LED question

[jym1]

Hiya all.
 
I bought a Bachmann 44-tonner and since I don't run DCC, I removed the decoder as soon as I got it. I wanted to add lights so I purchased some Ngineering LEDs and used resistors taken from an LED light-strip. After testing on my workbench, I soldered the units to the PC board on the loco. The critter ran smoothly with bright directional-lighting BUT when I ran the unit forward or reverse, the LED was bright, but there was some "flickering" on the rear LED! Any ideas about the cause? Thanks, Jim in Tokushima, Japan. 

[mmagliaro]

Usually this is caused by slight intermittent contact with the track.  The power to the motor shuts off for an instant, short enough that you don't even notice it, but in that instant, the field collapses in the motor as it shuts off and there is an inductive spike out of the motor.  In other words, when the motor is shut off, there is an instant where the power stored in the magnetic field is converted to a brief spike of current, in the opposite polarity.  So you see a flicker in the LED.
First, power the engine on your workbench, connecting power leads DIRECTLY to the motor (no wheels or wipers in the path).
If the problem goes away, this intermittent contact is definitely what's happening.
Connecting a small capacitor across each LED should get rid of the flicker, something about 0.1 uF.

I'm a little puzzled by the fact that you always get the flicker in the REAR LED, whether you run forward or reverse.  Are you sure about this, and you definitely have the LEDs wired correctly so that the forward one lights up in forward and the rear one lights up in reverse? 

[jym1]

Hiya Max. Thank you for your very quick response. In your reply you talked about intermittent contact with the track and that got me to thinking. My test track was squeaky-clean and I had cleaned the wheels before, but decided to see if the wheels were the problem. I took a piece of 1500-grit sandpaper to the wheels and tried the critter on the test track again. Hey! No flickering! I don't know what type of metal B'mann uses for their wheels, but it doesn't seem to be nickel silver. I'm gonna try to fit a capacitor to the LEDs, but there's not much room.
    BTW, do you remember me buying Neo-Lube from you last year and having it shipped to my stewardess daughter in California? Well, she couldn't fly to Japan until a coupla months ago so we had to bide our time. Finally, she was able to "smuggle" it in and mail it to me from the airport in Tokyo. Better late than never, eh? ha-ha So, a belated Thank You for the Neo-Lube and, of course, for the LED information. Stay healthy, Jim 

[peteski]

Jim,  Any way you can post a schematic of this circuit?  I don't clearly see where the traces are cur on that copper clad board.

Like Max mentioned the flicker is cased by brief disconnections of wheels from the track which create very brief  (microseconds) open circuit between the throttle and motor.  While that happens the spinning motor acts as a generator generating BEMF voltage which is opposite polarity than the track voltage. That voltage is enough to power the very efficient white LED (opposite of the travel direction).

Adding a small capacitor in parallel with each LED will shunt those small voltage spikes and keep the LED from illuminating.  Not sure why only one LED is doing that. Maybe the other LED is not quite as efficient.  I small 0.1 micro Farad 50V SMD multilayer ceramic cap will work quite well.  You can see such capacitors installed on some of the factory light boards.

Also, those 150 ohm resistors have resistance a bit too low for this circuit. While they are ok when the loco is run slowly (low voltages), if you crank the throttle closer to 12V, the current passing through the LED will likely be exceeding their rating.  I would recommend at least 680 or 750 ohm resistors.

I know you're in Japan, but here is an example of the cap and resistors I would use:
https://www.digikey.com/en/products/detail/kemet/C1206C104M5RACTU/411247
https://www.digikey.com/en/products/detail/nte-electronics-inc/SR1-1206-168/11655663

[mmagliaro]

Jim thanks (about the Neolube), and I totally agree with Peteski on those low resistance values, which I completely missed.  I really wouldn't run the engine, or at least, keep the throttle really low, below like 3 volts, until you change those.  You will blow up those LEDs.  And I only said 0.1 uF, but yes, you want at least a 50v cap  (0.1 uF, with 50 volt rating).  The spikes that kick back out of the motor when it shuts off can easily be several times higher than the supply voltage that the motor was running at.

[jym1]

Hiya, Max and Pete. Thanks for all your input! An old geezer with no knowledge of electronics is fortunate to have people like you to lend a hand. I had no idea the resistors were only 150 ohms! I tried one out and thought it was okay even though the light was VERY bright! I just checked and I have 510 ohm resistors from Ngineering, good enough for low-power slow-speed switching? I don't have any SMD caps so that will take some time to order. Pete, here's the schematic:

[/url]

[Steveruger45]

Jim, I use 1000 ohm resistors for white LED’s in my locos.  Plenty bright enough for me.  So, if you had a load of the 510 ohm resistors on hand and room to get them all in, you could just put two resistors in series with each led.   Anyways you could try using both one or two of those resistors per led and see what you like before ordering more resistors.

