Make a better MARS LIGHT . . .

I'm curious to see what you come up with. I have a Lionel Lackawanna Trainmaster that I would like to ad one to. The catalog said it would have it, but it doesn't.

Well, I've been thinking about this very thing for my GS4 Daylight.  I'm thinking of combining two small LED's side-by-side with a circuit to slowly cycle them on and off.  I'll also add capacitance so they gradually fade out and in to simulate the light moving back and forth.

I also have some Arduino uP parts in house, and I may build a fancier version from them, you can buy some of the processor boards for as little as $10.

For detecting when the train is in motion, I just finished building a circuit that triggers off the motor voltage and closes a relay whenever the train is in motion.  In my case I needed it for a smoke unit control because I didn't have that available on the TMCC board I was using.  It will switch anything you like after detecting the motor has applied voltage of about 1.5 volts, which is barely moving, if at all.

You can also change the relay for one from the same manufacturer to the DPDT model and have control for both moving and stopped, say to turn on cab lights.

Here's the schematic for my little sensor.  Click on the thumbnail to expand it to a readable size.

I did a Mars light circuit for one of the train manufacturers (unnamed). Yes, 2 LEDs side by side. As one comes on slowly the other goes off slowly. The circuit? Here we go again- a PIC.

Gary, I see no picture, I don't want a PIC.

Thanks to all for the feedback. I have dug around quite a lot looking for a ready-built solution to this as I don't have the experience to design and build my own circuit. Here's a couple of things I have turned up I'll mention in case they assist.

1.  There is a Mars light LED board from Drumheads 'N More that the maker says will take either AC or DC power input; see here for details: http://www.ebay.com/itm/Cool-W...;hash=item27c7a448ed

However this seems to be a combination of a steady and a flashing LED and not the more prototypical effect Gary mentions.This is not quite what I am after although I'd like to understand how it accommodates both kinds of power.

2.  If the LED used in the GS-4 just blinks I'd guess there must be at least 3 or 4 VDC of filtered current being supplied to it, and I wonder if that would be enough to drive one of the small simulator circuits and at least one LED if I simply wired the board to the wires that now lead to the bulb? I think I will get in touch with Ngineering to see if they've got any ideas on this or any alternatives.

3.  I have a couple of Lionel engines that include perfectly good Mars light effects; one is my 6-14588 TMCC F3 A-B-A set and the other is my 6-34536 Legacy Texas Special E7. The former and a number of other TMCC engines use dedicated Mars light PCBs. I have thought about trying to adapt one of these but for a start I can't see this working within the confines of the space in my GS-4, where the only ready access inside the boiler shell is right behind the boiler so I can't get at the main wiring.

Besides that, apart from the NPR Berkshire PCB mentioned in my original post, these boards might not be readily adaptable as they seem designed for specific models. FWIW I attach the photos from Lionel's replacement parts site of the Berkshire and F3 boards.

I assume that in contrast Legacy uses software control and/or built-in main board circuitry to allow for Mars light functions as I do not see a similar form of  separate circuit board listed for Legacy engines.

4.  Lastly, I can get inside the GS-4 tender and I have wondered about putting any necessary power wiring and circuitry there and connecting it to the headlights by a wire tether. Possibly a bit messy and a last resort but at least there's space to work IF I can find a suitable combination of parts.

I don't know what the inside of your GS4 looks like, the MTH model I have is cavernous inside, so I'm probably putting my TMCC and lighting controls in it.  As I see it, the ones you first identified seem like good candidates.  Some of those were pretty realistic.

It's easy to generate DC from rail power, so that's not a problem.

The best Mars light made was by Kemtron in the late 1970's. 

It consisted of a small circuit board and a dual filament bulb.  The downside, it ran off a 9v battery.  (Pre DCC days, folks)

Rusty

Hancock

I have the 6-14588 TMCC F3 set and the Mars light is a standout.

I understand that the 1:1 real  Mars lights were installed by the Santa Fe shops.

Not sure which pca board is related to F3. Is it the 691-MAR1-A0?.

Price/time wise it would not be practical-but I have have a friend who could

"reverse engineer" the circuit of the the pca. Purchasing a 3rd party board would be a lot easier. Either way-I would shy away from connecting any new Mars pca to existing power sources.

Floyd

Floyd,

I totally agree about the 6-14588; all its lighting is a gem and like certain Lionel models from 2005 I think that it has yet to be bettered. It's also never given me a single QC/service issue. These are the main reasons I looked at its light circuitry when I decided to try to upgrade my 3rd Rail model although obviously one is a diesel and the other a steamer.

The light control boards I have tracked down for that F3 set are the second two pictured in my earlier posting; the 691-MAR1-A01 is the Mars PCB for the 6-38050 NPR Berkshire. I have one of these PCBs but no TMCC engine that I'm prepared to experiment with using it.

I'll see what Ngineering says about using its simulators with a TMCC loco. Frankly I'd be very happy to have someone reverse engineer a setup for my GS-4 if that turns out to be the only option.

I second Floyd's suggesting that you avoid trying to connect to the 5V from the TMCC/RS4 boards.  They have very limited current capability, and you could easily cause more problems than you solve.

For command operation, getting DC power from track voltage involves only about four components, a diode, two capacitors, and a 5V regulator, such as the LM7805T for several hundred milli-amps of 5V with no heatsink requirement.  You can assemble these in a small space and just encase them in heat-shrink, which is what I do.

I'm thinking of trying to do a MARS light using three LED's and a microprocessor to drive them in sequence using the PWM outputs.  I can vary the brightness of the three in a sweeping pattern that should simulate the MARS light pretty effectively.  The same Microprocessor is going to be there anyway for a couple of other control functions, so this is just icing on the cake.

