SMD LED lights dim but don't go off

[ISW]

I'm retro fitting DCC and lights to an old Mainline Class 45, and am experiencing problems with the lights (F0 headlights). They work, but don't go 'off' properly. They are still lit dimly - and enough to be annoying.

 

I'm using 3v 0805 SMDs with a 1500 ohm resister to avoid blowing the LEDs (done that already!). They are wired to the Blue / Yellow & Blue / White DCC cables to give directional working.

 

The chip is a Hattons 8-pin harnessed version.

 

I checked the voltage on the yellow & white circuits and get ~13v DC when 'on' and ~3v DC when 'off'. Is this normal, and how do you avoid the LEDs being dimly lit at the 'off' voltage?

 

Any help appreciated!

 

Ian

[Lady_Ava_Hay]

I'm retro fitting DCC and lights to an old Mainline Class 45, and am experiencing problems with the lights (F0 headlights). They work, but don't go 'off' properly. They are still lit dimly - and enough to be annoying.

 

I'm using 3v 0805 SMDs with a 1500 ohm resister to avoid blowing the LEDs (done that already!). They are wired to the Blue / Yellow & Blue / White DCC cables to give directional working.

 

The chip is a Hattons 8-pin harnessed version.

 

I checked the voltage on the yellow & white circuits and get ~13v DC when 'on' and ~3v DC when 'off'. Is this normal, and how do you avoid the LEDs being dimly lit at the 'off' voltage?

 

Any help appreciated!

 

Ian

 

I might be wrong to suggest this but I do recall that the blue wire on the decoder is positive so yellow and white will be the return negative wire. I think you can also use the red ( rh side when loco going forward ) pick up wire as a substitute for the blue and that gives half the voltage. I do believe this difficulty with wiring actually persuaded the RTR industry to use the 21 pin connection and a small circuit board to mount all the resistors and return wires.

[Nigelcliffe]

 

I'm using 3v 0805 SMDs with a 1500 ohm resister to avoid blowing the LEDs (done that already!). They are wired to the Blue / Yellow & Blue / White DCC cables to give directional working.

 

 

 

You use the word "resistor" which is singular.   Have you fitted one common resistor, or one resistor for each LED ?    If its a common single resistor, that may be the cause of your problem.   

 

Also, 1500ohm is fairly small for a SMD LED, I'd expect a higher value to be more appropriate.  

 

- Nigel

[ISW]

I might be wrong to suggest this but I do recall that the blue wire on the decoder is positive so yellow and white will be the return negative wire. I think you can also use the red ( rh side when loco going forward ) pick up wire as a substitute for the blue and that gives half the voltage. I do believe this difficulty with wiring actually persuaded the RTR industry to use the 21 pin connection and a small circuit board to mount all the resistors and return wires.

 

I may not have made myself clear (easily done ...). The LEDs are wired to the Blue(+ve) + Yellow(-ve) and Blue(+ve) + White(-ve) pairs. 

Not heard the 'trick' of using the Red in place of Blue. Is this a common idea?

 

You use the word "resistor" which is singular.   Have you fitted one common resistor, or one resistor for each LED ?    If its a common single resistor, that may be the cause of your problem.   

 

Also, 1500ohm is fairly small for a SMD LED, I'd expect a higher value to be more appropriate.  

 

- Nigel

 

I did try the single 1500 ohm resistor on the Blue, but changed to having the same resistor on each of the Yellow & White wires at both ends of the Class 45 (ie: 4x resistors). At least I appear to have got that bit right.

 

What would be a 'common' value for the resistor with SMD LEDs? If it's a multiple of 1500 then I can double / triple up the resistors.

 

However, I still don't see why / how the chip is giving 'any' voltage when the circuit is meant to be 'off'.

 

Ian

[Dagworth]

I may not have made myself clear (easily done ...). The LEDs are wired to the Blue(+ve) + Yellow(-ve) and Blue(+ve) + White(-ve) pairs. 

Not heard the 'trick' of using the Red in place of Blue. Is this a common idea?

 

 

You can use either rail feed as return(red or black) not just the red wire. 6pin decoders for N gauge don't use a blue positive feed at all, they return to a running rail.

 

No idea why you should be getting a voltage when the function is off though.

 

Andi 

[Huw Griffiths]

I'm retro fitting DCC and lights to an old Mainline Class 45, and am experiencing problems with the lights (F0 headlights). They work, but don't go 'off' properly. They are still lit dimly - and enough to be annoying.

