Anti-Flicker Directional Lights on DC

[delticman]

I am currently respraying, weathering and detailing a Hornby bubble car Class 121 for a friend, part of the detailing involves fitting directional lights which will show white front marker lights and red tail lights according to the direction of travel and this is conventional DC operation. So far so good, now I would like to add capacitors to help alleviate light flicker and this is where I need help. I searched RMweb to no avail, plenty of advice on anti flicker constant coach lighting using a bridge rectifier for constant polarity and directional lights on DCC.

By my reckoning, as the capacitor is polarity sensitive, I would need one for each direction along with a diode to prevent the wrong polarity reaching the capacitor, is this correct?  I am appealing to the more knowledgable members on here for help in this matter so I have included a rough circuit diagram of my intentions and for clarity have only shown one end of the unit. I intend using 12volt bi-colour diodes to show either white or red in the marker lights.

[34theletterbetweenB&D]

Your circuit immediately discharges through the storage capacitors whenever the applied power is reduced or absent.

 

I suspect the easiest way is to use a DCC decoder that can have a small stay alive added, dedicated to running the lights; could be built from a design or kit such as those from MERG if you really want to do the soldering.

[Dave John]

Might be simpler to play about with non polarised electrolytic capacitors for straight DC operation.  The motor would always swamp the LEDs with regard to discharging the capacitor, but if its just to reduce flicker caused by very short ( fractions of a second) loss of supply it might help a bit. Try it and see how it looks to the eye. 

 

You could start getting sophisticated by using a bridge to charge a capacitor (thus losing the motor as a load) then detecting polarity and switching the leds in but thats getting a lot more complex. 

 

[Dagworth]

Scrap the two diodes, put the two capacitors series wired but anode to anode (or cathode to cathode) and then in parallel with the LEDs, but put another resistor between the power feeds and the whole set-up to prevent the discharge via the motor. 
 

Andi

[delticman]

Thank you so much for the replies guys, just to reiterate this is for a conventional DC system, so 34theletter I cant use a DCC decoder (it would be so much easier if it were DCC). Dave John I never knew that there was such a thing as non polarised electrolytic capacitors, this is how little I know! 

Dagworth (love your full length APT by the way), your solution intrigues me (I cant get my head round the capacitors being like pole to like pole but that is probably just my ignorance!), I think I will give this a go, I wonder if you could scribble a rough circuit diagram for me? What value resistor do I require for between the power feeds?

 

Geoff

[Dagworth]

The two capacitors back-to-back act as a single non-polarised capacitor. 
 

 

 

Thanks for the comment re the APT  can’t wait for my Hornby one to arrive! 
 

Andi

Edited July 9, 2020 by Dagworth

[delticman]

Thank you for the circuit diagram, it is much appreciated and I will give it a go when my leds arrive. I dont really know how connecting the caps that way round will affect the total capacitance, I know that connecting them in parallel you just add them together and in series there is a formula to work it out.

I will let you know when I have tried this out.

 

Geoff.

 

P.S. I dont think I can stretch to one of the new Hornby APTs  but judging by the engineering samples they look the business! I have an original 5-car set and a seperate power car that I have de-motored and fitted Kadees to for the APT Test Train that is next on my "to do" list.

[Dagworth]

From what I can find online it seems that the capacitance of a pair of inverse series wired caps (and the two caps must be identical) is half the capacitance of one of them.

 

Hope that helps.

 

Andi

[DCB]

You will need a resistor in series with the red LED (s)  Generally White's won't light on 3.6 volts and Reds smoke and die on 3.6 volts so there is no sweet spot where both will work off the same resistor.

I use electrolytics with bridge rectifiers, either off the shelf or just 4 diodes soldered up, for LED coach lighting on DC which is pleasantly flicker free but its not directional and only works well between about a scale 20 and 50 mph.

[Dagworth]

Thanks David, my personal preference is individual resistors for every LED, keeps things very simple. I had only stuck with the single shared resistor to keep some continuity with the OPs circuit diagram

 

Andi

[delticman]

David, re the resistor in line with the red LED, I intend using 12 volt bi-colour red/white LEDs as this is for a Class 121 where the marker lights and tail lights share the same lamp housing, so I wouldnt be able to add a resistor to just the red. My in line resistor in the original circuit diagram was to knock back the light intensity to a more prototypical "candle" output lol.

