Joe (or someone else):
Can someone please draw a simple picture of this, so I can see the whole chain from start to end. I really like graphical things… Makes it so easy for a newbee like me to understand how it all connects. I’m not good at all when it comes to connecting cables. [xx(]
Mark:
My friend Charlie Comstock, who is an EE, has written up a nice explanation of how the decoders do their thing. In short, it’s not by doing a lot of direct voltage variation because dropping the voltage dramatically means the components would need to disapate the unused voltage. This means you would need some hefty transistors, making the decoders far larger.
I’ll dig out Charlie’s nice explanation and post it on here. It’s on my computer at home so it will be this evening when I make the post.
Thanks, Joe! I’ll look forward to reading it tomorrow. I’m in Amarillo for a couple of weeks, so I don’t have web access after working hours.
Is this what we are talking about here? I found this picture on the net.
So the BUS WIRE are two cables that runs wherever your track runs? And at every 36 inch or so there are two TRACK FEEDERS soldered to the track?
Electrolove- That is exactly what we are talking about.
Though anotheer option is to use terminal strips for your bus wire. Instead of one wire running the entire length of the layout with the feeders soldered to it. You place terminal strips throughout the layout. Then you connect each terminal strip with your bus wire and run your feeders from the terminal strip. These would be a benefit if you ever plan on moving the layout because you would be able to disconnect the bus wire. For a picture go here-http://www.radioshack.com/product.asp?catalog_name=CTLG&product_id=274-658#
Andrew
george745:
Yes, must be a lot easier to wire with terminal strips. Thanks for pointing me to the page.
Can I use something else instead of the terminal strip to get the same function, maybe insulation displacement connectors? Or is this completely wrong?
El, yes insulation displacement connectors work a treat. Just be sure you get the proper size to connect your large gage bus wire to the smaller gage feeders without a: cutting so much bus it weakens the wire,or b: makes too little contact with the smaller feeder. The generic term here for these connectors is Scotchlock and they come in many different cofigurations usually through an electric supply house. like the terminal strips they will work but they are an expensive option for most of us.
Just in case I haven’t said so before I really appreciate the time so many of you spend patiently explaining all of this to us newbies. This is a great community to visit! J.R.
Joe’s got it right - decoders do not work by reducing motor voltage directly,t he way a typical rheostat throttle does it. The ‘reduce voltage’ method is simple but horribly inefficient, especially at lower speeds. All the excess voltage gets wasted as heat - not good in the tight confines of the innards of a locomotive. For example, if you had a 1 amp motor, and at slow speed it ran on 2 volts, with a track voltage of 14 volts. The comonents would be dissipating 12 of those volts at a 1 amp current level - that’s TWELVE WATTS. Plenty of heat to melt plastic shells - not to mention the size of the heat sink required to keep the component itself from frying. Maybe it would fit in G scale…
DCC decoders do what’s called Pulse Width Modulation - PWM. They feed the motor full voltage pulses - minus the internal component drop which is about .6 volt or less. The longer the pulse, the more the apparent voltage and the faster the motor turns. Now, we’ve all heard pulse power is potentially damaging - indeed it is, for sensitive coreless motors, but this is why better decoders have ‘high frequency’ or ‘silent running’ or any of other terms that mean the same thing. This is also why locos with cheaper or older decoders ‘buzz’, they use lower frequency pulses and the motor vibrates slightly. High frequency pulses, often above human hearing frequencies, don’t cause this problem. They DO have a problem getting a motor to turn at slow speed sometimes, which is why some decoder have a ‘kick start’ function that helps get the motor turning at low speed.
The advantage of the PWM drive is that the component only ever has to dissipate the internal voltage drop of the component. Very little excess is wasted as heat, enabling a smaller component to do the job without melting the plastic.
–Randy
WARNING!
Don’t drill extra large holes for feeder wires from the rails. Drill a hole only slightly larger than the wire so the wire is snug.
If the holes are too big, when you flood the track for ballasting the water/glue will drip through the MANY holes onto the floor or a lower level. Trust me on this (and don’t ask how I know).
Wiring is one of my weak suits, but what is the rule of thumb when wiring feeders? Is it a certain colour to the outside rail?
You can use whatever colors you like. But be consistent throughout your layout. That will avoid errors and shorts. I bought 500’ of thermostat wire which happens to be red and white. White is outside for me and red is inside. One note of caution though, “inside” and “outside” are relative, so A and B might be more appropriate.
-Tom
I think the suggestion of wiring feeders every 3 feet AND soldering the rail joiners is severe overkill. This would put a feeder no more than 18" from any locomotive. Waaaay over kill in my opinion.
When calculating the voltage drop on your system, when you’re combining feeders with rails, you have a very complicated calculation to make. Ever hear of a load-flow calculation? Probably not unless yo’re in the electric power industry. Until powerful PCs came along these calculations were done on mainframes. You’ve got multiple parallel paths to take into account, all with varying effects based upon where the loco is, feeder size & length, etc. The end result is that according to all of the calculations you’re likely to make here, you’re over-estimating the voltage drop.
DCC doesn’t like poor connections because ANY fluctuation can look like signals on the track, and screw up the decoder.
If you’re soldering the track, placing a feeder every 8 - 10 feet should be plenty. This will put a locomotive no more than 5’ from any feeder, plus you’ve got the current contribution from BOTH directions, which will make it look like the locomotive is even closer to a feeder, like maybe 3’?
