Skytop: Installing DCC and accurate tail light function Hiawatha Kato N scale

[u18b]

The LED dies at even 10k resistor.
voltage is whatever the function wire is on the decoder.

So it is dying connected to the decoder.

After 1 to 3 minutes or so.  Not at once.

I think they are just bad quality.

[nkalanaga]

Ron:  Anode is positive.  The easy(?) way to remember that is that old TVs used CRT screens - Cathode Ray Tubes.  Those are basically an electron gun hitting a light-emitting target.  Electrons are negative.  So, "cathode rays" are negative, and so is the cathode.  Thus, the anode is positive.

[peteski]

The LED dies at even 10k resistor.
voltage is whatever the function wire is on the decoder.

So it is dying connected to the decoder.

After 1 to 3 minutes or so.  Not at once.

I think they are just bad quality.

Something doesn't add up.   The function output will be approximately the DCC track voltage minus whatever is dropped by the bridge rectifier and function output transistor of the decoder. I usually allow about 2V for that.  So if the DCC voltage is 13V then the voltage between the blue wire and function output will be 13-2=11V.  The blue or white LEDs have voltage drop around 3V.  So with a 10k resistor and 11-3=8V across the resistor, the current passing through the LED will be 8/10000=0.8mA  That is very low (0402 LEDs are usually rated for 5mA.  There is no way that LED should be failing. Even low quality LED should not fail operated at 0.8mA. And since you are powering it from the decoder function there is no possibility of any reverse voltage spikes which could damage the LED.

I wonder if maybe you are overheating that tiny LED when soldering it? Blue/white LEDs are more delicate than other color LEDs.  Do you have temperature-adjustable soldering iron (lower it to about 550 deg. F)?  Could you then quickly solder some fine wire leads to the LED then use the wires to temporarily connect the LED to your circuit and test it? If you don't have any fine wire, just strip some insulation from stranded decoder wire, then take single strands and use them as the LED hookup wire.

[u18b]

I wonder if maybe you are overheating that tiny LED when soldering it? Blue/white LEDs are more delicate than other color LEDs.  Do you have temperature-adjustable soldering iron (lower it to about 550 deg. F)?  Could you then quickly solder some fine wire leads to the LED then use the wires to temporarily connect the LED to your circuit and test it? If you don't have any fine wire, just strip some insulation from stranded decoder wire, then take single strands and use them as the LED hookup wire.

That is the most likely thing.   I have a lower watt iron, but not self adjusting.

I remember from the N-gineering site, he recommended a lower temp solder too.

The white ones must be super delicate.

The other colors do fine.

[u18b]

I decided to bail out on the 0402 white.

I have a bunch of 0603s.  So I made a light last night that worked.

I'll post shots possibly tonight.

The 0603 is bigger than I wanted, but may/should work.

[u18b]

Now, in order for this two-color deal to work, we need a clear lens.

So I removed the shell- easy.  Just spread the sides a little.  Comes right off.
The stock lens is red.




The lens is 60 mil with a lip.




Unfortunately, I didn't have anything like that in my stash.
And searching ebay, I saw no hobby fiber optic that big.

I thought about buying a Kato HO scale lens to something- but that is expensive.

So I took my wife to the Dollar Tree.
Walked around the store looking for SOMETHING that I might use.

Then saw a pack of clear push-pins.  I'll give it a try.




After I started on one.... I saw that they were not all the same.   Note how the head of the pin sticks up so high.  I don't want to be anywhere near that.  So I looked for another.




This one is much better.
Chucked it in the motor tool.
Trimmed the flange on the edge.
Got a file.




With the motor tool running, I moved the file diagonally.
A LITTLE at a time!   To much builds up too much heat and melts the plastic.




Patience!
Go slow.




Getting close.  We're about at 75 mil right now.




We're there.



With the tool spinning again, I put an Xacto blade at the fat part to gently try to cut it off.



I then take a file and file the fat part down toward the thin.
Test fit.  This is close.




Filed some more.
Sanded lightly with ultra fine sand paper.
Then gently brought a flame close to it.  (too much and it melts too much and time to start over).




At present, still too long.  Needs to be trimmed.



This is it.




New one doesn't have the same lip, but it looks pretty good for an amateur.

[milw156]

Where there is a will, there is a(t least one) way!!! Nice work!
Rick

[lashedup]

Very nice Ron. I was rummaging through the bin-O-parts trying to find something clear and you stumbled upon a great solution.   

[u18b]

Now we can turn to the actual electronics.
I'm presenting my solution, but others are welcome to add their own approach.

THIS is how much room we have to work with if we mount the LED package in the rear- which is the way I'm going.
(someone else may wish to mount the LED further in inserted into the path of the light tube).
My way should provide the strongest light at the exit hole.




