Flicker-prevention for Kato Interior LED Lighting on DCC

[KenS]

I'd mentioned that I've been working on a circuit to add to my Kato EMUs to reduce the annoying amount of flicker the interior lighting has.  I'm using the standard Kato add-on lighting kits (11-209/11-210).  The reason for the flicker isn't clear, but I think it's a problem with the brass strips under the floor making intermittant contact with the "L"-shaped strips that bring power up to the lights.  I tried soldering the lights to the "L" strips, and that didn't solve it.  And with all-wheel pickup and relatively clean track, I don't think it's a pick-up problem.  And every car on the train flickers (well, the motor car is much less affected than the others).

So I decided to add some circuitry to bridge any interruptions in the power supply using a capacitor.  I chased down a few false trails (mostly thinking I needed to make things more complicated than necessary), and ended up with a very simple design (see diagram) that you can put together for about $2 and an hour or two's work.

Parts: (prices are for quantity 100)

Rectifier: Fairchild Semiconductor 58K1812 ($0.219)
Capacitor: Kemet  T491X107K025ZT, 100 μF Tantalum Surface-mount ($1.73)
Resistor: Vishay PR01000101500JR500, 150 Ohm, 1W, 350 V, 5% ($0.076)

Note: this circuit is specific to the low power draw of Kato's LED lightboard. A larger capacitor might be needed with other kinds of LED-based lights, and use with bulb-lighting is probably impractical.  The circuit may also work on DC, but not as well since the voltage is typically much lower.  See the design details on my site (link below) for the reasons I did things this way if you want to customize this for other uses.

Photo 1: the circuit diagram

Photo 2: the parts used to make the circuit

Photo 3: the assembled circuit before installation

Photo 4: the installed circuit

and a video showing it in operation. The lead car just has a normal lightboard, the second car has this circuit installed.


<a href="http://www.youtube.com/v/HiTG3c_57Ws" target="_blank">http://www.youtube.com/v/HiTG3c_57Ws</a>

I've a much longer writeup on my site, detailing the circuit design, some alternates, and the steps I went through to get to this point, for the curious, but the necessary details are included in this post. I should also note that I'm hardly the first person to design something like this, and the page on my site has a references section with links to similar designs (none exactly the same; everyone has slightly different ideas about what is most important).

[CaptOblivious]

FWIW, tantalum caps arent polarized...

[CaptOblivious]

Just saw the movie, looks good! It would be interesting to see if all this could be put on a small, easy to assemble and hide PCB.

I understand the necessity of the bridge rectifier, but it seems a shame that the one already in the light board is made redundant; Now I'll have to look closely at mine to see if there's a way to wire the thing up such that you could use Kato's rectifier instead of adding a second one.

This circuit should work just fine on DC, BTW. It just won't keep the lights at a constant brightness…but then, that's not what the circuit is designed to do. It certainly would stop flickering on DC though.

Interestingly, your Kato units flicker a whole lot more than mine do!

Anyway, nice work!

[inobu]

Ken,

The first N scale UPS system. That's the way to do it!

Inobu

[KenS]

FWIW, tantalum caps arent polarized...

Maybe not all of them, but the ones I bought are clearly marked with a polarity.  Check out the data sheet from the Digikey reference.

This is an older E231 (10-551, apparently from 2007), and it may be that it had a poorer design than later ones.  I never noticed this much flicker on the cars with factory-installed lighting on my trains that have it.

It might be possible to splice into the lightboard traces, but I wasn't entirely sure what that circuit was doing and I couldn't follow all the traces.  Plus, my eye-hand coordination isn't up to that precision of soldering (maybe with a bit more experience with this small stuff, but not today).

A small PCB would need to be small enough to fit into the very limited space at the end or between the benches, unless you want to remove some of the structure, which I want to avoid.

Inobu: Clearly there's a marketing opportunity there.  Call it a "DCC UPS" and sell it for $50 per car. 

[CaptOblivious]

I stand corrected about the polarity! You are right sir. (Interesting note: I apparently already had that datasheet in my downloads folder. I wonder why? Then again, I have several hundred assorted and various datasheets on my HD, so perhaps it shouldn't be surprising)

[quinntopia]

Ken, you are the MAN! Thank you so much for putting this tutorial together!  I don't have time to apply it right now, but its one of the projects I eventually want to get around too! 

[nik_n_dad]

I was always intrigued by something like a "dcc ups"....... http://www.lenz.com/techinfo/usp.htm

[KenS]

The Lenz solution sounds like a combination of memory to store the last DCC command for some period, and some kind of storage (probably a large capacitor based one the durations noted below) to store power to run the decoder and motor.  There's another page that makes it clear that the power storage module is rather large (in a physical sense):

http://www.lenz.com/products/decoders/currentdecoders/power1.htm

But yeah, that's a "UPS" for real.   I hadn't been aware of that before.  Pretty cool.

[Martijn Meerts]

The Lenz system works great. I have 2 of the Lenz "spur 0"locomotives, which are equipped with the USP system. You can put a piece of paper on the tracks, and the loco will just drive over it without problems. It can run about 4-5 seconds without track power (and that's with light and sound turned on.)

I also have one of the small USP things, which I bought for the Tomix railbusses, but the thing is too big for that. In fact, it wouldn't fit in any N-scale locomotives, and in only very few motorcars.