[woodone]

Looks to me like the resistors are not wired correctly I do not see a cut in the PCB that would let the resistor even work?

[peteski]

Jim, thanks for the diagram - it answers the questions I had.
When I saw the photo of your ligth board I was (like Jerry) contused.  Whenever I see surface-mount (SMD) components, like those resistors, I assume that they will be soldered to the PC board. And in your example, the copper cladding is not cut around one of the resistor's terminals.  That would not work properly.  But I also see some white "stuff" around the resistors.  I wasn't sure what that was, but seeing the diagram I assume it is some sort of insulating material (caulking?) that keeps the resistors from touching the copper cladding.

It appears that you actually glued the resistors *ABOVE* the copper cladding, then soldered leads to them. You "converted" SMD resistors into regular leaded components.  Seems counter-intuitive to me, but it works.  I would have just isolated small islands of the copper cladding, then soldered the SMD resistors directly to the copper. One of the resistor ends would be soldered to the large copper area, while the other would be soldered only to the small isolated copper island.  Then run a wire from the small island to the LED.  But your circuit works as-is, so all is well.

As for determining the resistor values, it is easy once you understand the code.  You can Google for "SMD resistor value calculator" to find many otf them out there.  Basically the first digit is the first digit of the resistor's value, the second digit is the second digit of the resistance value, and the third digit is how many zeros follow the first 2 digits (to be accurate, it is a multiplier).  Value is on ohms.

Your resistors show "151" on them.  So the resistance is "15" plus one zero, or 150 ohms.  A 1k ohm or 1000 ohm resistor will be marked "102".  A 15k ohm (15000) resistor would be marked "153", 560 ohm would be "561",  and so on.

[woodone]

Pete, you are a much better detective than I. I would not have ever guessed using an adhesive that was non-conductive to hold the surface mount resistors. While this will work, its no wonder the LEDs are bright! I would think 560 or 1K would dim them a bit.
A small cap would prevent flickering.

[mmagliaro]

I wondered about this too (with the way those resistors are mounted on the copper), but figured they must be glued ABOVE the copper or the LEDs would have burned out immediately.

Oddly enough, this discussion prompted me to go do one of these in my "get around to it pile" - an old Atlas RS-1.  Just like your case, the opposite-direction LED flickers from momentary lost contact, but I just never bothered to fix it.  After this thread, I finally got in there and put the caps on it.   Worked like a charm.  No more flicker.

As for the resistor value, I have to say, I'd go with at LEAST a 1k ohm.  Those little SMD LEDs don't need much current at all to light up very brightly.   The LEDs I used in my engine aren't even SMD, and I am using 1.6k resistors because they are blindingly bright by 9-10 volts, and still plenty bright at 3 volts when the engine is barely moving.  I woudl suggest buying some 510, 1k, 1.5k, and 2k, and see how they look to you.  Use the highest value resistor you can where you still get pleasing light at all speeds.

As an aside, when I got this RS-1, one of the LEDs was burned out.  I noted that they only used  180 ohm resistors, and between that and the motor kick-back flickering, the LED was doomed.

[peteski]

Jim,
Another thing we did not address is protecting the white LEDs from excessive reverse voltage (which can damage the LED).  These LEDs are very sensitive to reverse voltages (and sometimes even 12V can damage them).  Better safe than sorry.

I drew up a circuit diagram, and PC board layout where the LEDs are connected in such a way that each LED protects the other one from excessive reverse voltage.  The other advantage of this circuit is that it is even simpler than your original circuit. It only required 1 resistor.  ANd the capacitor will prevent the reverse headlight flicker.

All you have to do is make one more cut to the PC board, and you can even solder all the components directly to the PC board.  Even if you don't use the SMD type resistor and capacitor, you can still solder their leads to the appropriate areas of the copper.  I also recommend a 1k ohm resistor.

The first drawing of the PC board shows the extra cut, then the 2 other ones show where to solder the components.  One shows wired-in LEDs, while the other shows the SMD LEDs soldered directly to to the PC board for (the cleanest, neatest install).

[jym1]

Hiya, Pete. Thanks for your circuit diagrams. I'm thinking to make my own board to replace the original because it would be difficult to make another cut. I have some really thin pc board and it wouldn't be that difficult. I also found that I have quite a few SMD 510 ohm resistors from Ngineering and I wanna wire 2 of them together to get my 1k resistance. How would that appear on your circuit?

[peteski]

Here you go Jim.  Just make sure the copper cladding is cut through.

[glakedylan]

just wondering...was this not a dual mode DC/DCC decoder that would have run on DC?
i know it is not the best situation to run dual mode on DC unless you are able to program the decoder as such.
but would that not have saved all the work? just really curious.
i hope you get it to run well.

sincerely
Gary