Flashing LEDs

[Ray H]

I have fitted some flashing white LEDs to a club control panel as a warning devise to indicate when specific points are set in reverse. Each LED is wired in parallel with a relay via the SEEP PM1's inbuilt switch and powered when the point is reversed - the relay contacts set the frog polarity and determine which controller feeds a local track section based on the point's setting. The LED is wired in series with a 470R resistor. The LED/relay combination are powered from a 12v DC supply.

 

The relay operates correctly and the LED will light when it is supposed to (i.e. when the point is reversed/the relay energised) but the LED won't flash. However, a similar LED flashes if I test it across a 3v battery supply.

 

Does anyone have any idea why the LEDs might not flash on the control panel - there are four distinct LED/relay combinations associated with four different point. All are wired the same way and all have the same fault?

 

Many thanks.

[Gordon H]

It's possible that you may have damaged the LED's internal flashing circuit if you haven't included a back-emf protection diode across the relay coil as well.

Do any of the LEDs that have been fitted across a relay coil still flash properly when taken off and tried on the battery?

[Ray H]

Thanks for that suggestion.

 

I'll have to recover one of the LEDs from the club layout to try that out (and won't be there again until next Thursday evening).

 

I may try and save a bit of time by replacing one of the existing LEDs with a new one having first fitted a diode across the relay.

[cpman46]

Ray

Just an observation.

I don't know what the spec is for your flashing white LEDs but a quick look at Rapid Online lists a 5V flashing white LED which has a typical operating current of 25mA. I wonder therefore if your series resistance is a bit too high. From your 12V supply the resistor needs to drop the excess voltage at the typical LED operating current i.e. (12-5)/0.025= 280 Ohms.

 

The nearest preferred value is 270 Ohms and a normal 0.25W device should be more than adequate in terms of dissipation (I*I*R = 0.025*0.025*270 = 170mW). I agree with Gordon's observation that you should include a back-emf protection diode across the relay coil, a 1N4001 should be more than adequate.

 

Regards

Mike C

[Ray H]

Mike

 

The LEDs are reported to have an operating voltage of 3.2 - 3.8v and a typical forward current of 20mA (max 30mA) giving a resistor of around 390Ω.

 

I can try altering the resistor size as well and seeing what happens.

[cpman46]

Ray

 

Thank you for your reply and LED info. Looking at the operating voltage spread and typical forward current you could try dropping the resistor value down to 330 Ohms and see what happens. The resistor dissipation will still be less than 250mW.

[Suzie]

You cannot use resistors with this type of LED. When the light is on it will use 20mA or so and the resistor will drop the voltage, but as soon as the light goes off there is no current draw and the full 12V will be across the built in electronics which will fry it if it is not a 12V version.

 

These LEDs with integrated electronics have to be run from a fixed voltage (usually 12V or 5V) unlike a normal LED which runs from a fixed current.

 

If you need to run one of theses LEDs from a higher voltage than the rated voltage, just put a Zener diode across and use a series resistor following the normal LED calculations. This will allow you to use a 5V flashing LED on 12V for example with a 5v1 BZY88C Zener and 470R series resistor. You cannot reliably adjust the resistor value to adjust the brightness of the LED.

[cpman46]

Ray

 

Suzie is quite right! If your LEDs are specced at 3.2 to 3.8V and 20mA you could use a 3.6V Zener (Rapid Part No: 47-3006) together with a 390R resistor which should give about 21mA through the LED when on. The same current will flow thru the zener when the LED is off clamping the voltage across it to 3.6V. The zener will be dissipating about 78mW when conducting the full current.

[Ray H]

I'm getting confused now - not hard I know.

 

I presume that the reference to the light being off means between the flashes, is that correct?

 

Here I show my ignorance (also not hard). I have the resistor in line with the LED (because that's how I was advised to wire it) and both in parallel with the relay. Do I wire the zener in parallel with the LED - I never have really understood how a zener works or which way round to fit it (although I believe that it will only conduct when the voltage across it exceeds its stated voltage)?

 

Because the LEDs are staying illuminated (as opposed to not illuminating at all) are they now unlikely to flash again or do you think they'll start flashing when the zener is fitted?

 

I'm most grateful for all the help. Basic DC I can manage, electronics have always been a black art to me.

[cpman46]

Ray

 

Light OFF between flashes - Yes.

