PRR Position Light Signal - Chapter 2

[bman]

So what's the major difference between the original posting's circuit board and the one just mentioned besides about $19.50?  Besides their manufacturer.  I've never heard anyone mention oshpark before. That's a pretty big difference in price and while I can solder just about anything, I usually have to be shown what to solder where.  So my knowledge of electronics is pretty much contained to, "don't touch that."  Thanks.

[dcutting]

The OP's board had all of the wires going down into the pole... that's a lot of wires. My circuit board is operated with only 4 wires (VCC, GND, and two communication wires). I might be able to get it down to three wires eventually. It is very flexible because it is controlled by software. For testing purposes I will use an arduino board to control the microcontroller in the signal head. Once the prototype is done, I can move to a thinner board.

To solder the chip on the back of this board you will NEED a hot air rework station ($100 at sparkfun), because the board uses a QFN chip that does not have exposed metal. I will offer these as RTR signals because of this, a la BLMA.   

-David

[dcutting]

I guess the question here is what circuit system would be usable for this sort of thing?

-David

[C855B]

I guess the question here is what circuit system would be usable for this sort of thing?

The most popular protocol intended for conventional color-light ("Type D") MRR signals is common anode, one cathode lead per LED. Personally, I would design around the Digitrax SEC8 signal controller using their common-anode default, which should cover most other plug'n'play signal drivers. Atlas uses common-cathode, but it is my understanding (not firsthand) that Atlas detection has issues with DCC, and is also not nearly as popular as Digitrax (and compatibles) for signal control.

Frankly, I would not go in the direction you are going with a comm-based control scheme to the chip in the head. KISS should apply - one wire per aspect, to match the available off-the-shelf logic. You're already at four, or three if you use the mast as a common, so you are no worse/no better than the comm-based scheme. Otherwise, you are committing yourself to a second board to translate one-per aspect to your comm system.

[dcutting]

Hi guys,

Here is the next version of the signal head: https://oshpark.com/shared_projects/hR3dQ2gs

This version has holes going through the board to the back, so the 0603 LED's are inserted light-first, then soldered in. This will make it so that the .1 mm PCB will look like sheet metal. There are five contacts on the back of the board. One of them is soldered onto the pole for the common, the other four are for the center light, the pair of "stop" lights, the pair of "slow" lights, and the pair of "go" lights. they will run down the tube.

I ordered three of them, and they should be here very soon, along with my last not-so-nice attempt. 

-David

[bman]

I'm ultimately going to be in need of 100+ PL heads. I'm really sad that @TrainCat2 has gone offline. He had or had in the works everything I need. Unfortunately, just as much as I need the heads, I need masts and signal bridges. Not sure what I'm going to do if he doesn't come back, but these are a possibility.

Here's a link to the masts that nscaleSPF2 used.  Got the info of the original signal head post.

http://benscale.com/

Once I get a couple of other projects out of the way I need to submit an order to benscale for myself.

  So getting back to the original idea.  Is this still a possibility or is there not enough interest?  I do like the idea of having the led's pre-wired on a pcb as I studied nscaleSPF2's build of these signal heads on the thread he referenced.  Call me lazy.  It just appeals to me not to have to solder something THAT small.  Although if I had to so be it.  Been looking at the 1:1 and was wondering if the wires could somehow be made to resemble the skeletal framework(spider?) that the lights were mounted on. 

 

[eric220]

...Is this still a possibility or is there not enough interest?  I do like the idea of having the led's pre-wired on a pcb as I studied nscaleSPF2's build of these signal heads on the thread he referenced.  Call me lazy.  It just appeals to me not to have to solder something THAT small.

Pre-wired pcbs are just about a requirement for my application.  If I ever hope to get my signals done, I'm going to need to outsource some of the work.  I can assemble the bridges and heads, but doing all the fiddly bits on the circuit board (or learning to program micro-controllers) is not on my agenda.

I'm looking forward to seeing how this project progresses.

[nscaleSPF2]

dcutting,

I applaud your efforts, David.  The path that I was going down was not a good one; your path has a higher probability of success.  I just have a couple of suggestions, but first I need to answer some of the other's questions.

Somewhat OT.... what are you guys using for a tiny good yellow LED?

I had spec'd out this 0402 LED:

http://www.kingbrightusa.com/product.asp?catalog_name=LED&product_id=APG1005SYC-T

So what's the major difference between the original posting's circuit board and the one just mentioned besides about $19.50?  Besides their manufacturer. 

The quotes that I got included engineering design costs, tooling cost for a minimum of 100 pieces, the cost of the LED's, and the cost to attach the LED's to the circuit board.  Since I don't have the skills to do any of these, I would have to pay for them.  David apparently does have these skills.

The OP's board had all of the wires going down into the pole... that's a lot of wires.

For the record, my design has 5 wires going down the pole.  This is a lot of wires, but it works.

Here's a link to the masts that nscaleSPF2 used.  Got the info of the original signal head post.

http://benscale.com/
Once I get a couple of other projects out of the way I need to submit an order to benscale for myself.

You may be too late, bman.  The last I heard, Ben is out of business.  I hope this isn't true, because his parts are really nice.

To get back on track, here are a couple of comments about dcutting's revised design.

Unless the mast pole is going to be attached to the circuit board, the board will need some kind of cutout, to allow the Shapeways part to be attached to the pole.

Second, the orientation of the LED's is different from the current Shapeways part.  I can change the Shapeways part, or you can design a new one, David.  Either way is ok with me.

I definitely like the thinner circuit board.  Much better appearance.