Mark in Utah
Finally, here’s the promised explanation my friend Charlie Comstock gave to me about how DCC decoders control motor speed without using heavy components on a decoder.
Charlie Comstock running trains on the Siskiyou Line
Charlie’s explanation is a bit technical, but if you know just a little something about electronics, you should be able to follow it.
============================================================================
The motor always receives the voltage (for a given track peak to peak voltage) regardless of the speed step. What changes is the percentage of time the motor receives the voltage.
For each time period the motor spends sometime driven and some time coasting.
v v v v v v
-------_-------_-------_-------_-------_ - nearly full throttle
----____----____----____----____----____ - half throttle
--______--______--______--______--______ - 25% throttle
________________________________________ - throttle OFFThe time from “v” to “v” varies depending on the motor drive frequency. A 100hz drive frequency would have 10ms (1/100th of a second) between each “v”. For silent drive decoders this might be 40uS (just a guess) resulting in 25khz drive which is a higher frequency than humans can hear (but might annoy the family mutt).
The voltage applied to the motor when it is being driven will be slightly less (I’d guess about 1.5V or two bipolar junction voltage drops) than full peak to peak track voltage. One junction drop to rectify the power coming in on the rails and smooth it a bit with a small capacitor, Another junction for the transistor then turns the power to the motor on/off.
Back emf uses the time between the motor drive periods (that part of each drive cycle when the motor is coasting, not driven, to turn the motor into a gener
I think the main reason for running feeders to every rail section is reliability.
No rail section will ever be without power, and you don’ t need to rely on rail joiners at all.
Unsoldered rail joints can move with temperature and humidity differences in the layout room, meaning less chance of mysterious kinks showing up in your track later.
Don’t underestimate this last point. I’ve seen layout track kink badly because it was all laid with soldered rail joints in the winter. When the temperature and humidity change of summer came (yes, we’re talking about a heated/air conditioned basement), rail popped off the flex track ties in several places because the track could not “breathe”.
The distance to the nearest feeder with Nickel Silver track is secondary – although not insignificant if you run high amp boosters and space your rail feeders out 15-20 feet or more.
The practical considerations above means, for me at least, feeders to every rail section and not soldering rail joiners (except in a very few special cases) will give you more reliable trackwork.
Temperature and humidity differences in the layout room, do you guys think that is a problem even here in Sweden? Maybe a hard question for you to answer. The reason I ask is because at the moment I’m not really sure what the best thing is. Solered rail joiners or not. Seems that there are different opinions (and that’s good) but it makes it hard for a newbee like me. So I 'm a little lost, as usual [:D]. I want to make the electrical side work, and I don’t want to bother with it again after that. If I solder every rail joiner and solder track feeders to every rail section. Is that good or bad?
If you have a feeder to every rail section, the soldered joiners are more to keep alignment that electrical connectivity (thought they are still performing that as well). Depending on how things are constructed, they are going to move through their life. I have not done the experiment to prove this to myself, but I think it is not the track expanding and contracting, but the ‘stuff’ under it. I have an unproven theory that foam with a flexible mount to the wood below it (I’m thinking bolts with big washers through slightly oversize holes) would probably not change much with temperature and humidity, so you could probably solder every joint. I’d read that not soldering in the straight runs and allows enough movement. My layout is currently paper, so this is a lot of conjecture, but I’m tempted to go overboard on the feeders, and leave some joints flexible.
Jeff
Joe,
I heartily agree that it’s a good odea to run a feeder to every track section if you’re only using joiners. Eventually the joiners will fail electrically, either through loosening up or corrosion. Remember, the joiner / rail interface is between two disimilar metals, which will eventually result in corrosion and failure. True, they’re CLOSE to the same, but not quite the same.
My point is that when you solder the joiners (minimal chance of electrical failure) AND install feeders for every track section (3’) you’ve got severe overkill. You’re better off installing a heavier bus and stiffer power supplies.
Mark in Utah
I think that because of the properties of the track as a conductor the feeders are superior to the joiners to minimize voltage drop around the layout. On most layouts this is probably not that significant, but on a larger layout it might be. Joe had some numbers to support that earlier in the thread.
Jeff
You can try soldering rail joiners but the safer approach is to not, and just run feeders to every track section.
Not soldering rail joiners and running feeders to every rail section has the least risk, and is the general method I have used on the Siskiyou Line. We’re at 14 years now and doing well.
Now I do solder rail joiners in some special cases:
If the rail section is under a foot in length, I solder the railjoiner to the next section and don’t run feeders to the short section. There are maybe 6 of these situations on the Siskiyou Line in 1200 feet of track.
Because turnouts are short track sections, I solder the points end rail joiners to the adjoining track (remember the rule – always power feed a turnout from the points end). The Siskiyou Line has 122 turnouts, so this is a number of soldered rail joiners.
I solder two pieces of flex track together with rail joiners if the track is going to be on a curve so I don’t get kinks. This gives me, in effect, one 6 foot section of flextrack I can use to make a smooth curve.
All other rail joiners on the layout are left unsoldered and I leave a small gap of about 1/16" to allow the track to expand and not “pop” the rails off the ties.
Remember, my motivation for doing things this way is electrical reliabilty and allowing the track to expand/contract with the layout as temperature and humidity changes occur through the year.