 

I'm using 3v 0805 SMDs with a 1500 ohm resister to avoid blowing the LEDs (done that already!). They are wired to the Blue / Yellow & Blue / White DCC cables to give directional working.

 

... I checked the voltage on the yellow & white circuits and get ~13v DC when 'on' and ~3v DC when 'off'. Is this normal, and how do you avoid the LEDs being dimly lit at the 'off' voltage?

 

I'll admit that I've never used DCC on any of my models (probably not much point, when most of my modelmaking actually involves part-models - to test ideas and learn techniques before reusing the parts on something else).

 

However, what jumps out at me is the "~3V when off" - because a lot of single colour LEDs will conduct and light at this sort of voltage. As far as I'm aware, the resistor(s) aren't there to reduce the voltage across the LEDs, but to reduce the current through them to a safe value.

 

It might be interesting to measure the voltage across the LEDs - both when you're expecting them to be ON and when you're expecting them to be OFF - I suspect that, in both cases, you'll probably get somewhere reasonably close to 3V across them.

 

This might sound strange - it would have to me, were it not for some vague memories from my electrical & electronic engineering HND course back in the 1980s.

 

Like a lot of diodes, LEDs are based around semiconductors - which only really start to conduct when the voltage across them is close to their quoted forward voltage. If you increase the voltage across them, the current they'll allow to flow through them shoots up (if unchecked, to the point where it could destroy them) - this is why you add series resistors to ensure that this current doesn't exceed the maximum forward current quoted on manufacturers' data sheets.

 

 

There is a lot of theory available from various sources (some of them reputable!) - but I suspect you're probably more interested in what you can do here.

 

 

I may not have made myself clear (easily done ...). The LEDs are wired to the Blue(+ve) + Yellow(-ve) and Blue(+ve) + White(-ve) pairs. 

Not heard the 'trick' of using the Red in place of Blue. Is this a common idea?

 

 

I did try the single 1500 ohm resistor on the Blue, but changed to having the same resistor on each of the Yellow & White wires at both ends of the Class 45 (ie: 4x resistors). At least I appear to have got that bit right.

 

What would be a 'common' value for the resistor with SMD LEDs? If it's a multiple of 1500 then I can double / triple up the resistors.

 

However, I still don't see why / how the chip is giving 'any' voltage when the circuit is meant to be 'off'.

 

Would I be correct in assuming that there are either 2 or 3 headlights (lit as a group) - and 2 tail lights (also lit as a group) - at either end?

 

If this is the case, I'd be tempted to wire all the headlights at one end as a group - all the tail lights at the other end also as a group - and connect them, via series resistors (one per group) to the blue and either the white or wire.

 

I'd then be to do something similar with the remaining groups of head and tail lights - but connect these (and their series resistors) between the blue and yellow wires.

 

With any luck, only the lights at the correct ends would conduct and light. However, it might also be possible to connect a diode in series with each group of LEDs, to further limit the voltage across the LEDs and reduce the chances of them conducting and lighting. I suspect that some of the people here would know a lot better than me whether doing this would help or hinder.

 

 

Now let's turn to the value of the series resistors - the exact values aren't critical - and you certainly don't need to stick to multiples of any specific value. To be honest, 1K5 (or 1500 Ohm) doesn't sound ridiculous.

 

However, unless resistors of this value had been supplied with the LEDs, I would probably have experimented with different values until I was happy with the LED brightness (most likely 1K, 2K2, 4K7 and 10K - or anything in a similar "ballpark" that I happened to have lying around). Actually, it's quite possible that I might have eventually settled on 1K5 - but I doubt if this would have been the first value I tried.

 

 

Best of luck with your project,

 

Huw.

[ISW]

 

However, what jumps out at me is the "~3V when off" - because a lot of single colour LEDs will conduct and light at this sort of voltage. As far as I'm aware, the resistor(s) aren't there to reduce the voltage across the LEDs, but to reduce the current through them to a safe value.

 

It might be interesting to measure the voltage across the LEDs - both when you're expecting them to be ON and when you're expecting them to be OFF - I suspect that, in both cases, you'll probably get somewhere reasonably close to 3V across them.

 

This might sound strange - it would have to me, were it not for some vague memories from my electrical & electronic engineering HND course back in the 1980s.

 

Like a lot of diodes, LEDs are based around semiconductors - which only really start to conduct when the voltage across them is close to their quoted forward voltage. If you increase the voltage across them, the current they'll allow to flow through them shoots up (if unchecked, to the point where it could destroy them) - this is why you add series resistors to ensure that this current doesn't exceed the maximum forward current quoted on manufacturers' data sheets.