 

Fit the zener as in the attached diagram, band to the +12V. This should ensure that the LED never sees more than 3.6V across it when its not taking current i.e. off.

 

 

I think it likely that the LEDs flashing circuits will have been damaged and therefore will not flash but you could fit the zener across one and try it just to see.

 

If they don't work then the only option is to replace them but try out one first with the zener in place before doing the rest just to confirm correct operation. Make sure the back emf protection diode is fitted as well.

 

Let us know how you get on

 

Kind regards

Mike

[Ray H]

You Sir are a real gent.

 

Not sure how soon I'll be able to do the test but if it isn't later today then it could well be in over a week's time.

 

Thanks once again to you all for your assistance in trying to resolve this.

[Ray H]

I managed to sneak out this afternoon, retrieve one of the LEDs from the control panel and bring it home.

 

I wired the components in acordance with Mike's diagram, connected it to a 12v supply and hey presto, the LED flashed. I'm a little loathe the try any other combinations for fear of destroying to LED.

 

Many thanks once again for all the help.

[cpman46]

Ray

That's good news! Stick with the combination that works! Just as an aside I managed to find a 5V flashing LED and have tried it out with a 4.7V zener and 470R resistor as per my diagram and it too works (as it should!). I guess this should put your problems to bed!

 

Regards

Mike

[Ray H]

I do have one further query associated with this.

 

I can fit the diode across the relay without trouble - the relays have been fitted with ease of exchange in mind with all their pin-outs connected to a choc bloc. I shall simply connect the diode to the appropriate choc bloc terminals.

 

As originally fitted the LEDs had their legs trimmed and either a wire (anode) or the resistor (cathode) soldered thereto (and a wire then soldered to the resistor, which had also had its legs shortened). All exposed metal is thus covered with heat shrink.

 

I now have to fit the zener and would welcome suggestions as to how I could best achieve this within a fairly confined space (because of surrounding wires to other parts of the control panel) and still cover as much exposed metal as possible without buying yards of heatshrink - I'm thinking of the cathode LED leg and the connection between it, the zener and the resistor in particular.

 

There are several of these LEDs on the panel. Could I connect the cathode of each LED and the anode of each zener to a small bus wire that is then connected to 0v via a single 390Ω resistor? I could then replace the existing 470Ω resistor connected between each LED cathode and 0v with a length of covered wire connected to this (new) bus - I can position the bus well away from other exposed metal.

 

Or am I being a bit paranoid about exposed wire?

[beeman]

Not read all the posts but originally fitted red/green led's to each direction on my turnouts, changing with turnout direction. Eventually thought this was overkill and confusing took the reds out. not really needed, why make it complicated, if it aint Green you cant Go. Beeman.

[cpman46]

Ray

I am assuming the SEEP point motor switch is in the 0V line to the relay and LED combination i.e the 12V supply is connected permanently to the other side of the relay coil and LED? If this is so you cannot use a single resistor for all the LEDs. You will need one resistor and one zener per LED as shown in the attached diagram.

 

 

As for fitting the zener I think you will have to carefully remove the heatshrink where necessary and solder the zener in. If you need to insulate perhaps you could use small pieces of insulating tape where necessary.

 

Having just seen Beeman's post whilst composing this reply I have to say that I completed a control panel for a client doing exactly as per his comment. Green LEDs set into the mimic and controlled by relay or switch contacts. The route set was shown by the illuminated LEDs.

 

however if you wish to retain your red flashing LEDs go with the comments / diagram above.

 

Kind regards

Mike

[Ray H]

Mike

 

Our circuits are basically the same. I feed the 12v supply through the switch on the PM1 to the relay and the LED. The other side of the relay is connected to 0v and the relay will have the reversed diode fitted across it.

 

Currently the LED is connected to one side of the 470Ω resistor and the other side of the resistor is connected to 0v. I shall be replacing the 470Ω resistor with a 390Ω resistor and wiring the zener in parallel with the LED.

 

The only difference between our two circuits is that I have the switch before the components and you have the switch afterwards.

 

Would I be right in assuming that if I used a common (single) resistor for all the LEDs, as soon as more than one LED is illuminated, the effect of the resistor (in terms of reducing the voltage across a single LED) would be halved?

