Author Topic: Power distribution for model railroad electronics  (Read 4680 times)

0 Members and 1 Guest are viewing this topic.

Offline rrinkerTopic starter

  • Super Contributor
  • ***
  • Posts: 2046
  • Country: us
Power distribution for model railroad electronics
« on: November 14, 2016, 08:50:15 pm »
 As part of designing and planning my next model railroad layout, I am considering methods for distribution of power. The layout will fill a 20x25 foot area of my basement, with two levels and somewhere around 300 linear feet of track. Track will be DCC powered and have its own bus. Turnouts will be servo operated, with a circuit I am designing. Signals and detection will be a mix of a controller circuit I am designing and some commercial bits for current detectors.

 There are several components of the DCC system that need between 14 and 18V DC to operate. My initial plan was to put a fairly high current 12V AC power bus around the layout and tap off with small fused full wave rectifier circuits at the point of power need. That nets about 15V DC after rectification and filtering - perfect. Now, all my OTHER circuits are 5V DC - servos, the ATTiny85 servo controllers, and the Arduino-based signal controllers. Rather than run another 5V DC bus, I was considering using inexpensive buck converter modules (with a rectifier and filter) at the point of use. 15V down to 5 seems a bit much for a linear regulator, as in an awful lot of wasted power int he form of heat. Or instead of 12VAC and rectifier modules, I could run 15V DC which saves building a whole lot of full wave rectifier modules.

 As I type this it starts to make more sense to do the multiple power circuits, but there is another possible catch. The servos can be rather noisy and I'd like to isolate their power feed from the one that drives the micros. That would mean TWO 5V power busses. Now I'm up to 3 aux power lines plus the at least two track power busses in any given area and I haven't even started on structure lighting yet. That is, assuming pulling one drop of a 5V bus for the servos and a second for the circuit controlling the servo (said circuit to have reasonable bypass capacitors in place on the power input and directly across the controller power pins) isn't good enough.

 A third option is a single 5V DC line and use a boost converter for the devices that need 14V DC (maybe 4-6 total, and they each draw no more than 150ma).

 The design goal is to minimize the number of wire runs while not using a ton of expensive regulators and converters, but maintaining reasonable enough isolation to ensure reliable operation of the various micros. The power supplies, regardless of which way it ends up getting done, are no problem, I can build or buy whatever I need.

 Any and all suggestions appreciated. One thing I absolutely do not want is dozens of individual power supplies all over the place. Otherwise, whatever makes the most sense. I'm probably overthinking this already.

 TL;DR - 1 higher voltage AC supply, with local rectifiers and regulators, 1 higher voltage DC supply with local regulators, 1 lower voltage DC supply with boost converters where needed, or multiple DC supplies of various voltages. Or some other option I haven't thought of.

 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16615
  • Country: us
  • DavidH
Re: Power distribution for model railroad electronics
« Reply #1 on: November 15, 2016, 09:25:07 am »
For the endpoints which need tight regulation, I would use a DC bus which is just high enough and low dropout linear regulators at the point of load.  Even at 5 volts, the efficiency can be competitive with a switching regulator.

The major complication I see is ground loops.  If the microcontrollers and servos are using ground as a reference and the voltage difference between grounds is significant enough, that can cause problems.
 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4136
  • Country: gb
Re: Power distribution for model railroad electronics
« Reply #2 on: November 15, 2016, 09:53:56 am »
I thought the only reason for using AC for power transmission was that you could use step up and step down transformers to utilise the efficiency of high voltage transmission lines?

If that's the case does it make sense to have AC running all over the place when you could just distribute a 'high' voltage DC rail and regulate it down with a pile of those cheap buck converters?

If voltage drop is an issue then a few smaller 15V switchers feeding 'zones' may be better?

Unless I've missed something?
 

Offline rrinkerTopic starter

  • Super Contributor
  • ***
  • Posts: 2046
  • Country: us
Re: Power distribution for model railroad electronics
« Reply #3 on: November 15, 2016, 01:43:29 pm »
 You're right, I shouldn't have even considered the AC option any longer. That harkens back to a different type of point motor, plus the fact that some of the commercial items that will be connected have multiple input options, one that will accept AC and one that will accept DC. So there is no reason to run an AC power bus at all.
 Likely there will have to be multiple feeds rather than one huge length of power bus, even with the relatively low current draw, losses over one long line would probably be too much, even with the source located in the middle. The track power is likewise divided into multiple sections with multiple power supplies for the same reason.
 Ground loop should not be a problem. The controlling micro and the micro for the servos controlled by it, will be fed from the same power tap, so both micros involved plus other supporting circuitry like the 74HC595 shift registers to expand the IO, will all have a common power source. One Arduino controller will not talk directly to another one, so if they are on separate power supplies, there should be no issues as there are no common wires between them. They will all talk on an RS484 bus back to the central computer, with some shielded twisted pair, the screen of which will only be connected at one point, not on each board.
 The basement is being refinished, with new mains circuits as well, and you can be sure I will be checking to make sure none of the outlets is wired backwards. Also, all power to the railroad will come from one circuit. Things not electrically attached, like the lighting, will be other circuits. I'm having a separate subpanel put in the basement since running wires above the fire rated ceiling int he garage is rather difficult plus it makes every run from breaker to outlet or light fixture 25 feet shorter.