 

Would I be correct in assuming that there are either 2 or 3 headlights (lit as a group) - and 2 tail lights (also lit as a group) - at either end?

 

If this is the case, I'd be tempted to wire all the headlights at one end as a group - all the tail lights at the other end also as a group - and connect them, via series resistors (one per group) to the blue and either the white or wire.

 

I'd then be to do something similar with the remaining groups of head and tail lights - but connect these (and their series resistors) between the blue and yellow wires.

 

With any luck, only the lights at the correct ends would conduct and light. However, it might also be possible to connect a diode in series with each group of LEDs, to further limit the voltage across the LEDs and reduce the chances of them conducting and lighting. I suspect that some of the people here would know a lot better than me whether doing this would help or hinder.

 

 

Now let's turn to the value of the series resistors - the exact values aren't critical - and you certainly don't need to stick to multiples of any specific value. To be honest, 1K5 (or 1500 Ohm) doesn't sound ridiculous.

 

However, unless resistors of this value had been supplied with the LEDs, I would probably have experimented with different values until I was happy with the LED brightness (most likely 1K, 2K2, 4K7 and 10K - or anything in a similar "ballpark" that I happened to have lying around). Actually, it's quite possible that I might have eventually settled on 1K5 - but I doubt if this would have been the first value I tried.

 

 

Best of luck with your project,

 

Huw.

 

Huw,

 

Thanks for the above.

1. I think the idea of an additional diode in series with the SMDs is a good one. Maybe that will a; limit the voltage even more and b; cut the voltage off below its threshold.
2. I'll have to find out what the forward limiting voltage of a typical 0805 SMD actually is below which it does not light.
3. I already have each end of the loco as 2x red in parallel (on blue / white) and 1x white(on blue / yellow) with the 1k5 resistor in series on each of white & yellow (obviously the other end of the loco has the white / yellow reversed, but otherwise identical). So quite why this results in 'dimmed' and not 'off' LEDs is causing me the trouble.

Once I've tried the additional diodes to have 'off' LEDs, I will then 'fiddle' with the resistor value until I get the correct brightness.

 

Might take a while though ... 

 

The strange thing is that I had some 0603 SMDs with inline 1k5 from Layouts4U, and they didn't exhibit this problem (with the same DCC chip type). It was the cost (~£1 each) that took me to buy 0805 SMDs (~£1.50 for 5). Am I missing something?

 

Ian

[Dagworth]

Other option is to put a resistor in parallel with the LEDs as well as the series current limiting resistor, this may act to shunt the power away that is causing the LED to glow when it should be off, but high enough value that it will still permit current through the LED when it should be on

 

Andi

[Salmotrutta]

Digitrax decoders can be set to illuminate the "tail light" at the same time as the head lights. I don't have a Hattons decoder to see if they also have hat function but if you reset the decoder (CV8 usually) that will stop that happening if it is the problem. I had a similar issue to you using bi-colour LEDs with an ESU decoder. However the same set-up with the same model of decoder in another loco did not do it. So it might just be something to do with that particular decoder behaving strangely. 

[Huw Griffiths]

Other option is to put a resistor in parallel with the LEDs as well as the series current limiting resistor, this may act to shunt the power away that is causing the LED to glow when it should be off, but high enough value that it will still permit current through the LED when it should be on.

 

I seem to remember something vaguely similar being done with the inputs on some CMOS logic chips (I think it might have been to ensure they defaulted to a known state when there was no input, or something like that) - if I recall correctly, the resistors used for this tended to have high values (megohms) but the internal resistance of the logic chips was a lot higher.

 

Huw.

[Dagworth]

I seem to remember something vaguely similar being done with the inputs on some CMOS logic chips (I think it might have been to ensure they defaulted to a known state when there was no input, or something like that) - if I recall correctly, the resistors used for this tended to have high values (megohms) but the internal resistance of the logic chips was a lot higher.

 

Huw.

Are you sure you aren't thinking of standard 10k pull-up resistors on TTL inputs?

 

Andi

[Lady_Ava_Hay]

 

Not heard the 'trick' of using the Red in place of Blue. Is this a common idea?

 

 

 

 

Just a memory. I personally have never used it. Lights were a rarity back then!

[Huw Griffiths]

Are you sure you aren't thinking of standard 10k pull-up resistors on TTL inputs?