 

I'm using a single flashing LED to highlight when specific points are reversed. The points are controlled by an electric pencil and the contacts for this are positioned on the control panel where the LED would ideally go. The reason for a flashing LED is that one of the operators is dyslexic and has trouble interpretting steady lights/colours and because it is located slightly away from the crossover on the diagram.

[beeman]

My mimic led's are switched by the frog rails becoming live depending on the turnout direction set, so of course uses the DCC track power, generally no need for relays or switches. Must admit sometimes if the turnouts have not operated for a while they need to be operated a few times to 'clean' the switch blade contacts. Similarly some of the layout is interlocked using these to avoid any collisions where out of sight. Beeman

[cpman46]

Ray

 

Switching the positive rail is a you say just the same as my circuit switching the negative (0V). Operationally there is no difference.

 

As long as only one Relay/LED combination is energised there is no problem with having one common resistor. The problem arises, as you surmise, when more than one Relay/LED combination is energised. As both now take current the voltage drop across the resistor increases hence starving both of sufficient voltage to work properly.

 

The only way to ensure proper operation of each combination is to use individual resistors for each. This ensures that however many are energised at the same time each will have the proper current limit via its own resistor.

 

Mike

[Ray H]

Thanks Mike.

 

I'll work on that basis - good job I just stocked up on some more resistors.

 

I suppose that I don't need to worry too much about the exposed conenctions as they're on the underside of the panel and there's no need for anyone to go near them once we've got everything working.

[Ray H]

Latest update:

 

I took the previously tested set up to the club last evening and apart from using a different relay (of the same type/model) and a different 1N001 diode, I re-installed the LED, fitted the zener that I'd used for the test and did likewise with the 390Ω resistor. I changed the point to power up the relay, the LED illuminated and stayed that way i.e. it didn't flash!

 

I tried numerous other LEDs and zeners and even connected them direct to the power source i.e. ignoring the relay. The result was the same; flash the LED would not.

 

Then I removed all the extraneous components so that I was left with just the LED and I connected this to the variable output from the power source - our 12v DC supply is being obtained from a Gaugemaster controller that, like the old H & M controllers, has a variable output, a fixed 12v DC output and a fixed 16v AC output. I turned the variable DC output down to about 3v (measured on an analogue meter) and connected the LED. Still it wouldn't flash. I tried another LED with the same result.

 

Time for tea!

 

Out of the toolbox came a couple of AAA cells which I managed to arrange in series and connect the LED to them. Geronimo! a flashing LED (and not just one, thankfully).

 

And then a brainwave (?). I'd been using a kit-built bench power supply for the tests at home.

 

I'd recently added an extra 2200µF electrolytic capacitor to the CDU so that it could drive a few more points without grumbling. I removed said capacitor and placed it across the fixed 12v supply, re-assembled the test circuit - LED, resistor and zenor - and powered it up. The LED flashed. Out of interest and with the capacitor still in situ I changed the other points that have relays/flashing LEDs (indirectly) attached and they all flashed - yes, all the ones where I haven't altered the circuitry!

 

And so the problem all along seems to have been that we were using a DC supply that is smooth enough for ordinary LEDs but not for the flashing LEDs.

 

The 2200µF electrolytic capacitor has now been restored to the CDU and the flashing has (not surprisingly) stopped again!

 

The question is, what size capacitor do I need to use to smooth the supply? Can I get away with a few pFs or do I need something like an 1000µF electrolytic capacitor or bigger?

 

Although I have the components to change the circuitry of all the (flashing) LEDs I plan to leave the test one in situ for the time being (having added the capacitor across the supply) and do likewise to the unmodified circuits, which I shall only change if the original circuitry fails.

 

I thought that you'd be interested to learn the outcome.

[cpman46]

Ray

Very interesting!

 

I can't say for certain but it would appear you may have been powering the LEDs from an unsmoothed half wave (or full wave) rectified ac supply that is notionally 12V dc. Adding the 2200uF capacitor would smooth out the rectified ac and hence allow the LED to flash. I suspect the integrated circuit in the LED needs a fairly smooth supply as you have found out.

 

I don't know how many circuits your 12V supply feeds worst case but I would suggest a 470uF 35V capacitor would be adequate to smooth the supply. The normal LEDs and relay coils should not be worried by this extra smoothing.

 

It would be interesting to measure the volts using your meter before and after fitting the capacitor to see what the difference is. I suspect the dc reading will be somewhat higher.

 

Regards