Keep in mind on most decoders, ditch lights are done using 2 functions, one for the left ditch light, one for the right ditch light, and when set up in the decoder, the decoder just alternates which one is on to make them flash. And not all railroads flash the ditch lights - so if you were to model one fo those railroads, you could hook the citch lights up like BLI did the numebr boards - two LEDs in series, and one resistor and one function.
Ditch lights are far too new for me, too. In fact, based on an employee rulebook I have from close to my specific year, they just started even turning on the headlights during the day in my era. ANd no flashers on the cab roof, no Mars lights, nothing - makes installs easy for me. And means any extra functiosn a decoder might have, I can do some of the more esoteric lighting, like cab lights or ground lights. One of these days - so far I've done nothing but basic headlights.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
Thanks, Randy, for staying with me all through all of this. I get what you are saying about the electronics, so I will put this issue to rest. Like you, I look for consistency so this back and forth between positive common and negative common really threw me at first.
By the way, I got so intrigued by the workings of the ditch lights that I started to wish that I owned a diesel locomotive with ditch lights. But when I looked up to start date for ditch lights, it turned out to be in the early 90s, much too late for my modeling era. Oh well.
Rich
Alton Junction
You'd have to ask whatever engineer at BLI decided to use positive logic for some functions and negative logic for others.
I shoudl draw it out, but I suspect if they used a positive common for the headlight it might have allowed some sort of sneak path with the ditch light, or maybe it just made it easier to lay out the board traces - stuff liek that comes into play with electronic design too. The actual use of active high or active low is really just a software thing in the microcontroller. A given pin can be set high or low. Which represents the "on" state from the end user perspective is entirely a software choice: If F0=On, Set (HeadlightPin) = Low or If F0=On, Set (HeadlightPin) = High. You, runnign the train, just know that trurning F0 on turns ont he headlight. Inside the loco - the code could eb making the pin go low, or go high. Other factors drive which the developers choose.
Though that really is a mess - I'm too OCD to have half the things work one way and half work the other. I thought MTH was bad, but at least they consistently use a negative common. Seems like BLI used whatever makes the circuit easier. Since it's all the fashion among the youngsters these days - I'm offended. This all over the place design offends my engineering sensibilities. And furthers my understanding of BLI as a bunch of penny-pinchers, as worst case wirign the headlight the same way as the number board lights would have required 2 vias in the PCB. If they made 10,000 of those locos, I'd be amazed if it added even a penny to the cost of each one. Just speculating...
rrinker Why would +3.3 to gnd light the headlight? The headlight + side is connected to +fl, not +3.3.
Why would +3.3 to gnd light the headlight? The headlight + side is connected to +fl, not +3.3.
Yes, the diode test in a meter tends to use less than the suppyl voltage - in fact the ones that use a 9V are the elast likely to light LEDs in the diode test mode because the first thing they do it knock the 9V down with a regulator since the multimeter chipset generally does not run on 9V. My two better handhelds use AAA in the smalelr and AA in the alrger, 4 batteries in each. They can light a white LED, which is usually 3V or more. If yours barely lit a red/green bicolor LED, it likely won't work with a white LED, or it will REALLY be dim. For a basic diode test you don't need much voltage - a silicon diode is .7V. A little more won't hurt, and then works with the hight forward voltage of an LED. A suspect there may be some sort of current source used, to keep from blowing LEDs - which is why the el cheapo meters can do a diode test but can't light LEDs. Extra cost for the extra circuitry and all.
Randy, I can't help myself. I am still at it. A couple of questions for you (and the OP).
One, I think that I can see how the OP approached this. Using +3.3v as the common, +3.3v to dit illuminates left ditch light. Next, +3.3v to gnd illuminates both ditch lights. Then, +3.3v to num illuminates both number boards.
That leaves the headlight and the right ditch light. The OP wired +fl to gnd to illuminate the headlight. But wouldn't +3.3v to fl also illuminate the headlight? It would seem so, leaving the right ditch light to be wired. So, the OP wired +dit to gnd to illuminate the right ditch light. I assume that worked as a result of the three-state logic circuit. Is that correct?
Two, I pulled out my digital multimeter and dialed the diode test setting. I tested some bi-polar LEDs after first testing them with resistors attached on a 9v battery. On the 9v battery, the red/green colors lit brightly, but on the multimeter diode test setting with the resistors bypassed, the bi-polar LEDs lit, but barely.
That said, it was interesting to use the diode test setting which I had not been aware of on the multimeter. My multimeter is powered by a 9v battery, but I assume that the voltage is far lower on the diode test setting. What say you?
Love the sounds and the lighting effects.
Did you happen to take any photos when you rewired the lighting?