 

Andi

 

Quite possible - I haven't had much sleep recently, so my memory could well be playing tricks.

 

Anyway, that's more than enough about me.

 

Back to the trains.

 

 

Huw.

[John ks]

While this wont solve ISW’s problem

 

With regard to where the blue (+) from the lights is connected

 

1:- you can connect it to the blue connection on the decoder

 

2:- you can connect to either rail to get half wave DC for your light

 

3:- you can connect via 2 diodes to get full wave DC to your lights (small diodes in 3^rd diagram )

 

The diagrams show a very simplified decoder

 

The last drawing shows the current path for ½ wave DC

 

 

Back to IWS’s problem

I am fairly sure that there is a transistor (or similar) between  the PIC chip  & the function output.

 

If this is not switching off completely then the leakage current might be sufficient to dimly light a LED

 

Have you tried a different decoder in this loco?

Have you tried the decoder in a different loco?

If the problem follows the decoder then it is a decoder problem

If the problem stays with the loco then the is something in the loco wiring

 

John

[RAF96]

The use of the rails as a positive is common to 6-pin decoders and as stated can have a diode or two or voltage regulator in there to control the DCC wobbly voltage.

 

Brian Lambert’s site shows colour coded schematics for all variant pin decoders if you need to see how they are wired.

Rob

[ISW]

 

Have you tried a different decoder in this loco?

Have you tried the decoder in a different loco?

If the problem follows the decoder then it is a decoder problem

If the problem stays with the loco then the is something in the loco wiring

 

John

 

John,

 

Sadly I only have Hattons 8-pin harness chips at present (I do have a Hornby one, but it's hard-wired into the insides of an Bachmann 00-gauge Class 08!).

 

I did try a few Hattons chips and got the same problem. 

 

As noted previously, this problem did not happen with 0603 SMDs from Layouts4U (with in-series 1k5 resistors) on the same Hattons Chips.

 

Ian

[ISW]

The use of the rails as a positive is common to 6-pin decoders and as stated can have a diode or two or voltage regulator in there to control the DCC wobbly voltage.

 

Brian Lambert’s site shows colour coded schematics for all variant pin decoders if you need to see how they are wired.

Rob

Rob,

 

I did take a look through Brians webpage (found from a Google search ...) and that's what convinced me to put 1k5 resistors on each leg of the White / Yellow instead of a single 1k5 on the Blue. Didn't seem to help though ...

 

Ian

[ISW]

Digitrax decoders can be set to illuminate the "tail light" at the same time as the head lights. I don't have a Hattons decoder to see if they also have hat function but if you reset the decoder (CV8 usually) that will stop that happening if it is the problem. I had a similar issue to you using bi-colour LEDs with an ESU decoder. However the same set-up with the same model of decoder in another loco did not do it. So it might just be something to do with that particular decoder behaving strangely. 

I did try the CV8 reset, but it appears to make no difference. And according to the Hattons CV listing, there is no CV to 'modify' the brightness (as I see some, more expensive, chips do).

 

My problem is not with the tail & head light coming on at the same time, one is always lit while the other is 'dimmed'.

 

Ian

[ISW]

Many thanks for the responses chaps, they are much appreciated. Just seems odd that no one has reported this 'dimming' problem before. Maybe its just me being dim? (hope not ...).

 

Anyway, I have sourced some diodes from the local Maplin (while they are still trading ...) and will try those in series with my 0805 SMDs to see if that cures the dimming problem.

 

The other thing I need to 'investigate' is if there are difference 'flavours' (voltages) of 0603 & 0805 SMDs. I've got 3v SMDs (it says on the pack - and they do light brightly with a 3v button cell) and I'm wondering if there are 12v versions of the same out there. That might explain my problem if the DCC standard is for ~13v 'on' and ~3v 'off' on the blue / yellow / white wires.

 

Ian

[Lady_Ava_Hay]

I did try the CV8 reset, but it appears to make no difference. And according to the Hattons CV listing, there is no CV to 'modify' the brightness (as I see some, more expensive, chips do).

 

My problem is not with the tail & head light coming on at the same time, one is always lit while the other is 'dimmed'.

 

Ian

 

I have been cogitating about this. As a rank amateur when it comes to electrics, is it possible that a capacitor is somewhere in your lighting circuit and that is 'leaching' power through the dimmed bulb when the loco sees track power? If that dimmed state is not exclusive to one LED then that capacitor is in the blue circuit. Otherwise for only one LED to do this would suggest that the capacitor is in that circuit.

