JNSF DIY lighting project (PCB based)

[cteno4]

It’s possible (but may make layout hard) to have a lead that would route around the switch so you would either solder your main dropping resistor from the input to the switch pad to the runaround lead. Or if you put the switch in then the circuit routes thru switch and it’s resistors.

 

theoretically, the voltage regulating circuit should be the same brightness 6-12v.

 

what are the simpler circuits? With dcc only you could get rid of the voltage regulator, but you still need the rectifier and a current limiting resistor. Caps of course are optional. I think the caps for buffering the voltage regulator are just icing as with low power being put thru and the fact the leds are being driven way below limits and any surges are not of any consequence. Even the resistor to slow the charge of the caps could probably be eliminated.

 

jeff

[chadbag]

One reason is that you may want to A/B two different values in a given train and not want to take it off and put it back on etc.  You probably would not want to put the switch on every board.

[chadbag]

1 minute ago, cteno4 said:

It’s possible (but may make layout hard) to have a lead that would route around the switch so you would either solder your main dropping resistor from the input to the switch pad to the runaround lead. Or if you put the switch in then the circuit routes thru switch and it’s resistors.

 

Yeah, maybe something like that would be better than requiring a trace to be cut.

 

 

[gavino200]

2 minutes ago, chadbag said:

One reason is that you may want to A/B two different values in a given train and not want to take it off and put it back on etc.  You probably would not want to put the switch on every board.

 

I'm pretty sure I would never do that. I'd just think it through, choose my values.And leave it. No reason you'd need to choose the same resistance for each coach in a train. 

[chadbag]

1 minute ago, gavino200 said:

 

I'm pretty sure I would never do that. I'd just think it through, choose my values.And leave it. No reason you'd need to choose the same resistance for each coach in a train. 

 

 

But you may want to A/B two values for a new train.  Maybe 2.2 and 2.4 or something.  Maybe the interior is a different color or brightness and moderating the LEDs differently may make that train look better compared to a different train with different values.

 

The good thing is, if you don't want it, you don't have to use it.

 

[gavino200]

8 minutes ago, chadbag said:

But you may want to A/B two values for a new train.  Maybe 2.2 and 2.4 or something.  Maybe the interior is a different color or brightness and moderating the LEDs differently may make that train look better compared to a different train with different values.

 

I may use ten different values on ten different trains. Or multiple different values within the same train. Or I may just find a value that is fine for all trains. But I'd sort it out by trial and error. ie. install a strip, see if I like it. Then change or move on. 

 

I wouldn't solder up a gizilion in advance. I'd still be in a fix if neither of the A/B values turned out to work for a particular train. 

 

I do think it's cool that you worked out how to do this. But I think it adds unnecessary complexity, and a moving part that can fail. 

 

Quote

 

The good thing is, if you don't want it, you don't have to use it.

 

 

This is true. 

Edited March 20, 2019 by gavino200

[chadbag]

1 minute ago, gavino200 said:

 

I may use ten different values on ten different trains. Or multiple different values within the same train. Or I may just find a value that is fine for each train. But I'd sort it out by trial and error. ie. install a strip, see if I like it. Then change or move on. 

 

I wouldn't solder up a gizilion in advance. I'd still be in a fix if neither of the A/B values turned out to work for a particular train. 

 

Wait until you want to start removing the ones you put on 🙂

[cteno4]

I can see this being something you will probably want to set train by train depending on interiors, Windows and just how the train looks.

 

im curious to try putting the smd pots in and see how close they can be tuned, I expect they can. Then it’s dirt simple to set things. 

 

switch failure would not be a worry unless you are planning on switching it a thousand times then maybe a small chance...

 

jeff

[gavino200]

7 minutes ago, cteno4 said:

I can see this being something you will probably want to set train by train depending on interiors, Windows and just how the train looks.

 

I totally agree.

 

Quote

 

im curious to try putting the smd pots in and see how close they can be tuned, I expect they can. Then it’s dirt simple to set things. 

 

I think you're asking for pain here. KISS principle. In any case you have to open and close the car with every adjust. Also you'd need to do it for each car. Whereas with ordinary resistors once you chose one value for a simple train (similar coaches) you just rinse and repeat for the rest of the carriages. Also what looks right for an E5 shinkansen will likely look right for an E6. With pots you're going to have a lot more trouble. Also, it would be hard to get uniformity. Doable but a PITA.

 

Quote

 

switch failure would not be a worry unless you are planning on switching it a thousand times then maybe a small chance...