I thought I'd revive this old topic, since it may still have relevance.   I've been tinkering with an Adruino processor and seeing if I can create a MARS light.  I've actually gotten what I feel is a pretty good looking one with just a single LED.  The great part about having a processor is you can do pretty much any sequence you like.  So, I use the PWM feature of a port to vary the brightness of the MARS LED.  The sequence I use is:

Ramp up at a 10ms rate from 0 to 64 (0 to 1/4 drive).

Turn on the LED full for 100ms at 255 (full drive).

Turn off the LED for 100ms at 0 (no drive)/.

Turn on the LED full for 100ms at 255 (full drive).

Ramp down at a 10ms rate from 64 to 0 (1/4 to 0 drive).

REPEAT.

This makes a really nice looking MARS light.  I'd post it here, but for some reason I can't get PhotoBucket to give me a URL that work here.

Got it over on YouTube, here's my prototype MARS light, this is my first cut.   I think it looks a lot better than what K-Line originally did.

Wow, that does look good!

I thought it worked pretty well for a single LED.  In my GS-4, I don't see how I'd fit multiple LED's in the MARS light anyway.  The only way I could do that is with light pipes from LED's inside.

gunrunnerjohn,

Your prototype Mars light looks really good. The oscillation effect is so much better than the regular alternate-blinking LED's or incandescent bulbs. Until now I've never seen any Mars light effect in O gauge that was as good as the Richmond Controls EZ-Lite circuits for N and HO scale. I think you've really nailed it.

Thanks, I fooled around with two and three LED's, but then I looked at pictures of a real one running, then I came up with the algorithm that I used.  It's also very easy to fine tune the timing when you have a microprocessor controlling the lights. 

I'm going to have a motion sensing circuit wired to the motor so I can tell when the locomotive is stopped.  When that happens, I'll turn off the MARS light and turn on the cab lights. 

That really does look good. My own efforts to achieve something similar have been stymied by lack of (a) space in my Sunset GS-4 to fit anything but the smallest control board and (b) time to work on the parts I have got from Ngineering, which are supposed to simulate a Mars or Gyralight using very small LEDs. A single bulb would certainly look better.

I'm also still perplexed about wiring a power supply and motion sensor to whatever board I use. I've got space for a bigger board in the Sunset E7s I am also hoping to modify. These seem to use a different light control system than the GS-4 although the effect (a blinking LED that comes on when the loco moves forward) is the same.

P.S. What Arduino processor are you using and how is it powered? I've also found that Miniatronics make a very good adjustable single-bulb LED Mars light for HO but again the problem is powering it in a 3-rail model and motion control. 

I do my debugging using the Arduino UNO board, but I was going to use the Arduino Mini or Nano when I install the lighting.  I like the Nano boards because they have a USB on-board, so loading new versions is much easier than with the Mini, even though they're a tad larger.  I threw in a picture of the two boards so you can see the difference.

As far as powering a board, with command you can just use track power.  A small bridge rectifier, three terminal regulator and a couple of caps and you have regulated DC power. 

For motion sensing, I connect a Schottky bridge rectifier across the motor and feed it into an opto-isolator.  I get a digital indication that the locomotive is moving in either direction.  If you want sensing only in one direction, you could just use a Schottky diode to drive the opto-isolator, then you'll only get an indication in a single direction.

Here's the circuit I'm working on that provides power and does the sensing I mentioned.  I'm also using it to control the cab lights and I am throwing in Rule 17 lighting since it's basically free.

You need to right click on the diagram and use View Image to expand the diagram to full size.

Thanks John. I think I'll confine my efforts to putting in a Mars light although I can see the sense of Rule 17 lighting in the Sunset E7 I hope to modify as of course it does not have Legacy.  (Pity about that as the Legacy E7 I have is a great runner and smoker with a more than acceptable Mars feature although not a patch on yours!)

Rule 17 was basically free, so I figured I'd put it in.  I also don't like how the K-Line leaves the MARS light running when the locomotive is stopped, that's certainly not prototypical I would imagine.

:"here's my prototype MARS light, this is my first cut."

i think that looks fantastic!

Maybe I should start building these for sale.  

Sorry but I like these better

http://ngineering.com/marslight_video.htm

They are very nice Chuck, I'll have to see one and maybe I can emulate it.  It's a computer, I can do anything I need to.

I have some of Ngineering's kits and you really have to squint to look at them as the components are tiny given that they are made primarily for N and HO. I'll have to report back next month once I have actually tried building one for use in an O scale model. I will say to their credit that the guy who runs Ngineering was extremely helpful to me and is trying to come up with his own O scale solution. I don't pretend to understand the science behind it however.

Since the boards drive LED's, why is it important that they be made for a specific scale?  I see no reason you couldn't use those in an O-gauge locomotive.

John,

It's the power supply. The Ngineering boards are made for DC or DCC systems whereas I am trying to adapt them for 3R O scale models. I also may have complicated matters by telling the maker about the existing Sunset Mars light wiring, which I was hoping to tap into in order to avoid having to run an entirely new track power link.

Well, generating DC from the track voltage isn't all that complicated, so I don't see the power supply being a major issue.  You'll have to whack a few components together, but no big deal.

I invested in a couple of Ngineering MARS light simulators to do some actual in-hand comparison.  Here's the head-to-head comparison of mine and the N8031A.  Both of them look pretty good, I'm not at all displeased with either of them.  In case you can't tell, mine is on the right, the Ngineering model is on the left.

I also have the N8031B that uses three LED's to do the simulation, but I haven't wired that one up yet.  I may actually do a small bundle of fiber optic from each of three LED's to try to make a unified lamp for testing with this one.  Watch this space, but it'll be after I get back from our Christmas vacation.