No Capacity For Capacitors

4,700 microfarads isn't enough. 

If you want to stop lights blinking in moving train cars going across switches, it takes impractically large capacitors to sustain incandescent lamps for more than a fraction of a second. I've tried it.  Might work for LED lights that use much less power. Or add a second pickup roller to the car if it only had one center-rail pickup to start.

The larger capacitors needed to help with flickering will be very large in size for what you want to do and may not fit inside the passenger cars. I tried something similar for some pre war passenger cars and finally gave up on the idea.

Like ACE mentioned try using a second pick-up roller.

Lee Fritz

Trying to get capacitors to eliminate flickering incandescent lights is like trying to fill the ocean by spitting in it.   If you want non-flickering lights, consider a conversion to LED lighting.  Then a reasonable sized capacitor will do wonders for the blinking.

The time constant of 4700uF and an 18v bulb is roughly like .005 times 75 = .375 seconds. That's the time it takes the charged voltage (18 vdc) to discharge down to about a third or about 6 volts. The actual time constant relationship is more complex and the resistance of the bulb is non linear with voltage, but this should give you some idea. If you were committed to an incandescent bulb, a starting point might be more like 500,000 uF which would take 100 times longer to get down to a third of the starting voltage. Or the best way to go would be with LEDs and specifically the Lighting circuit kits at Henning's Trains. I'm using them with LED strips (about 9 or 12 LEDs) in some IC passenger cars and it's really neat to see them not blink when they go around the layout over switches, etc.

Why not use tethers and wire all of the cars together?  There's no way ALL of the rollers will be on a dead spot at once.  Besides that, sometimes lights in the real trains do blink.  I rode SEPTA (ex-Reading) electric commuter trains for years.  On occasion (I guess when the pans lost contact with the wire) the lights went out 

I'd like to see a self-contained battery-power LED light that turns on for a couple minutes when it detects motion, and put one in each car. Seeing some of the electrical toys and gadgets you can get at the Dollar Store (like solar-powered dancing flowers), it would be cool if something like that could be made for cheap. Of course it would help if you had easy access to the interior of the car for installation.

Ted, IMO it's a far easier task to convert to LED's and have better lighting that is also flicker-free.  If you're wiring tethers, you have to take into account protecting the wiring in the case of a derailment, hence something like PTC's in each link. 

It wouldn't be difficult to do a motion triggered lighting scheme, I just prefer no batteries, so I'm not all that motivated to build the module that would do that. 

Try the 2.5F supercapacitors from Digikey. 

I made some circuit boards for lighting circuits back in 93-94 that replaced the red lens in the observation cars of postwar and mpc passenger trains.  One version came with no capacitor which meant the lights went out after the power was cut off.  The other two versions had 0.5 Farad and 1.0Farad capacitors in them.  The 1.0 Farad circuit would keep the lights on for between 30-60 minutes after track power was removed.  I would recommend a small value resistor in series with the larger capacitors to prevent them from drawing large currents during their charge time.  The higher value capacitors take a while to charge up.  I also used a 1.0Farad capacitor inside my VCR which meant it never lost programming during a power outage.

Let's put aside the LED alternative for a moment and take your initial question at face-value.

To "engineer" a solution, some math is required.  Let's say you have a couple incandescent bulbs running at 16V and drawing 5 Watts total.  You want to sustain a, say, 1/2 second loss of power without the lights blinking/flickering.

Compute the total energy needed.  Power (watts) x Time (seconds) = Energy (Joules).  Example: 5 Watts x 1/2 sec = 2.5 Joules.

The Energy (Joules) stored in a capacitor is 1/2 x Capacitance (Farads) x Voltage x Voltage.   Example: 1/2 x 4700uF x 16V x 16V = 0.6 Joules.

Your experimental capacitor is under-sized by a factor or 4 or 5 times.  But on top of that, as alluded to by others, a capacitor has the "annoying" characteristic of dropping its voltage as it dumps its energy.  Compare this to a rechargeable battery which, relatively speaking, holds is voltage relatively constant; then the voltage drops fairly rapidly at the end of the "charge".  As voltage drops, the bulb brightness dims so you need a circuit to boost the rapidly decaying capacitor voltage up to 16V DC...I'd look into an eBay "boost" voltage regulator for a few dollars (free shipping).

Peeling the layer of the onion a bit more, you cannot fully recover all the stored energy in the capacitor and there are transaction fees (losses) in converting voltages.  So if you choose this route, I'd plan for a larger capacitor of, say, 10 times what you have now...or 47,000uF which could be ten 4,700uF caps stacked in parallel if that's all you have available.  Obviously this could take up quite a bit of space.

So now what?  OK, as RJR suggests above there are newer capacitor technologies that have tremendous energy density relative to the conventional "electrolytic" technology of your 4,700uF cap.  A 1F (1,000,000uF) 5V supercap (less than $5) is about the same size if not smaller than your 4,700uF/25V electrolytic.  It can only be charged up to 5V but using the formula above, that's a stored energy of 1/2 x 1F x 5V x 5V =12 Joules!  That's 20 times the energy of a single 4,700uF capacitor charged up to 16V.