 

 

Murphy's law. You will encounter it.

Edited March 20, 2019 by gavino200

[gavino200]

Kiss principle guys!

[cteno4]

59 minutes ago, gavino200 said:

I think you're asking for pain here. KISS principle. In any case you have to open and close the car with every adjust. Also you'd need to do it for each car. Whereas with ordinary resistors once you chose one value for a simple train (similar coaches) you just rinse and repeat for the rest of the carriages. Also what looks right for an E5 shinkansen will likely look right for an E6. With pots you're going to have a lot more trouble. Also, it would be hard to get uniformity. Doable but a PITA.

 

I disagree, just make them all the same, then you can set any to what you want. You are going to have to do the car open close to figure out the value of the limiting resistor. Once found just set that for the train. Plus if a while later you decide you don’t like it it’s simple to reset w,o unslodering and resoldering. You are soldering a single component just like the fixed resistor, so no more work there. Just adjustable if needed w,o resoldering.

 

probably best to make a board with the limiting resistor pads soldered to some heavy magnet wire and then connected to a pot to figure out what value the resistor should be for the train being able to adjust it outside the car easily 

 

Quote

 

Murphy's law. You will encounter it.

 

I think thats taking that a bit too far. A switch here is not a failure item if you want/need it. Dont bother running any of your trains as they will go up in flames if you take this path.

 

many paths to kiss. I think the circuit will probably do fine with just rectifier - voltage regulator - cap - limiting resistor - leds and eliminate the voltage regulator caps and the cap charging resistor if you want to simplify.

[gavino200]

28 minutes ago, cteno4 said:

 

I disagree, just make them all the same, then you can set any to what you want. You are going to have to do the car open close to figure out the value of the limiting resistor. Once found just set that for the train. Plus if a while later you decide you don’t like it it’s simple to reset w,o unslodering and resoldering. You are soldering a single component just like the fixed resistor, so no more work there. Just adjustable if needed w,o resoldering.

 

How small is the pot that you're thinking of? It needs to be very small in order to not obstruct or be visible. 

 

Soldering an SMD resistor is seriously not difficult. They're minuscule and close to 100% reliable. 

 

 

Quote

 

probably best to make a board with the limiting resistor pads soldered to some heavy magnet wire and then connected to a pot to figure out what value the resistor should be for the train being able to adjust it outside the car easily 

 

That's not a bad idea. Is that what you were talking about earlier? I thought you were considering incorporating a pot into the circuitry of the LED strip. If you're just talking about using a pot to decide on the value of the dimming resistor then we're in complete agreement. In fact it doesn't even need to be an SMD pot. You can use a larger one and rig up a little test board.

 

Quote

 

I think thats taking that a bit too far. A switch here is not a failure item if you want/need it.

 

It's unnecessary. I really don't think it adds anything. I am however impressed that Chad can pull it off. 

 

Quote

 

Dont bother running any of your trains as they will go up in flames if you take this path.

 

I wouldn't if they were made from scratch by one of us. 🙂

 

(Not without some careful testing anyway)

 

Quote

 

many paths to kiss.

 

Yes, and all paths should be followed. KISS is the place to be.

 

Quote

 

I think the circuit will probably do fine with just rectifier - voltage regulator - cap - limiting resistor - leds

 

I agree.

 

Quote

and eliminate the voltage regulator caps and the cap charging resistor if you want to simplify.

 

I probably agree with eliminating the cap charging resistor and the voltage regulator caps. I definitely agree with testing the strip without these and only including them if they make an observable difference. 

Edited March 20, 2019 by gavino200

[Ochanomizu]

Hello Mr chadbag and Mr gavino200,

 

I continue to watch this thread with excitement.  Personally, I have 191 consists.  All passenger, no freight.  All Japanese.  I have 1225 cars that need lighting.  I have a variety of consists from the big 4 manufacturers in Japan.  I have a few questions:

 

1. Will your light board fit in all? 

2. Will it fit bilevel cars, such as MicroAce 0020 and 0021, 100 and E4 series shinkansen etc: Kato 10-292 10-354/5/6 , Tomix E1 shinkansen 920059 etc ?

3. Can the board be trimmed down to shorter cars, such as Kato 10-864 Ginza Line, Odakyu Romance cars and the like?

 

If the lighting board is the magic bullet for all manufacturers then I'd be inclined to place an order for 1500.  I run DCC but I'm not interested in DCC'ing my interior lights.  My layout has the ability to turn power to sidings off remotely.  This is an adequate solution for me.

