Using a capacitor with LED lighting strips

[Guest]

Hi

 

I have decided to have a go at installing some LED lighting strips in my carriages. The lighting strips are those that come in reels and you can cut them into sections.

 

I intend the lights would be powered from the DCC track. My track is set at 14 volts, although it seems to be closer to 13.5 to 13.7 volts (according to my Z21). A home made stay alive unit with a bridge rectifier and capacitor should then modify the charge to dc.

 

However I need some advice or tuition on a couple of things.

 

I understand there is a voltage loss with the bridge rectifier of about 1.4 volts. So my 14 volts DCC would come back to, say, 12.6 volts. Most shops or ebay sites simply rate their lights at 12 volts and don't mention any tolerance.

 

My question is would the LED strip handle this extra voltage?

 

Various threads and sites I have read have suggested a capacitor rating of 16 volts. 

 

My question is that if the lighting strips are rated at 12 Volts, would the voltage from the capacitor be too 'strong' and affect the LEDs.

 

Thanks for any help

[John ks]

The voltage rating of the capacitor is the maximum voltage that it will handle. If your supply voltage is higher than the rated voltage then you risk the capacitor exploding

I would use a capacitor rated at 30V or even 50V

When the DCC track voltage is rectified to DC I suspect that the Dc voltage will be more than the 13.5 to 13.7 volts (according to Your Z21).

 

The LED strip will probably be able to cope with this extra voltage

 

The biggest problem with these LED strips is that they are very bright,

 

A resistor in series with the led strip will dull them down to a more acceptable level

 

Start with a 500ohm resistor & increase or decrease its value until you get the light  to an acceptable level 

 

Order of parts is Rectifier, Capacitor, Resistor, LED strip

 

 

John

[Guest]

Thanks John for your advice.

 

cheers

[Tricky Dicky]

A capacitor wired in series will not pass current through, it needs to wired across the output of the rectifier.

 

Richard

[Pete the Elaner]

 

 

My question is that if the lighting strips are rated at 12 Volts, would the voltage from the capacitor be too 'strong' and affect the LEDs.

 

No.

That is like using 15A cable for track current which you never expect it to exceed 1A. This is not a problem at all since the wire itself cannot introduce any current. It only allows more to flow & bigger cable = less resistance, which is why some use chunky cable for their DCC bus.

 

The only voltage supply in your case is the supply itself. If this is 13.7v & your LEDs are 12v then in theory you could have a 1000v capacitor & this would do exactly the same job as a 25v one because it can't introduce any voltage. It would be a lot bigger though, probably impractical to fit into a OO coach & awkward to hide in a O gauge one.

 

On the other end of the scale, you don't want to use components rated at lower than the applied voltage or current. These would run too hot, which can cause them to melt plastic and/or fail.

This can be useful in some circumstances. A safety cutout could be a bi-metallic strip which bends when hot to break the contact. A fuse wire is designed to melt at its capacity to protect the device 

I understand that electrolytic capacitors bulge then explode when overloaded, leaving a sticky mess. I have seen one bulge but I don't want to see one explode!

Edited December 13, 2018 by Pete the Elaner

[John ks]

A capacitor wired in series will not pass current through, it needs to wired across the output of the rectifier.

 

Richard

Tricky 

You are correct, in my statement "Order of parts is Rectifier, Capacitor, Resistor, LED strip" I assumed(we all know what happens when you assume  ) it was obvious  that the capacitor was in parallel across the rectifier

After reading your post I see that I could been misunderstood 

Here is a diagram which should clear any misunderstanding

 

John 

[Guest]

Thanks for the additional comments.

 

John’s diagram was my understanding. But it helps me to see shown like that.

 

I have done some more travels around the net and now seen 2 variations on that plan.

 

The first had the capacitor connected at the other end of the lighting strip. I imagine that would be good if there was limited space to hide the items at one end of the carriage.

 

The second used a variable resistor/potentiometer.

 

I would appreciate any thoughts on those

 

Cheers

[dhjgreen]

Capacitor after the resistor would give a greater smoothing effect. Do not use a variable resistor, anything small would have a low power rating.

