Hi
Didn't see this until now. Keeping it simple is best, but installing many small pots around the layout seems like overkill. The current duty cycle values are user selectable and stored in eeprom, as are the periods of the square wave for each of the different head aspects. Thought was - given the different manufactures of LED's, the inherent differences between even the same manufacture LED; let alone the color differences - it seemed easier to allow end users the ability to program the intensities to their liking. Probably not a perfect solution but one that gives users more options. The day/night thing now can be turned on either by a 'master software switch' or a 'dusk detector' on the layout somewhere.
Don't get me completely wrong as solder IS my favorite programming language!
Todd Hackett
Libby, Montana 59923
I take only pictures then leave footprints on railroad property that I know is not mine, although I treat it as such...
Coding MPUs for different luminosities by duty cycle control is cool, but I would locate a simple Day-Night toggle switch at the control panel. (could be a DCC controlled switch or relay)
Here's th' deal..... go to the LED or LEDs in question and put two inexpensive, tiny, surface mount pots in the return lead of the LED. Have the switch (arm) throw the LED return line between either pot's contacts. Now put the switch in daylight position and adjust the daylight pot for a pleasing brilliance.
Now turn out or dim the room lights to simulate night time condition and throw the switch to night time postion and adjust the night pot for a pleasing Brilliance. (let your eyes dark adapt first).
Setting pots are easier than going back into the code and reprograming an MPU should you wish to change intensities in future or change out the signal for a different type with different LEDS.
In this fashion, any color or type led can be tamed to equal or useful apparent brilliance to the eye under all conditions.
There are variations of this, of course, to suit the tastes and skill sets of the user.
Richard
If I can't fix it, I can fix it so it can't be fixed
hon30critter ctclibby I agree that many of the LEDs used in the signals that I have seen are too bright. I also agree with the implication that signals should not be so bright during night time running that they dominate the scene or interfere with your vision. They should be visible but not leave dots before your eyes.
ctclibby
I agree that many of the LEDs used in the signals that I have seen are too bright. I also agree with the implication that signals should not be so bright during night time running that they dominate the scene or interfere with your vision. They should be visible but not leave dots before your eyes.
Prototype info:
In recent years the main line where I work has received LED signal upgrades. They're nice in the day time because they're easier to see. At night they're nice to have from a distance as well. At night when you get close up, it's best to put your hand up to shield your eyes because they are overpoweringly bright. No dimmer switches on those!
For the signals at my club, the plan is to model and orient the signals along the track the way they _should_ be and instal "repeater" signals on the fascia that duplicate the signal's indication to the operator aisle. So the operator in the aisle should be able to read the signal regardless or his or her angle, and the model signal isn't canted off at some weird angle just to be visible from a certain angle from the aisle.
Chris van der Heide
My Algoma Central Railway Modeling Blog
Yes, thought of that. I was at a club and they had blind signals - train engineer could see them, but the operator could not. They turned the signal about 180 degrees so that the operator could see. I figured that would be the last resort kind of thing as I hope to position small mirrors where the operator can see the reflection. I can see that this method could cause problems with double track or control point as the signals would be backwards. We will see
One other thing that I have found on model signals. On the prototype, the engineer is looking into the signal. On our layouts, the operators are looking at the signals from the side quite a bit of the time. Because of this, I used 3mm LED's and let them stick out from the shades a little bit so the operators could see them better. I know that this may seem to be wrong and may look funny, but it sure works better. I have seen model signals that used smaller LED's that you had to look into them to see the signal aspect. This type is really hard for the operators to see as they are controlling their trains.
Elmer.
The above is my opinion, from an active and experienced Model Railroader in N scale and HO since 1961.
(Modeling Freelance, Eastern US, HO scale, in 1962, with NCE DCC for locomotive control and a stand alone LocoNet for block detection and signals.) http://waynes-trains.com/ at home, and N scale at the Club.
Yes, you got it! I guess that I could have done more xplain'n about the code and such, but I wanted to ask a general question without mudding the waters.