Let's keep going down this path and see where it will lead us.

[drgw0579]

I would need 10 the first time, they if you ever did this again, 10 more later.  Am using CMRI, so  common anode preferred.

Bill Kepner

[C855B]

... For the record, my design has 5 wires going down the pole.  This is a lot of wires, but it works. ...

Somebody please tell me why this won't work. These are diodes, after all. Four wires (common + three aspects), no funky logic:

[ Guests cannot view attachments ]

[dcutting]

dcutting,

I applaud your efforts, David.  The path that I was going down was not a good one; your path has a higher probability of success.  I just have a couple of suggestions, but first I need to answer some of the other's questions.

I had spec'd out this 0402 LED:

http://www.kingbrightusa.com/product.asp?catalog_name=LED&product_id=APG1005SYC-T

The quotes that I got included engineering design costs, tooling cost for a minimum of 100 pieces, the cost of the LED's, and the cost to attach the LED's to the circuit board.  Since I don't have the skills to do any of these, I would have to pay for them.  David apparently does have these skills.

For the record, my design has 5 wires going down the pole.  This is a lot of wires, but it works.

You may be too late, bman.  The last I heard, Ben is out of business.  I hope this isn't true, because his parts are really nice.

To get back on track, here are a couple of comments about dcutting's revised design.

Unless the mast pole is going to be attached to the circuit board, the board will need some kind of cutout, to allow the Shapeways part to be attached to the pole.

Second, the orientation of the LED's is different from the current Shapeways part.  I can change the Shapeways part, or you can design a new one, David.  Either way is ok with me.

I definitely like the thinner circuit board.  Much better appearance.

Let's keep going down this path and see where it will lead us.

Thanks man! I added a new page to my website about the PRR signals, it will be updated as time goes on.

Although the 0402 might work, I already ordered 0603 LED's with an 0603 board. Everything fits well, perhaps that could be a future improvement.

I did not intend to use any shapeways parts. instead, everything will be etched brass. It is cheaper, thinner, and stronger. Another possibility is to make the cutouts in the board for the LED's round, and use the circuit board as the face plate itself, eliminating that part and giving it a more prototypical appearance.

Somebody please tell me why this won't work. These are diodes, after all. Four wires (common + three aspects), no funky logic:

(Attachment Link)

Ooooohhhh... that is an idea! I'll impement that in the next revision, unless someone tells me that it won't work. I knew there was a way to do it, just exactly how evaded me. Thanks!

-David
 

[peteski]

Somebody please tell me why this won't work. These are diodes, after all. Four wires (common + three aspects), no funky logic:

Brilliant!  I love it!
If you go back to Jim's original thread, he used my wiring diagram to simplify his original wiring.  But this is even better.  I don't know why I didn't see it that way back when I posted my circuit diagram.  And since each aspect uses 3 LEDs, there is only 1 resistor required.,

I am also leery of installing a microprocessor in the signal head. Innovative idea but . . . overly complicated and I'm not sure how it would look (the components would stick out of the board increasing the overall thickness of the target).  Plus it would not be compatible with any of the currently available signal driver circuits. Plus soldering all those components by hand is not something I would expect every modeler to do (I am one of the very few modelers who owns a hot-air jet SMD rework station).  In order for thse to be commercially viable, they would have to be sold assembled (including the hardware and software required to drive them all).

BTW Jim, I'm back from vacation. 

[peteski]

Hi guys,

Here is the next version of the signal head: https://oshpark.com/shared_projects/hR3dQ2gs

This version has holes going through the board to the back, so the 0603 LED's are inserted light-first, then soldered in. This will make it so that the .1 mm PCB will look like sheet metal. There are five contacts on the back of the board. One of them is soldered onto the pole for the common, the other four are for the center light, the pair of "stop" lights, the pair of "slow" lights, and the pair of "go" lights. they will run down the tube.

I ordered three of them, and they should be here very soon, along with my last not-so-nice attempt. 

-David

David,
I looked at the OSHPark PC board specs and those boards are all standard 1/16" (0.063") thick!  That is rather thick for what we are trying to do.  The board Jim is designing is much thinner (around 0.010").

[Mark5]

I was gonna suggest Alkem for the bridges and masts but I see he discontinued the PRR ones ...

[VonRyan]

Thanks man! I added a new page to my website about the PRR signals, it will be updated as time goes on.

Although the 0402 might work, I already ordered 0603 LED's with an 0603 board. Everything fits well, perhaps that could be a future improvement.

I did not intend to use any shapeways parts. instead, everything will be etched brass. It is cheaper, thinner, and stronger. Another possibility is to make the cutouts in the board for the LED's round, and use the circuit board as the face plate itself, eliminating that part and giving it a more prototypical appearance.

Ooooohhhh... that is an idea! I'll impement that in the next revision, unless someone tells me that it won't work. I knew there was a way to do it, just exactly how evaded me. Thanks!

-David

So will you be replacing the electrickery bits with wires? I'd rather have a loom of wires than deal with extravagant bits and bobs.

For signaling I want to use big rotary switches in a nice hefty oak-framed panel. Basic wiring all around rather than a ton of components.
My ancient brain works better with ice-cube relays and microswitch contacts on Lemaco/Fugulrex switch machines.

Even though it's a thick circuit board with bigger LEDs, I can live with it if I can get assembled/wired boards with 6" wire tails for a better price than $20 each.
But the prospect of having to form and solder on all those shades is where I say no thanks.
At that point I'd rather put out the $20 each for the .010" thick boards that have smaller LEDs that I can use with the Shapeways heads.