 

I would want the option to control brightness, if that is on the agenda.  I think switches are to bulky and expensive.  I would suggest 3 or 4 resisters in series, all of which are bypassed with a piece of wire.  Simply snip off the wire to force the current through a resistor.

 

If I'm understanding this thread correctly, and if you can accommodate a simple brightness control along the lines I've mentioned above, I'd commit to 1500 units at around $0.20 - $0.25 each.  I would first want 2 or 3 to try, for which I'll pay for product and postage.

 

I mention this today, because it might influence the size of your order.

[gavino200]

6 minutes ago, Ochanomizu said:

 

1. Will your light board fit in all? 

 

The goal has been to make the strip as universal as possible. A caveat is that they don't have a direct connection to the pick up strips of a coach. A small length of wire will be needed to connect the LED strip to the car's pickup strip. 

 

6 minutes ago, Ochanomizu said:

2. Will it fit bilevel cars, such as MicroAce 0020 and 0021, 100 and E4 series shinkansen etc: Kato 10-292 10-354/5/6 , Tomix E1 shinkansen 920059 etc ?

 

I don't believe we've discussed this question. I have been thinking about it a bit as I have some bilevel commuter cars and have a pre-order for a Sunrise Express. My bilevel commutor cars are TOMIX. JR500 recently posted detailed pictures of a interior detailing and lighting project on a Sunlight Express in the "improving your rolling stock" thread.

 

We should give this some thought.

 

 

6 minutes ago, Ochanomizu said:

3. Can the board be trimmed down to shorter cars, such as Kato 10-864 Ginza Line, Odakyu Romance cars and the like?

 

Amazingly Chad has been able to incorporate this feature into his design. (hats off to Chad) You should be able to score and cut the strips to length. 

 

6 minutes ago, Ochanomizu said:

 

If the lighting board is the magic bullet for all manufacturers then I'd be inclined to place an order for 1500.  I run DCC but I'm not interested in DCC'ing my interior lights.  My layout has the ability to turn power to sidings off remotely.  This is an adequate solution for me.

 

I have the same goals as you. 

 

I like the idea of DCC interior lighting but it's too much hassle for me. Chad is more tenacious. The exception is with European cars. If I use magnetic couplers to route power it should be possible to control a whole train's lighting with just one decoder (I believe). In that case it would be worth it for me.

 

 

6 minutes ago, Ochanomizu said:

 

I would want the option to control brightness, if that is on the agenda.  I think switches are to bulky and expensive.  I would suggest 3 or 4 resisters in series, all of which are bypassed with a piece of wire.  Simply snip off the wire to force the current through a resistor.

 

I want'to emphasize, that the idea is that the strips would come blank - just a circuit board with solder pads and a circuit path. You'd have to add all the components yourself, though suitable components can be specified and sourced. You could just add the resistor that you chose. 

 

Jeff has outlined a nice method for choosing resistor value in a post above. Basically using magnet wire to include a potentiometer into the circuit. Then close the car and adjust the potentiometer to the desired brightness. Read the resistance with a multimeter. Then use this value and choose the closest resistor available. You can tailor to taste.

 

 

6 minutes ago, Ochanomizu said:

 

If I'm understanding this thread correctly, and if you can accommodate a simple brightness control along the lines I've mentioned above, I'd commit to 1500 units at around $0.20 - $0.25 each.  I would first want 2 or 3 to try, for which I'll pay for product and postage.

 

I mention this today, because it might influence the size of your order.

 

The more the merrier. Please continue to contribute. It's sort of a crowd source project, though Chad is definitely the heart of the operation.

[chadbag]

19 minutes ago, Ochanomizu said:

Hello Mr chadbag and Mr gavino200,

 

I continue to watch this thread with excitement.  Personally, I have 191 consists.  All passenger, no freight.  All Japanese.  I have 1225 cars that need lighting.  I have a variety of consists from the big 4 manufacturers in Japan.  I have a few questions:

 

1. Will your light board fit in all? 

2. Will it fit bilevel cars, such as MicroAce 0020 and 0021, 100 and E4 series shinkansen etc: Kato 10-292 10-354/5/6 , Tomix E1 shinkansen 920059 etc ?

3. Can the board be trimmed down to shorter cars, such as Kato 10-864 Ginza Line, Odakyu Romance cars and the like?

 

If the lighting board is the magic bullet for all manufacturers then I'd be inclined to place an order for 1500.  I run DCC but I'm not interested in DCC'ing my interior lights.  My layout has the ability to turn power to sidings off remotely.  This is an adequate solution for me.