But now you need some circuitry to manage the charging of a supercap to keep it to 5V, and follow it with a boost regulator module to pump it up to 16V DC when needed.  Again, a few eBay voltage modules for a few bucks a piece can handle the circuitry side. 

So if you want to stay with incandescent bulbs, you can solve the blinking problem! 

Stan's post is an excellent reason to consider the LED upgrade.

Alan,

Just  a suggestion: Put two rechargeable  9 volt batteries in series to the 18 volt lamp. when the power goes off. The lamps will stay lit until the power comes back on..  

Good Luck ,John  

NYC,SUBWAY TRANSIT SIGNAL posted:

Alan,

 

Just  a suggestion: Put two rechargeable  9 volt batteries in series to the 18 volt lamp. when the power goes off. The lamps will stay lit until the power comes back on..  

Good Luck ,John  

 

Will this work with AC track power or would you need to add a DC converter circuit as well?

George

Batteries are DC devices and take very unkindly to being fed AC!  You'll need a charging circuit, and unless you want to cook the batteries, you'll have to arrange to limit the charging.  This is more complicated that it's made out to me.

Personally, I wouldn't touch the 2-9-volt battery option without a means of automatically keeping power supply and batteries separate. And be sure to shut the switch when done running.   I doubt the practicality of this approach.

I go with GRJ on this.  I converted all 15 of my Lionel & Wiliams passenger cars to LEDs.  Lighting in the cars is even; no bright spots.  Can run multiple passenger trains on a TIU circuit.

I had a feeling the battery solution was not an easy answer.  You could convert the lights to battery with a switch to isolate the batteries.  It is possible that 9 volts would be sufficient power.  Adding a charging circuit with a power isolator would be interesting but probably is not cost justified.  LEDs seem like the best answer given the trouble.  For me, I like the flickering lights.

Again, if we leave the pros/cons of LEDs as a separate discussion, the rechargeable battery approach can be made to work with the incandescent bulbs.  As GRJ mentions it's not a simple as just tossing in the 2 batteries.  If you are going to "engineer" a solution, you need to work the numbers a bit.

So if using, say, a typical 160mA-Hr rechargeable 9V battery (Harbor Freight has one on sale now for $6) you indeed have a whopping amount of stored energy.  In round numbers you have 9V delivering 160 mA for 1 Hour.  So using the equation from my previous post, that's 9V x 0.16A = 1.44 Watts for 3600 seconds (1 Hours) or over 5000 Joules of energy.  So no question it can run the lamps for several minutes since you're only look for a handful of Joules to ride-out a track voltage drop-out. 

But 9V rechargeable batteries can be squirrel-ly creatures when called upon to deliver large currents as you would when driving incandescent bulbs.  You'll still need a voltage regulator circuit to maintain the voltage or else you'll still see momentary flicker when track power is lost and regained.

And, as RJR mentions, there's so much stored energy that you really want to turn off the battery when it's done protecting against the blinking or else the lamps will stay on for minutes when you're done...and conversely require an inordinate amount of charging time when you start an operating session with discharged batteries before it can protect against blinking.  But, yes, you can add a circuit to limit the battery output to, say, 1 second following a track voltage drop out so the battery is only used for blinking/flicker protection.  Point is, by the time you add charging circuitry, voltage regulation circuitry, shut-off circuitry, etc., you're out several dollars in parts and assembly hassle...plus over $10 for two 9V rechargeable batteries.

Believe me, I can understand the interest in using incandescent bulbs (vs. LEDs).  For one they have a prototypical light spectrum that LEDs cannot yet duplicate.  If you are seriously interested in pursuing blinking/flickering protection from an electrical perspective for you bulbs, I'll cobble together a prototype using eBay modules and a super-capacitor for you.  It would be less out-of-pocket than 2 9V rechargeables plus the battery-management circuitry . 

I assume you've already followed the earlier advice of cleaning the wheels/tracks/rollers to rule out or demote the effects of those connections.

If the bulbs are like those that come in cars, the batteries will die really fast.

The problem with the flickering lights is that they lead to lawsuits.  The passengers claim vision problems resulted from the flickering and sue the operator.

He said he using DC. for the power Here

That's how Alan, Started  off Here! 

(I'm experimenting with capacitors to stop lights from blinking.  I put a 4700microF 25V capacitor in parallel  with a DC current going to an 18V bulb.  I get it up to about 16V and let it sit for a while (30 seconds to 5 minutes).  All I seem to get is a very slight amount of slow dimming when I cut the current.  If I do a rapid open-closed circuit experiment the light still blinks.)

I  also  agree about using Led's too.

Or am I missing something Here.  (DC) 

Good luck, John

I'm amazed at some of the answers here.  As a physician I realize the depth of knowledge that a field can command.  Other than for the relatively simple stuff these discussions are beyond me.

However, I enjoy reading them as they portray a knowledge base that's pretty much in another world from mine.

The funny thing is that my bulb experiment was just an experiment to gauge the operating characteristics of a capacitor.  I'm at the Capacitors For Dummys stage.  I was thinking about putting capacitors into lighted cars, but I really didn't get beyond the "just try it on a bulb first" stage.  I'm pretty much in the messing around stage.

Thanks for all of the posts, and I'll sock away these ideas for future reference.

Alan