[Crosland]

Capacitor after the resistor would give a greater smoothing effect. 

 

It would negate the effect of the resistor.

 

The resistor doesn't just limit the current to the LEDs, it drops some voltage. If the capacitor is fitted after the resistor it will simply continue charging up to the full track voltage until there is no current through, and no voltage drop across, the resistor.

[dhjgreen]

It would negate the effect of the resistor.

 

The resistor doesn't just limit the current to the LEDs, it drops some voltage. If the capacitor is fitted after the resistor it will simply continue charging up to the full track voltage until there is no current through, and no voltage drop across, the resistor.

I generally agree with you but not this time. The load of the LEDs will prevent that. Otherwise there would be no volt drop with the resistor after the cap. The thought experiment is: leave out the capacitor, would there be a volt drop, answer yes. Edited December 13, 2018 by dhjgreen

[sharris]

When the DCC track voltage is rectified to DC I suspect that the Dc voltage will be more than the 13.5 to 13.7 volts (according to Your Z21).

 

 

Surely the DCC output is a square wave centred around 0V with the length of a cycle determining whether the code contains a 1 or 0 (1s are shorter than 0s). For square waves the RMS of a signal that is symmetrical about 0 equals the peak voltage, so I'd expect the rectified and smoothed voltage to be the same as the peak voltage or the RMS voltage minus the rectifier voltage drop. 

Edited December 13, 2018 by sharris

[Vecchio]

Looking at that I would also put a diode with a resistor in parallel before the capacitor (to limit the charging current). This way you avoid flickering LED when the cap is charging but have full power when needed (similar to the loco decoder stay alive functions)

[cliff park]

If the voltage is too high don't forget the simple option of half wave rectification. In other words one diode in series with one of the supply leads, and forget the bridge rectifier. It may work better with a bigger capacitor, depending on how much current the LED's draw.

[dhjgreen]

Surely the DCC output is a square wave centred around 0V with the length of a cycle determining whether the code contains a 1 or 0 (1s are shorter than 0s). For square waves the RMS of a signal that is symmetrical about 0 equals the peak voltage, so I'd expect the rectified and smoothed voltage to be the same as the peak voltage or the RMS voltage minus the rectifier voltage drop.

 

I nearly said that earlier, in fact the ouput of the rectifier could be pretty much flat DC, the resistor and cap (if after the resistor) providing decoupling of any residual ringing from the DCC signal.

[sharris]

If the voltage is too high don't forget the simple option of half wave rectification. In other words one diode in series with one of the supply leads, and forget the bridge rectifier. It may work better with a bigger capacitor, depending on how much current the LED's draw.

 

Assuming the DCC signal is symmetrical (which it should be) a single series diode will actually increase the voltage at the LEDs and increase the peak current consumption too. 

 

The reason for this:

With a full wave rectifier (FWR), you've essentially got 2 diodes in series for each direction of current flow, so a voltage drop of about 1.4V total. The DCC signal is nominally a square wave, so with a FWR the only time the output voltage will return to 0 (apart from dead track areas) is at the rise or fall of the signal, which should be very short for a well designed DCC system, so almost all of the time the voltage will be peak-voltage - 1.4, and the capacitor won't be discharging much (or charging) - the only time it will be charging significantly is after recovering from a track voltage loss. 

 

With a single diode half wave rectifier (HWR), you've got 1 diode (obviously) for one direction of current flow, so half the time (before thinking about the capacitor) the output voltage will be peak-voltage - 0.7 (i.e. 0.7V volts higher than for the FWR) for forward current or 0V for the reverse current direction. When the capacitor is in the circuit, the LEDs will be relying on it to power them during the 0V half of the cycle, causing it to discharge somewhat, and it will then charge again, drawing current, during the +ve halve of the cycle. 

Edited December 13, 2018 by sharris

[Crosland]

I generally agree with you but not this time. The load of the LEDs will prevent that. Otherwise there would be no volt drop with the resistor after the cap. The thought experiment is: leave out the capacitor, would there be a volt drop, answer yes.