 

I am convinced that the fault is somewhere on the decoder PCB. They are all wired differently and only the PIC coding is common.

 

If I am right then some research as to which decoder to use is necessary.

[markw]

Rob,

 

I did take a look through Brians webpage (found from a Google search ...) and that's what convinced me to put 1k5 resistors on each leg of the White / Yellow instead of a single 1k5 on the Blue. Didn't seem to help though ...

 

Ian

Your description still implies that you are only fitting one resistor per end, you need one resistor for the red led and another for the white led at each end, so a total of 4 resistors.

Perhaps a diagram of how you have wired them would help

[Crosland]

The other thing I need to 'investigate' is if there are difference 'flavours' (voltages) of 0603 & 0805 SMDs. I've got 3v SMDs (it says on the pack - and they do light brightly with a 3v button cell) and I'm wondering if there are 12v versions of the same out there. That might explain my problem if the DCC standard is for ~13v 'on' and ~3v 'off' on the blue / yellow / white wires.

 

Forget about them being SMD, they are just LEDs. It makes no different to operation or resistor requirements.

 

Bare LEDs have a specified "forward voltage" or "Vf" which depends on the colour. What colour are yours? I suspect the "3V" refers to the maximum rated Vf.

 

LEDs specified as 5V or 12V include a suitable current limit resistor in the package. I've not come across a "3V" LED with integrated resistor, see above.

 

You should never connect a bare LED directly to a power supply. You only get away with it using a coin cell as they have such high internal resistance. In effect, you are relying on the cell's own current limit. One day you will regret it.

 

 

I'm using 3v 0805 SMDs with a 1500 ohm resister to avoid blowing the LEDs (done that already!). They are wired to the Blue / Yellow & Blue / White DCC cables to give directional working.

If you blew the LEDs on the decoder then you may have blown the decoder output as well.

Edited March 2, 2018 by Crosland

[ISW]

 

Bare LEDs have a specified "forward voltage" or "Vf" which depends on the colour. What colour are yours? I suspect the "3V" refers to the maximum rated Vf.

 

LEDs specified as 5V or 12V include a suitable current limit resistor in the package. I've not come across a "3V" LED with integrated resistor, see above.

 

You should never connect a bare LED directly to a power supply. You only get away with it using a coin cell as they have such high internal resistance. In effect, you are relying on the cell's own current limit. One day you will regret it.

 

 

If you blew the LEDs on the decoder then you may have blown the decoder output as well.

I used to use some of the '12v integrated resistor' type packages (using 1k5 + 0603 SMD) that came from Layouts4U with success. No dimming issue.

 

I am now using 'bare' 0805 SMDs that came from Direct Train Spares in Burnley. Both supplies were red LEDs. Soldering the SMDs is fiddly, but not a problem under a magnifying glass.

 

And in fact, I'm using the resistor off the Layouts4U package on the Direct Train Spares LEDs. So, you'd expect them to act the same, but they didn't. Maybe the resistor value is just a tad to small?

 

Yes, I know about blowing SMD LEDs ... I used a pair of AA batteries as a test supply first (err, 3V ...) and that blew the LED right away. It was bright though - briefly. Now I use a 3V button cell ... (less current) and no problems.

 

I have 3x DCC chips from Hatton to use at present. One is clearly blown, distinguished by the small hole burnt through the insulating plastic shrink wrap. I therefore have 2 others to verify it's not the chip causing the problem.

 

Ian

[ISW]

Your description still implies that you are only fitting one resistor per end, you need one resistor for the red led and another for the white led at each end, so a total of 4 resistors.

Perhaps a diagram of how you have wired them would help

Mark,

 

My 'acme' sketch is attached below. All my previous (successful) attempts at loco lighting were 'hard-wired' and looked a mess. This time I wanted to 'tidy things up' by using some veroboard.

 

When built it looks like this (with apologies for the biscuit crumbs that crept onto the photo).

 

And I'm using connectors at each end for the LEDs. As you can see there are 4x 1k5 resistors in place. All the 'tracks' are on the underside.

 

Ian

[Salmotrutta]

I did try the CV8 reset, but it appears to make no difference. And according to the Hattons CV listing, there is no CV to 'modify' the brightness (as I see some, more expensive, chips do).

 

My problem is not with the tail & head light coming on at the same time, one is always lit while the other is 'dimmed'.

 

Ian

The Digitrax makes the tail light dim so that's why I thought it might be worth a try. Sounds like it could be the transistor "leakage" that someone else suggested.