This project fascinates me, and I would love to see photos of how you rewired it. I am still trying to get my head around the whole thing.
Here is the video of the final project :)
For most of the LEDs on that board it is.
Again, rewiring it can take many forms. The most obvious, though maybe not the neatest, is to siply cut traces alongside each LED - now all you have is a bunch of individual LEDs soldered to a board, but not actually connected to anything. You can then wire them up as if you had a bunch of wired LEDs and were installing them in a loco that had none to begin with.
Since a few of the LEDs are wired the "right way" for DCC - positive 'common' and the signal side goes negative to turn on - you can use those as-is, and just cut traces to rewire the ones that are 'backwards' for DCC. If I had a board in hand, perhaps further analysis would reveal a 'simpler' (but really - since it requires a whole lot more thought on circuit theory - it's a more complex solution) way to cut fewer traces and make it work. But just isolating each LED on its own 'island' is the easiest - it leaves you with no circuit to figure out.
MTH is prettu consistently common negative, meaning all their light boards are completely backwards for conversion to DCC. Luckily my FAs really only have a headlight. Unless there's room behind the light pipe to change the class lights to be red/green/white/ogg instead of the factory red/green/off, I'm not going to bother even hooking them up. But the more modern power, with headlights, warning beacons, ditch lights, etc - it's not that easy a job to swap to a standard DCC decoder.
Hmm, it's possible to build a simple inverter circuit - and probably room inside those bigger locos once you strip out the FCS stuff - maybe a little circuit that goes between the DCC decoder and hooks to the stock MTH light boards..... ok, no, I have enough projects to do that actually apply to equipment I own.
rrinker Rather than one wire linking all the points that are common, like the gnd pin, there might be 2, 3, or more traces coming from the connector pin labeled gnd and goign to the various LEDs. Have to try and find a more zoomed-in picture than the one you posted - I'm not buying an AC6000 just so I can figure this out
Rather than one wire linking all the points that are common, like the gnd pin, there might be 2, 3, or more traces coming from the connector pin labeled gnd and goign to the various LEDs.
Have to try and find a more zoomed-in picture than the one you posted - I'm not buying an AC6000 just so I can figure this out
It is disappointing that you won't buy a GE AC6000 to figure this all out. I had hoped you would.
I'll close out this discussion with this comment. Although your post about the tri-state circuit was the breakthrough for wolf10851, I am amazed that he was able to take that comment alone and successfully rewire the lighting to the ESU decoder including cutting the traces on the bottom of that lighting board.
At this point, knowing what I know and what I don't know, I am not at all sure that I could rewire the lighting on that loco to a different brand of decoder. I wonder how many others could.
Almost certainly a double-sided board, so if you were running into issues with wires crossing over one another, that would take care of it. I haven't tried layoing out something based ont he schematic of how they would all be hooked together. But double sided PCBs take care of issues like that. There's probably plenty of ways to run traces arounfthe outside as well - say the LEDs for the class lights aren;t right at the very edge of the board, a trace can then run around them to get to the headlight on the front center of the board. Rather than one wire linking all the points that are common, like the gnd pin, there might be 2, 3, or more traces coming from the connector pin labeled gnd and goign to the various LEDs.
rrinker Well, you can easily strip out the headlight and number boards, there's nothing special there: +fl---->|----gnd is the headlight (backwards from a standard decoder, function is the +) +3.3---->|----num is the number boards (actually two LEDs in parallel most likely, decoder switches num to ground like an ordinary DCC decoder function) Then the ditch lights as described: +3.3---->|---dit left ditch light dit---->|--gnd right ditch light So the ditch lights are pretty close to what I guessed +3.3----->|----dit----->|----gnd dit tri-stated, neither ditch light is on. dit goes to ground, left light on, dit goes to +3.3, right light is on.
Well, you can easily strip out the headlight and number boards, there's nothing special there:
+fl---->|----gnd is the headlight (backwards from a standard decoder, function is the +)
+3.3---->|----num is the number boards (actually two LEDs in parallel most likely, decoder switches num to ground like an ordinary DCC decoder function)
Then the ditch lights as described:
+3.3---->|---dit left ditch light
dit---->|--gnd right ditch light
So the ditch lights are pretty close to what I guessed
+3.3----->|----dit----->|----gnd
dit tri-stated, neither ditch light is on. dit goes to ground, left light on, dit goes to +3.3, right light is on.
The confusing thing for a non-engineer is probably the whole concept of tri-state inputs and outputs, but you can basically think of it as an open circuit, not connected to either + or -, it's just done in the internal semiconductor, not by physically disconnecting the wire to the pin of the device.
