independently on the 12V.

GGG:  The link GRJ sent to the current calculator is to a British web site that contains interesting discussions of the LM317.  You can set it up to provide constant max current regardless of load or voltage, or as constant max voltage regardless of current.  All within its capabilities, of course.

Yep, the LM317 is a pretty flexible little device, I use them a lot.

Where is the thread to which you refer?

 

https://ogrforum.com/t...-rope-light-question

 

It does kind of wander all over the place but there's a lot of good nuggets in there if you have the patience...

Stan, if I want straight 12 volts, I use the LM7812, no extra parts.  However, I normally like to run with constant current, allows me much more intensity adjustment.

RJR,  Thank you.  I was just trying to clarify that you can't put an unlimited number of LEDs in series with a 12V source.  Each LED will drop 2-3V depending on the spec.  Same for current.  If each individual 3 LED segment is using 10ma, and you have 6 segments in the passenger car, then the regulator is seeing a 60ma load.  G

I fabricated a few of the voltage regulators.  Photos are attached.  After finishing the first, I added a 470mfd capacitor; the second I built with ojne.  Unfortunately, it is not adequate to compensate for the lousy Williams rollers losing contact with the track.  I used 2 strips in the one car I've finished.  The LEDs I got were the cool white, which has a blueish cast I don't care for.  I've ordered some warm white.

Just add higher value filter capacitors, it's possible to filter them and remove the flicker.

It's more than a simple flicker.  The lights actually go off and on.  

RJR, what values did you use in your LM317T circuit?  It may be the color-rendering on my screen but is the circled resistor marking Brown-Red-ORANGE-Gold (12,000 ohms)?  Do you have a multi-meter?

 

Also, I assume you at least saw a change in flicker behavior (slow fade vs. on-off) when you added the 470uF cap? 

Stan:  The LEDs don't fade out; they just go on & off.  The 470mfd cap made no difference in this.  Checking the Williams car that was the first installation, I find that it has a lousy ground.  If I connect my ohmmeter between wheels on each of the 2 trucks, there is spotty contact.  The Williams trucks don't have a pick up finger to the axles, and a loose connectioon to the car frame.  I'll have to study this more.

 

Sublinked to a link gunrunnerjohn had posted, is a info sheet on resistor sizes for various voltages:

http://www.reuk.co.uk/LM317-Voltage-Calculator.htm

 

I used the schematic (below) on the first page of that link, and per the table later on in the link, selected 470 ohms (R1) and 3300 ohms (R2) to yield 10.03 volts.  Can't vouch for the colors as they appear on the screen, but the R-S packs said 33k.

 

Schematic:

 

I assume you meant 3.3k (not 33k).

 

Did you confirm you get 10V with a multi-meter?

Over what track voltage range did you operate/test this with?

Can you post a photo of the inter-connect back-side of the circuit board in question?

First off, the cap goes to the input side of the LM317 for best results.  Second, the combination of 470 and 3,300 should yield pretty close to 12 volts.

 

I still think the constant current solution is superior for most lighting uses.

I don't mess around using analog regulators as Buck Converters are much more efficient.

http://www.ebay.com/itm/1pcs-LM2596S-Power-Module-DC-DC-Buck-Converter-Step-Down-Module-LM2596-/121129615011?pt=LH_DefaultDomain_0&hash=item1c33e33ea3

Infact been kicking around the idea of mounting the Buck in a baggage car and then tying all the cars together with mini 2 pin plugs. Also in the baggage car will be a LiOn battery powering all the lights. Then can remove all the power pickup rollers to get rid of the drag. We do this in the large scale world and it works great along with using ball bearings in the journal boxes.  

 

Stan, yes 3300 ohms.  I checked with voltmeter and got 10.1 volts.  Given the tolerance of resistors and the unit, that's close enough to the 10.03 volts the table shows for those resistors.

Grj:  not 12 volts.

 

Am working on a Williams observation car & found it has grain-of -wheat incandescent bulbs in marker lanterns, both burned out.  Had a few gow bulbs on shelf, so am also feeding them from the regulated 10.1 volt supply for led's.  Now will tackle improving ground feed from wheels.