Good stuff guys! Let me play devil's advocate a little. Jeff points out that his example is probably a worst case scenario. He also states that the track MAY give a lttle help. Here is my question, if you are running your longest length at about 25' (which is really 50' according to Jeff) and have feeders all throughout your layout (say about 6'-10' apart) won't the track help ALOT? (in this case I am using Kato Unitrack which I have found to have excellent conductivity).
I must admit that I have a very limited understanding of electrical engineering. Jeff's resistance formula is an excellent resource. But as he pointed out, there is no way to really calculate this problem because of all the variables involved (track, etc.)
The reason I am wondering all this is because I do have a readily available supply of 18 gauge wire. I'm thinking I might give it a try and if I have problems, then just replace the 18 gauge with 16 gauge or 14 gauge. Does this seem to be a stupid idea and a waste of time? Be honest, you won't hurt my feelings and may save me a lot of trouble.
Adelie wrote:What Nigel said!The concern for our application is not as much power loss as a loss in the integrity of the DCC signal.
In essence they are the same thing. The DCC signal isn't added onto the power 'signal', it IS the power signal. The reason you might lose the DCC signal is because of the power loss. Because of the nature of the DCC signal, I expect that the decoder might still be decoding long after the level has dropped to the point that it can't drive anything.
Jeff But it's a dry heat!
- Mark
In general terms I (also an EE) disagree that using too large a wire can be as bad as too small (in this case), unless you go absurd. Also, it is important to know what the purpose of that table was, since the voltage drop through the wire is what is important, and the table doesn't tell you what the criteria were for determining the max length.
Anyway, here is a little calculation:
12V, 5A, 18 gauge wire 1.3 Ohms per 100 feet, 16 gauge .818 Ohms/100ft, 14 gauge .516 Ohms/100ft.
A 25 foot bus has 50 feet of wire (it also comes back, and the voltage drop is the same). We are ignoring the help the track MAY give here, but gaps or bad connections reduce that, so we'll be safe.
So for 18 gauge wire you lose 5A x 1.3Ohms / 2 = 3.25 Volts in the wire, leaving less than 9 at the loco. 16 gauge: loses about 2 Volts, 14 guage loses about 1.25.
These are worst case, five Amps draw at the extreme end, etc., so it probably is never that bad. But it shows that 18 is almost certainly a bad idea, 16 might be ok, 14 is a good idea, and if you are going longer you might want even more.
There is no question that this is overkill, to an expent, and I'm not looking to argue that, but the extra expense and difficulty in going to 14 gauge isn't much to ensure you don't have a problem later, when there are more sound locos, etc., and things we haven't even though of yet hanging on decoders in the tracks.
Ray Seneca Lake, Ontario, and Western R.R. (S.L.O.&W.) in HO
We'll get there sooner or later!
I was researching the gauge of wire for my DCC bus and I came across this handy table. If I read it correctly, I should use 18 gauge wire for a DCC bus up to 25' for a system using 12V 0-5 amps. My understanding is that most DCC systems use 12V 0-5 amps - is this correct? (I know you can get 8 amps too but that's overkill for me). I am curious because I recently posted a thread about this and most modelers suggested 14-16 gauge wire for a DCC bus. If the chart is correct then 14-16 gauge wire is probably too much and 18 gauge is just right. Any thoughts on this? Am i misreading the table?
Here is the chart:
WIRE GAUGE SELECTION TABLE
Circuit Amperes
Circuit Watts
Wire gauge (for length in feet)
6V
12V
3'
5'
7'
10'
15'
20'
25'
0 to 2.5
0 to 5
15
30
18
3.0
6
36
16
3.5
7
21
42
4.0
8
24
48
5.0
10
60
5.5
11
33
66
14
6.0
12
72
7.5
45
90
9.0
54
108
20
120