Andi, if the total capacitance works out at half of one of the caps then I am stuffed, I was using 1000uf caps as the bare minimum due to space limitations under the chassis, I think halving that value would have no effect on the flicker 

 

Geoff

[Dagworth]

With the resistor in the feed line to the caps then 1000uf caps should have a significant and noticeable effect. 
 

try it and see?

 

what make of bubblecar is it? 

 

Andi

[delticman]

The bubble car is the Hornby one (ex Lima with a neater little power bogie) and has no lighting as supplied. I have one that I fitted directional lighting to and resprayed as Route Learning Car DB975042 but mine is DCC so no problem. I will certainly try out the set up and let you know how I get on.

Thanks Andi for taking the time to help, much appreciated.

 

Geoff

[Dagworth]

57 minutes ago, delticman said:

The bubble car is the Hornby one (ex Lima with a neater little power bogie)

Good, that one should have reasonable motor suppression built in that will solve some of the light flicker issues anyway. There was a big problem with Heljan 58s when first released where a pair of capacitors had been fitted to help with light flicker but the feed line resistor wasn't fitted and with the big motor that Heljan use any dirt caused back-emf from the motor to light the wrong lights!

 

Going back to the capacitance of the inverse series caps, adding a pair of diodes in parallel with the caps will actually increase the capacity of the arrangement back up to the full amount of one cap.

 

 

Andi

Edited July 10, 2020 by Dagworth

[John ks]

19 hours ago, delticman said:

David, re the resistor in line with the red LED, I intend using 12 volt bi-colour red/white LEDs as this is for a Class 121 where the marker lights and tail lights share the same lamp housing, so I wouldnt be able to add a resistor to just the red

If you alter Dagworth's circuit as below then The value of R2 will affect the brightness of the White LED & R3 will affect the red LED

 

 

John

[Suzie]

You can use a DCC decoder, most have the capability to work on DC and will operate the directional lighting, and make use of the stay alive. The better the decoder the better it will work on DC so you really need to spend at least £20 and get a Zimo (MX600p12) or similar.

 

If you fancy having a go at making your own the DIY Decoder Project function decoder will do the job well and you should be able to make it fit in the 121 and fit a small capacitor after the rectifier to give stay alive capability. The directional lighting will then light correctly at full brightness as long as there is a small voltage on the track.

 

http://dccdiy.org.uk/function.html

 

In practice it is hard to get good results for directional lighting with stay alive on DC without using a computer to control the lighting so using a small PIC like the 12F629 used in the DIY decoder is a good way to do it, and operating the lighting on a lower voltage such as 5V means it can take a while for the capacitor to discharge before the voltage drops low enough to dim the lighting.

 

 

[delticman]

Wow, thanks guys for all your interest and replies. Suzie I had totally forgotten that decoders can work on DC (I always disable DC running on mine as a matter of course), wasnt there some issue though with motors buzzing on DC?  I have spare decoders in my spares box although not expensive ones. I cant go to the expense of investing in a quality decoder and a stay-alive.

Dagworth thank you for the updated circuit diagram and along with Davids solution for control of the differing brightness of the red and white LEDs could well be a winner though a bit more soldering is involved!

When the LEDs arrive I will try the Dagworth/John circuit and also try Suzies suggestion of a decoder.

Many thanks to all of you for your expertise.

 

Geoff

Edited July 11, 2020 by delticman

[justin1985]

I'm trying to do something similar in a much smaller space (a Z gauge driving coach) - this thread was really useful.

 

But what is the minimum capacitance that would have a noticeable effect in keeping LEDs on?

 

Is there a formula for working out how long a certain capacitance could power an LED (or whatever) at a certain voltage and current?

 

I searched for non-polarised capacitors, but everything I could find at e.g. Farnell was either really really low in capacitance (picofarads) or really big! 