It seemed to me that LED's are really bright at night and wanted opinion of dimming them. Although if you look at a 1:1 signal, the green is almost blinding. Anyway, the human eye helps here and lets us see an 'apparent' brightness which can be accomplished by varying the on/off time ( duty cycle ) of LED's -or- varying the current through them. I haven't found ( and not really looking for ) a small uP with a digital to analog output, so varying the duty cycle is the way ima going. So far, having an on time of about 2.5mS constitutes the 'dim' brightness, while 8.5mS seems to be good for the 'bright'. The period of the square wave is about 20mS, which is about 50 on's per second. I played with this as some of the combinations induced a 'blinking' led where the eye could determine that the LED was blinking ... fast, but blinking.
I started down this road from looking at the different LED colors and wanting to change the intensity - Red and Yellow always seemed to be brighter. I haven't done much reading on this, but suspect that our eyes perceive those colors different. Of course, different mfg of the LED's made for different intensities given the same circuit, and there is a difference between surface mount and T1 1/2's. The T's have more plastic incapsulation and are kinda focused more than the surface mounts.
Fun Fun!
CharlieM90
Thanks for the explanation! It makes sense, but boy do I really feel like a dinosaur now. Heck, I'm one of the guys who never really figured out how to program a VCR! Stop laughing please!
I'm afraid that I will have to stick to the more basic methods of controlling LED brightness like using higher value resistors or dropping the input voltage. Not perfect but workable.
Dave
I'm just a dude with a bad back having a lot of fun with model trains, and finally building a layout!
hon30critter I am curious to know what you mean by "varying the duty cycle", and what are you referring to when you use the phrase "easy to do in code". Bear with me if the answers are obvious to you.
I am curious to know what you mean by "varying the duty cycle", and what are you referring to when you use the phrase "easy to do in code". Bear with me if the answers are obvious to you.
He's using a microprocessor of some type to alter the on/off times ("duty cycle") of the LED. "Easy to do in code" means that it's easy to write the software for the MPU to accomplish.
When adjusting your duty cycle (you didn't say if all of these LEDs are of the same wavelength or not), take into account photopic sensitivity.
The human eye will read some wavelengths of light as brigher/dimmer than others. To give them the (apparent) appearance of uniform brightness, you'll need to alter the duty cycle depending on the color wavelenght emitted by the LED. Peak is around 550nm (Green) and falls off on either side (moving towards Red or Blue).
Brief explanation here http://en.wikipedia.org/wiki/Luminosity_function
Common way to dm LEDs is to modualte the power to them on and off, since the range of dimmign with altering the current with resistors is rather limited. This is how DCC decoders to flashing and dimmign light effects, for example. The LED is actually blinked on and off, too fast for you to see the actual individual blinks, but the greater the percentage of 'on' time to off time, the brighter the LED appears. A side effectof this is that you can also drive more LEDs than the chip cna handle by a scheme called multiplexing, where you individually blink each LED on, so no more than one is actually on at any given millisecond, yet it appears to the human eye as if they were all on at the same time.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
I am curious to know what you mean by "varying the duty cycle", and what are you referring to when you use the phrase "easy to do in code". Bear with me if the answers are obvious to you. I am electronically challenged as it were, so I am drawing a bit of a blank here with your terminology.
Thanks
If you are going to operate in a night setting, then I think you should lower the light output setting of your signals if it isn't going to be hard to do.
what size resistor are you using? Also, what is the voltge of the LEDs?
Samuel A. Kelly
I can draw pictures with my keyboard!
-------- ( It's a worm)
HI All
I find myself attempting to create a signal control system and currently working on the head portion that contains the LED/LED's. I have noticed that when viewing other folks railroads that have a signal system, those signals look *really* bright. I am using LED's and varying the duty cycle will change the apparent intensity and is fairly easy to do in code. My question is - Do you think that there should also be a 'night time' mode where the signal is less brilliant than during day operations? Those layouts I was talking about above - the signal system is really really bright at night or dusk.