I measured that hole roughly with an SMD resistor I had on hand.
I guestimate that hole is less than 2mm wide and more than 3mm tall.

I used the resistor in this picture.  It would easily fit up and down.  However, when turned sideways, the resistor was too big.
And you can see the 0603 LED in the middle and the 0402 on the right.



Here (below), the bigger 0603 is the white LED.
The red LED is the smaller 0402 (which is clear in this photo).
You can see that they are soldered with a common positive (bottom side) and separate negatives on the top side.




I'm using three colors of magnet wire (sorry if my color scheme makes no sense).
Red (for "hot" for the positive).  It goes to both LEDs on the positive side.

The natural copper wire goes to the white LED, negative side.
and the green wire goes to the red LED, negative side.




Everything works.
White shows solid when in reverse.
Red varies as a Mars light when forward.

[Sokramiketes]

Just an FYI, Plastruct has various sizes of clear acrylic rod available.  I use it for headlight lenses all the time, and it would work if you want a more homogenous plastic rod.

Pressing on a hot, non-stick pan will get you a lip.

[peteski]

Just an FYI, Plastruct has various sizes of clear acrylic rod available.  I use it for headlight lenses all the time, and it would work if you want a more homogenous plastic rod.

Pressing on a hot, non-stick pan will get you a lip.

Also, placing the end of the rod very close to a side of the soldering iron barrel shoudl mushroom the end and leave a nice and smooth lens-shaped end.

[u18b]

Thanks guys.

Time to move in the direction of installing after all this testing and getting ready.

First, I used some very thin copper clad fiberglass board for the resistor board for the LEDs.




I shaped it to fit under the clear light pipe.




The light pipe was tall enough to still get SMD resistors under there.



Used a motor tool to cut channels.




And soldered my resistors.
I labelled them positive (+)
White
and Red.

The extra channel is just extra in case I want to do something else some day.  Hated to wast the extra space.

[u18b]

This is the support for the light pipe where the "lamps" are.

Since I am running wires, I need to cut the bottom support.

Before



And after



Ready to do more!

[u18b]

After a lot of debating in my mind, I decided to mount the LED array right on the light pipe.
I made my barrel lens a bit shorter than the stock, so it should be OK.  If not, I'll shorten it some more.

Here is the stock piece.



I oriented the array with the red LED on top and superglued it with the red LED toward the top of the light pipe.
So the white LED hangs lower.




Here is an end view.
Looks like I could have trimmed the left side a bit more.
But that thing is pretty TINY.




And wires soldered to the board.




Getting to this step was a bit tricky.

In the end, I superglued the LED wires to the outer edge of the light pipe since it was too hard to route everything properly otherwise.



Sort of nearing the end.
The last major hurdle will be installing the decoder.

[u18b]

I test fit the interior and found I needed just a little bit more room in the hole.
So I opened it up.



Earlier in this thread, I removed the hatch from the bottom.
Here I'm showing how I removed that inner brace that's on the right side.




With that brace removed, a Digitraks DZ126T decoder will fit in the toilet tank (assuming that's what it is).



Time to wire the decoder.  White shrink wrap removed.  I already had removed the stock wires.
These are the LED wires.
The middle left is the positive.

I actually reversed the F0 wires (so much for my color scheme).
I reversed them after this pic.



Again, I needed a path to route wires.
So since the toilet tank covered most of this area, I cut what is shown.
All this is totally hidden when assembled.

Also note I cut a channel in the two braces on the left side for wires.


Everything fits now.
I soldered the LED wires to the resistor board.



I did one last test.  Took this to the layout and used jumper wires with alligator clips to the track.
Still worked.  Woot!

Now time to get power to the decoder.



I soldered the red and black track wires to the triangular springs that go to the light board.
Notice I soldered on the inside/top edge of the pad so there would be no obstruction when assembled.
Red is right rail.   


All fits nicely now.



I'm not totally sure this was necessary, but I did it anyway.
In the past, I have found it impossible to program a decoder without some resistance on the motor leads where there is normally a motor!

So I used a 12v light bulb.
Instead of a switch, I just installed a TCS micro plug and socket.
I soldered the socket to the orange pad.  And then ran the gray wire over to the other pin on the socket- as shown here.
(notice the function wires have been moved.  The green wire is on the top pad.




I soldered the plug to the bulb.




The bulb can be plugged in when I need to make a programming change.
Otherwise it can just be left off.




Here is the car on the main line with the power turn up all the way.
I set Vmax at only 5 and Vmid to 2.  So wide open is only a couple of volts.



But I don't want the bulb to get warm.
So I leave it unplugged.
And the toilet tank is big enough to store the bulb so I don't lose it.




Now, all I need to do is remove the toilet tank piece and the decoder (and bulb)  is accessible.


Well, I'm done.

Show and tell photos and video will have to come later.

All I can say until then is....
Awsome.