 

You are correct. Don't know what I was thinking.

[DCB]

I find the strips too inflexible to allow the LEDs to be positioned in every compartment. My DC LED coach lights are wired as per post #5 mainly to minimise flickering but they don't exactly stay alive for long.  I find battery lights to be much more practical and I wonder if a charging circuit for a rechargeable battery cells might be a better solution.

[shiny]

Hi,

I tried a few experiments with coach lighting a couple of years back, including using exactly the same kind of cheap LED strips from a well known auction site.

 

The problem I found with those LED strips, apart from LEDs not lining up with the compartments, is that they tend to be wired with three LEDs in series. Once the voltage in the capacitor drops to below 6(ish) volts, it isn't enough to drive the LEDs - and the lights will dim and go out pretty quickly. You end up with 6 volts of power left in the capacitor that is going to waste because the LEDs cannot make use of it.

 

These light strips are also designed to be as bright as possible - which really is overkill for lighting up carriages. I have a Hornby MK3 TSO wired this way that I can use as a reading light.

 

That being said..

 

The approach that worked best for me was to leave the LED on the backing tape, but cut them up singly - and solder a resistor to each individual LED. Forget about the little black SMD resistor already on the tape. Use the back of a stanley knife blade to scrape away some of the paint from the LED side and this exposes enough copper track to solder connections to without melting the LED. The adhesive backing will also usually survive soldering so can still be used to mount the LEDs to the roof. A bit of surgical tape to hide/stick the extra wires and resistors down flush to the roof wouldn't go amiss though.

 

It is a bit of an ugly hack, but once the coach is sealed up it is almost invisible and quite effective. The lights fade out far more gently and stay on far longer, around 5-10 seconds, depending on how many LEDs and how bright you want them because they stay lit as long as the voltage remaining in the capcitor is above 2(ish) volts. The smoothing capacitor can then be a smaller size so it can be discreetly hidden away inside a toilet or guards compartment.

 

It is normally possible to line up the LEDs so that they sit directly above the partition between compartments, and you can light both compartments effectively using one LED. For a typical 7 compartment coach with end doors, you can usually get away with around 4 or 5 LEDs. It just depends how subtle (or not) you want the effect to be.

 

You can use resistor values between 1.5K and 3.5K for each LED, depending on the coach and the lighting effect you want. Lower values for brightly lit modern stock, and higher value resistors giving dimmer lights are probably more appropriate for older pre-80s stock.

 

I think it is worth experimenting with, and if you mess up - just try again - the parts are cheap enough. IIRC I think my LED strips worked out to just over a penny per light.

 

I ended up buying a bulk bag of capacitors and a reel with a couple of hundred bridge rectifiers from RS components to do a the rest of the stock - it still ended up pennies per coach. The rectifiers are tiny little things - but even with my ham-fisted soldering - really easy to do and definitely far less messing than using individual diodes. The bridge rectifier also allows the lights to work on DCC and DC in both directions as long as power is applied, the LEDs are sensitive enough to come on even at low speeds.

 

I would also recommend investing in some miniature plugs and sockets so that the LEDs in the roof can be separated from the circuitry in the base in case you ever need to open the coach up in the future. I learnt that one the hard way!

 

One more thing, think about putting a resistor between the rectifier and the capacitor, say around 100 ohms or so, perhaps a bit less - depending on how many lights you're running, which will limit the "inrush" current when you power up. The lights themselves don't use a lot of current on their own, but if you have a lot of coaches on the layout with empty capacitors and no way of limiting the inrush current - all of those thirsty coaches might trick the overload protection on your controller into shutting down when it is first powered on. Once the capacitors have had a chance to charge the current draw is not excessive.

 

I ended up painting the insides of some of the coaches because of light bleeding out through the roof and sometimes the body, especially older Lima stock.

 

Overall I was more than happy with using cheapo light strips - well worth the money!

 

I might have pics of some of my experiments knocking about somewhere.

 

Hope this helps.