 

I would want the option to control brightness, if that is on the agenda.  I think switches are to bulky and expensive.  I would suggest 3 or 4 resisters in series, all of which are bypassed with a piece of wire.  Simply snip off the wire to force the current through a resistor.

 

If I'm understanding this thread correctly, and if you can accommodate a simple brightness control along the lines I've mentioned above, I'd commit to 1500 units at around $0.20 - $0.25 each.  I would first want 2 or 3 to try, for which I'll pay for product and postage.

 

I mention this today, because it might influence the size of your order.

 

Hi!

 

The boards (without components) in quantity are probably about 0.23 US$ based on a sample I ran.  I will provide you the files necessary and you can order them yourself from any PCB manufacturer (I'll give you the URL of the one I have been using, which I got from the forum member who did the EM13 board).

 

Since it is the board only, you will need to get your own components (caps, resistors, etc) for them but most of that can be gotten for really cheap in qty.  So a completed board should be doable for under $1 with components I think.  I have voltage regulators and rectifiers for a few pennies, LEDs for under a penny each, resistors etc as well.  Caps for about 4-5 cents each (and you use like 6-8 of those).    You will be able to use whatever resistor values you want, so no need to put a string of them and bypass with a wire.  The board is able to be shortened so it should fit a variety of wagons, from long Euro wagons to shorter Japanese EMU.    I don't have any bi-level cars myself, but as long as you can stick a shorter board in it should fit.

 

I am working on a better design for a V2 board.  It should also be able to fit in more trains as it should be a bit narrower.  

 

I am also working on a V3 board which won't have any LEDs on it -- just the caps etc so you can hook it into existing lighting.

 

[chadbag]

Here is the basic schematic for V2.  It does not show the switch, extra solder pads, or anything.

 

Schematic-V2-v0.1.pdf

 

The idea is you want the caps before the voltage regulator, so that they charge to the full DC/DCC voltage so when they discharge, that higher voltage gets converted to the lower 3.3V or 5V so as long as the cap voltage is above the activation voltage (6.5-7V or so on a 5V converter) you get the full 5V out for your lights.   The RC time constant for the decay of the cap is to 60% or something close to that of the full charge voltage.  The V1 circuit on the boards that are being made start to get dim right away but do last the time of the RC time, but they just start dimming since the caps are charged to the 5V level of the regulator since they come after the regulator, and so as they start to discharge, that RC time constant is a drop off decay from that 5V.  It still helps with flicker as most flicker is just momentary and a slight dimming before it brightens back up won't really be noticeable.   However, with this new circuit, you should get constant full brightness for close to the full RC time constant amount (assuming 12V as 60% of 12V is 7.2 volts which should still trigger the full 5V out the voltage regulator).    Sadly, I did not think of this myself.  I was looking at other circuits for this sort of thing on some model railroading sites and one example circuit did just this and once I thought it through it made sense.

 

On the schematic, values shown are just representative,   The 47uF are the actual power caps for the circuit.  You can use whatever size you feel works for you and as many (up to the space provided) as you want.  Leaving blank the ones you don't need (ie, 6 of 9 used or whatever).   The 2 smaller ones are the power stabilizing ones for the voltage regulator.  If you leave them off the circuit will still complete no problem (you may have dirty voltage which could lessen the life of your LEDs etc).  However, any resistors left off will have to be jumpered since a resister is in series so needs to be there to not break the circuit.   Any cap pads left empty don't hurt the circuit as they are in parallel.  Same with the LEDs.  You can leave off LEDs.

 

The schematic does not have the rectifier bypass pads, the switch, etc. in it yet.  This is just the basic circuit.

 

[chadbag]

49 minutes ago, gavino200 said:

I probably agree with eliminating the cap charging resistor and the voltage regulator caps. I definitely agree with testing the strip without these and only including them if they make an observable difference. 

 

I don't think it is worth leaving them off, but you can.  However, if you leave the resistor off you need to bridge it with a jumper wire soldered in its place.  The caps can be left off.  However, I don't think it is worth it.  They are recommended by the manufacturers and help stabilize and clean the power.  It may affect the LED lifespan or something with harder "power edges" hitting them etc.  For the couple pennies it costs for ceramic caps it is not worth it IMHO.