 

Realistically in the space I have available, I could only accommodate something in the kind of format that tantalum capacitors come in, or possibly a single small surface mount cylinder type (the silver ones).

 

Can tantalum capacitors be used in the "back to back" circuit illustrated above? I know they have a reputation for blowing up if mistreated!

 

J

[Il Grifone]

I've come a bit late to this....

The red and white LEDs should have separate feeds and not be connected together. A single car will need two sets - red and white at each end. Unfortunately, AFAIK bicolour LEDs do not come in white/red, so some sort of fibre optics will probably be involved.

The formula for calculating the capacitor discharge time is Q=CV, but it is complicated by the voltage falling as the capacitor discharges so heavy maths is involved.... Something like 1000μF is necessary, but see below. I don't see why tantalum capacitors could not be connected back to back (protected with diodes to be safe), but our practice when I worked in electronics service was to replace faulty ones (common) with the normal type.

 

 To avoid undesirable effects like lights going off when the vehicle stops, back EMF causing the wrong lights to come on etc. I intend to use a small disc battery to power the LED lights* and control their function with transistor switches, these being held on by the capacitor(s). This will allow the lights to come on when voltage is first applied before the motor has sufficient voltage to start and stay on when the train stops again. Reversal of polarity will then switch the directional lighting while the train is stopped. I have ideas about using the couplings** to detect additional vehicles and switch off unwanted head and tail lights***.

 

* CR2025 etc. available cheaply from 'Poundland' (usual disclaimer). I've a couple of mini-torches which illuminate white LEDS without the benefit of series resistors. They were bought several years ago, when white LEDs were expensive as  a (relatively) economical source of the things (£1 each IIRC).

** These will need to be conductive, but decent couplings only come in metal so it's not a problem.

*** For example American trains have headlights on each locomotive and a habit of running power units in multiple, luckily there is plenty of room inside an F7A!

 

Sorry about any typos, but I had 25kg of Maremano puppy jump on my lap while I was typing!

 

[smokebox]

2 hours ago, Il Grifone said:

I've come a bit late to this....

The red and white LEDs should have separate feeds and not be connected together. A single car will need two sets - red and white at each end. Unfortunately, AFAIK bicolour LEDs do not come in white/red, so some sort of fibre optics will probably be involved.

 

 

 

Red & white bicolour LEDs are available from many sources.

[Dagworth]

7 hours ago, smokebox said:

 

Red & white bicolour LEDs are available from many sources.

https://www.expressmodels.co.uk/catalogue/the-led-store/3mm-bi-colour-red-white-led-for-dcc-common-pos-1-9-leds-detail

 

https://www.railwayscenics.com/electrical-and-electronics-lighting-and-leds-3mm-led-lights-3mm-21v-bi-colour-red-white-led-resistor-required-p-993.html

 

https://www.rapidonline.com/truopto-osrwpa3132a-3mm-red-white-bi-colour-led-common-anode-3-pin-diffused-76-0017

 

https://www.dcpexpress.com/led19-colour-redwhite-leds-resistors-pack-p-699.html

 

Andi

[Il Grifone]

Thanks!

I'm obviously behind the times!

[delticman]

Just as a belated update on this matter as I originally asked about the anti-flicker circuit for LED directional lighting on conventional DC operation. I had to respray a Hornby Class 121 bubble car from chocolate and cream to banger blue for a friends DC layout. He had asked me to install directional lighting so I had asked about the anti-flicker circuit as an added luxury really. After assembling all the required components I found very little room in the unit to install them without something being visible in the saloon area of the unit. One of the capacitors was able to masquerade as the cylindrical reservoir underneath the chassis (it was almost identical dimensionally). 

So, frustrated, I thought I would try the lighting with only the current limiting resistors and diodes to prevent reverse voltages and amazingly it all worked really well with no flickering.....problem solved!

My friend was so happy with the finished model that he sent me a short video of it running on his layout over dead frog pointwork with absolutely no light flicker so I have included it here.

May I just add my thanks for everyone who contributed with their great knowledge in particular Andi ( Dagworth). The circuit is now filed and may well be called upon in the future.

Thank you,

Geoff.