 

 

 

 

 

 

 

 

 

[shiny]

As promised, some pics of my experiments with cheap lighting strip..

 

 

The centre coach (aka the reading lamp) has an unmodified lighting strip, way too bright. The other coaches look bright because I tried a few different resistor values out and the camera insisted on overexposing.

 

 

Started by cutting the strips into individual pieces and scraping off some of the paint covering the copper track.

 

 

Resistors added.

 

 

Wired together. There are way too many lights here - could have gotten away with half the number.

 

 

Backing peeled off the LEDs and stuck to the roof. I used fewer lights in this full brake, but it still gave good illumination. The roof of the Lima in the photograph was given a quick spray with Halfords grey primer to stop light from bleeding through.

 

 

Putting the rectifier and capacitor onto a piece of stripboard. The rectifier is placed over the two breaks in the copper conductor and then soldered into place. The capacitor is a 25 volt 1000 microfarad type. For this one I placed the inrush resistor (47-100 ohms - whatever I have laying around) in one of the wires coming from the bogie, unfortunately no photograph. The rectifiers came from RS Components in a reel of 125, very small but rated at 1 Amp or thereabouts, and far easier to solder than four diodes. I think I soldered this one by throwing molten solder at it from across the room, apparently.

 

Placed in the toilet compartment of a Railroad Gresley. A little bit of plastic was trimmed out so that it would fit snugly. I ended up painting the ends and innards of the coach because of the amount of light bleeding through the bodywork.

In the toilet compartment of an Airfix Stanier composite and wired in through a hole in the floor.

Placed above the toilet compartments in a Hornby MKII with a bit of subtle butchering where it cannot be seen. I totally forgot to put the inrush resistor in this one. Should be fine as long as I don't forget too often..

 

 

After reassembly. Light is coming out of the Airfix through the join between the roof and the body - the gap isn't obvious until the lights come on. I ended up sticking a bit of card inside the roof to hide the gap. The Lima DMU trailer has two extra diodes on the same board and some fibre optics to give directional lighting on DC, but I'll be fitting a DCC chip at some point. The guard's van is one that Hornby made with the dirty big red lens at one end, and this was given the extra diode treatment as well some fibre optics, a coat of white paint inside and some roughed up glazing from that clear stuff that comes with new shirts to hide the 5km of wire I left inside. The tail lamp is directional on DC and stays permanently lit on DCC, but I'm not particularly bothered about that as it was done more for practice with lighting methods than anything else.

 

A Lima Western after receiving a splash of Humbrol matt black inside to prevent light bleed. Fibre optics used to light up the lamps at either end and red/white LEDs attached using heatshrink tubing so that the LEDs can be (fairly) easily separated from the fibre optics when the body is removed. Fibre optics were used to avoid fouling the motor bogie, and it meant that I only needed one bi-colour LED for lighting all three lamps at each end. An LED from the cheapo lighting strip glued to each cab roof illuminates the trailing cab when the headlights come on at the other end. This one is wired for DCC using an 8 pin socket made from leftover connectors - the holes in ordinary stripboard exactly match the pins of a standard 8 pin DCC decoder. The extra connector is for the 4th (purple?) function wire. I have left space on the circuit board for a rectifier and a capacitor so that constant lighting LEDs can go inside the engine bay - once I've tidied up the mess of wires..

 

 

All three lights are powered by one red / warm white LED using 1mm and 0.5mm fibre optic cables. The trailing cab is lit up by cheap lighting strip when the light at the leading end is white so that the cab stays lit up even if the rear lamps are out - if that makes sense. I used to see the diesels taking their trains out of Lime Street in the 70s and 80s and the trailing cab lights were always on - it seemed quite common to see a BR employee sitting there doing his paperwork/tea/crossword in the rear cab.

 

The lighting has cost literally pennies per coach/loco/wagon, the parts are easy to get and to fit - the only real expense comes from replacing the plastic wheels fitted to some of the coaches and wagons with decent metal ones - around a pound per axle.

 

It's all a bit crude, but it works. Hope that this has given you some ideas.

Edited January 22, 2019 by shiny