 

 

Also FYI:  In my buying of components, I've noticed the 6.3V (for 3.3 system) and 10V caps are more expensive (same size) as the 16V caps which are more expensive than the 25V caps.  All the same size (2D size -- height may be different).  The 47uF caps in 1206 size I was looking at were 3-5 cents each at 25V, 5-6 cents in 16V, and up to 10 cents each in the 10V or 6.3V.  (This is from memory so I could be off a penny here and there).  So I will just be buying the 25V or 16V ones for most cases.  If the height is different (I have bought a small number of low voltage ones to compare; then if I have an installation that is particularly height sensitive I might spring for the more expensive ones on a limited basis if they save a hair of height).

[chadbag]

28 minutes ago, gavino200 said:

 

The goal has been to make the strip as universal as possible. A caveat is that they don't have a direct connection to the pick up strips of a coach. A small length of wire will be needed to connect the LED strip to the car's pickup strip. 

 

It should be possible to buy brass strips and bend them (and solder to the input pads) such that you can fit them in a KATO wagon, theoretically speaking.  I have not tried it.  I may as an exercise (I don't like to notbe able to do things I think should be able to be done 🙂 ).

 

[chadbag]

With regards to the V2.  I ordered some breadboards, jumpers, and non-SMD component assortments for caps and resistors, and i already have a  bunch of normal LEDs, so I will be able to test the V2 circuit more easily than soldering it up using wire.

[cteno4]

58 minutes ago, gavino200 said:

 

How small is the pot that you're thinking of? It needs to be very small in order to not obstruct or be visible. 

 

3.1 x 3.8 x 1.6 mm so below the height of other components

 

Quote

Soldering an SMD resistor is seriously not difficult. They're minuscule and close to 100% reliable. 

 

For all intensive purposes a variable is as well in this case

 

Quote

 

 

That's not a bad idea. Is that what you were talking about earlier? I thought you were considering incorporating a pot into the circuitry of the LED strip. If you're just talking about using a pot to decide on the value of the dimming resistor then we're in complete agreement. In fact it doesn't even need to be an SMD pot. You can use a larger one and rig up a little test board.

 

Both. Use an external one to set your resistance for the light needed then just set it on your boards with the pot instead of a fixed resistor. Can the easily be changed if needed later. One size fits all.

 

whrn ever I’ve played with scale lighting it’s been a lot of fiddling to get it looking right and so I’ve gone down the path on structures to always use pots to allow a lot of fiddling at any point. I am going to experiment moving that into car lighting when I go there for sure.

 

Quote

 

It's unnecessary. I really don't think it adds anything. I am however impressed that Chad can pull it off

 

Switch can be added to the board and you just add it and two different resisor values or single resistor and no switch. If he wants it there its just the path design to make it so.

 

i can see why they did it on the poppendetta boards as you are buying it all made up so they wanted to give two light levels to use instead of just one.

 

jeff

[Ochanomizu]

Hello,

 

Thanks for the responses to my last post.  I must admit, overnight the penny dropped and I realised the price was for unpopulated boards.  Can you obtain a price for complete boards with components?  I'm not keen on soldering up 1500 boards ... I'd prefer to purchase completed boards, I think.

[chadbag]

4 minutes ago, Ochanomizu said:

Hello,

 

Thanks for the responses to my last post.  I must admit, overnight the penny dropped and I realised the price was for unpopulated boards.  Can you obtain a price for complete boards with components?  I'm not keen on soldering up 1500 boards ... I'd prefer to purchase completed boards, I think.

 

I don't know where to look for completed boards, but I would guess it would be a bit more.  I've been researching and using the so-called "hot plate" method, it should go pretty fast to solder up the boards.

 

You can google "hot plate smd" and find a bunch of articles and videos on it.  Basically, your typical portable hot-plate is used to heat the PCB up to hot enough to melt low temp solder paste.  So you apply low-temp solder paste to each board (using a syringe or template), place all the elements, and stick it on the hot plate until the solder melts and flows.

 

 

[cteno4]

It’s the frying pan method! I saw that a few years ago and meant to try it. Don’t have to worry about components moving and can do a number of boards at once. Small oven is a couple hundred bucks.

 

Im guessing 1500 would be the beginning of getting a very small commercial run. China or India. From the numbers I remeber Ardesh was doing with pcb work it was starting around that number in India and China.

 

Jeff

[Ochanomizu]

Hello Mr chadbag,

 

I found a China manufacturer: www.pcbway.com

 

Could you check it out?  I tried their instant quote based on your V2 schematic.  I may have answered some of the questions incorrectly, but 8000 units appears to be around $2000.

[Martijn Meerts]

I have soldering tweezers, great for SMD stuff.