Be careful if you try to replicate this to see how it works - the voltage has to be well under the sum of the voltage drop of the two LEDs - say you used 12V instead of 3.3 - you'd actually just blow out both LEDs since 12V is well above the total voltage drop across the two. The LEDs don't blow with 3.3V because it isn't enough voltage to make it through the two series LEDs. An ulta low current meter might detect a slight current flow, in the microamps, but far too low to make either LED light up.
wolf10851 the wiring is pretty crazy and it appears that the ditch lights are indeed using a tri-state circuit. for anyone in the future that is trying to figure this out this is what I found +3.3v to dit illuminates left ditch light +3.3v to gnd illuminates both ditch lights +3.3v to num illuminates both number boards +fl to gnd illuminates headlight +dit to gnd illuminates right ditch light the way I resolved this was to cut all the traces on the BOTTOM of the light board on the top of the light board I left all the traces in tact. I was able to keep the dit wire then I added a seperate wire to the rest of the leads of the LEDs. Then I wired everything normally to the ESU decoder.
the wiring is pretty crazy and it appears that the ditch lights are indeed using a tri-state circuit.
for anyone in the future that is trying to figure this out this is what I found
+3.3v to dit illuminates left ditch light
+3.3v to gnd illuminates both ditch lights
+3.3v to num illuminates both number boards
+fl to gnd illuminates headlight
+dit to gnd illuminates right ditch light
the way I resolved this was to cut all the traces on the BOTTOM of the light board on the top of the light board I left all the traces in tact. I was able to keep the dit wire then I added a seperate wire to the rest of the leads of the LEDs. Then I wired everything normally to the ESU decoder.
wow Randy do you work for BLI? lol Your suggestion of a multimeter with a diode checker was great and helped me figure this out! the wiring is pretty crazy and it appears that the ditch lights are indeed using a tri-state circuit.
Again Randy thank you for your suggestion without that I would have never been able to figure this out!
It's possible they did some non-standard thing with the ditch lights as well - it IS their own hardware after all, not necessarily a standard DCC decoder.
See, the little microcontrollers used in these things actually have three states on most of their wires. On, Off, and tri-state - which is effectively disconnected like you took the wire off.
So if you took a pair of LEDs and wired them
+3.3--->|----D----->|---GND
When 'D' is at 0 volts, the left LED lights. When 'D' is at +3.3V (most modern micros are 3.3V, not 5V any more), the right LED lights. When 'D' is in a tri=-state mode, it's liek the wire is off, and you have 2 white LEDs in series across 3.3V - not enough to light them, since each needs over 2 volts.
(I left out any resistors for simplicity)
So who knows - you'd have to remove the light board and examine the circuit traces and see what connects to what wire. Maybe they did it the easy way - one control wire per LED, maybe they did some sort of simple multiplexing like my example. Only BLI knows for sure.
One way to safely test it would be to use a multimeter that has a diode test setting strong enough to light LEDs - not all do. You cxould then probe the connections on the board (disconnect the plugs from the decoder first!) and see what combinations light up which LEDs. A multimeter in diode test limits the current so is safe to use in this manner.
This post has really triggered my intellectual curiosity. I do not own any Paragon 3 decoders, but I have been recently installing decoders in several locomotives, and so wiring issues are fresh on my mind.
When I look at a wiring diagram for a Paragon 3 decoder with front light, lighted numberboards and ditch lights, it seems to me that the "gnd" wire is the other ditch light.
The front light will have a common wire and a negative wire. The two numberboards, lights both on or off at the same time, will join into one common wire and one negative wire, That leaves the two ditch lights which need to have the ability to flash alternately, so you need one joined common wire and two separate negative wires.
So, there are your five wires, a single joined common, a negative front light, a paired negative numberboard lights, and two separate negative ditch lights.
I stand prepared to be corrected, but that is how I see it.
I have a BLI AC6000 that I am swapping out the paragon 3 decoder to an ESU v5 decoder I have a question on the lights that I need some help with if anyone out there knows. in the front of the locomotive I have 2 LEDs for number boards, an LED for the headlight, and 2 ditchlights that oscilate all of those lights are on the same board and there are 5 wires coming off the board. the wires are in this order
+3.3v black wire coming off of it
dit red wire coming off of it (these 2 go to a common 2 pin connector)
fl black wire coming off of it
gnd grey wire coming off of it (these 2 go to a common 4 pin connector with the other 2 wires going to the rear light)
num brown wire coming off of it
now my assumption is that all of these are a common positive with individual negative wires. what is throwing me off though is I would assume that the ditch lights would need seperate wires since they do oscilate when the horn sounds and all other times they both are steady burning but I do not see a second wire for the ditch light only 1 wire. Also if it is a common hot then what is the gnd wire for since normally that would be ground. Has anyone done this swap or have any insight on what wires I should be connecting to the ESU decoder and where to hook them up?