 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4136
  • Country: gb
Re: Power distribution for model railroad electronics
« Reply #4 on: November 15, 2016, 02:36:04 pm »
looking forward to seeing the finished layout.
 

Offline tszaboo

  • Super Contributor
  • ***
  • Posts: 7384
  • Country: nl
  • Current job: ATEX product design
Re: Power distribution for model railroad electronics
« Reply #5 on: November 15, 2016, 03:22:20 pm »
I would design / find a small breakout board for Cat 5e ethernet cables. The board has this screwdriver terminal block connector. Cat5e It has 8 wires, capable of some 0.6 A. You can buy the cables at the local computer store, at the required length. You dont have wiring fault. And twisted. Other than that, including a small power supply, on the board should not be an issue.
 

Offline rrinkerTopic starter

  • Super Contributor
  • ***
  • Posts: 2046
  • Country: us
Re: Power distribution for model railroad electronics
« Reply #6 on: November 15, 2016, 04:40:19 pm »
 Not sure what you mean here - use the CAT 5 as the taps along the power bus to drop to each board, or use CAT5 AS the power bus? I will be using RJ45 connectors for the actual signals since it's an easy way to make them removable. Making cables is no problem - being in computer networking since the days of Arcnet and RG58 thinnet Ethernet I can make pretty much any sort of cable, and I already have the tools - the DCC system uses 6p/6c RJ11 to interconnect the components. ANd I have a part spool of CAT5 stranded just laying around, only need to buy the ends.



 

Offline rrinkerTopic starter

  • Super Contributor
  • ***
  • Posts: 2046
  • Country: us
Re: Power distribution for model railroad electronics
« Reply #7 on: November 15, 2016, 04:43:41 pm »
looking forward to seeing the finished layout.

 Hopefully you plan to hang around on this earth as long as I do, because it will NEVER be finished. It never is. However, once there is actually something interesting to show (and I start updating my web site to show it) I will certainly post a link for anyone interested to follow along on my construction progress. These electronic bits I am working on will be done much sooner and I'll show that stuff off here.

 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19505
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Power distribution for model railroad electronics
« Reply #8 on: November 15, 2016, 05:02:11 pm »
looking forward to seeing the finished layout.
Hopefully you plan to hang around on this earth as long as I do, because it will NEVER be finished. It never is.

Just so.

There are three rather good layouts in the Pendon Museum, near Oxford; they've been well over half a century in construction.

I saw pictures of the Dartmoor Scene back in the late 60s/early 70s, and was stunned. Many decades later I finally visited there and both I and my daughter were stunned. But the Vale Scene is even better.

Pictures won't do it justice, but this video is a reasonable introduction:
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline C

  • Super Contributor
  • ***
  • Posts: 1346
  • Country: us
Re: Power distribution for model railroad electronics
« Reply #9 on: November 15, 2016, 05:32:55 pm »

You may be missing something important in your thinking.

Think of where DCC came from.
DC power on track with some signal tones to supply control of engine. The problem with this is loss of signal at times. DCC just makes the signals so large that the signal can be used as power. A High pass filter = control signals, a Low pass filter = DC power. In the engine, space limits size of filters. Still good idea to think Signal first.

With your servo's, cheap and size = more noise. Do not think you have a size problem under track so you could add a cap to remove some of the noise. An inductor before cap can remove more noise and can be as simple as some wire rapped around something that is magnetic, a steel bolt for example.
if needed a cmos logic gate or differential input to control servo. This reduces problem with logic 0 return not being same at both ends of cable.  And in process reduces noise on power. With cmos or differential, you could use a common mode choke to contain more noise. A linear regulator per servo might be cheaper with more noise reduction an benefit of higher power source variation. 
This is fix/reduce problem areas at source instead of letting problem grow to a level that is more costly to fix.

If this next model railroad layout is going to be around for a while, you might think of making a large number of blocks as far as track is wired. You can reduce number of blocks for now by jumpers under track that can be removed at later date.
Do what is cheap an easy now to allow best future at lower cost.

 

Offline CJay

  • Super Contributor
  • ***
  • Posts: 4136
  • Country: gb
Re: Power distribution for model railroad electronics
« Reply #10 on: November 15, 2016, 05:33:21 pm »
looking forward to seeing the finished layout.

 Hopefully you plan to hang around on this earth as long as I do, because it will NEVER be finished. It never is. However, once there is actually something interesting to show (and I start updating my web site to show it) I will certainly post a link for anyone interested to follow along on my construction progress. These electronic bits I am working on will be done much sooner and I'll show that stuff off here.

I realised my mistake shortly after hitting 'Post' having had relatives and their friends who were model railroad enthusiasts, I should rephrase it to 'Looking forward to seeing the progress towards a finished layout'  ^-^
 

Offline tszaboo

  • Super Contributor
  • ***
  • Posts: 7384
  • Country: nl
  • Current job: ATEX product design
Re: Power distribution for model railroad electronics
« Reply #11 on: November 15, 2016, 06:14:33 pm »
Not sure what you mean here - use the CAT 5 as the taps along the power bus to drop to each board, or use CAT5 AS the power bus?
I would use CAT5 with RJ45 as the power bus, and the data bus. Topology is irrelevant, can be star, daisy chain, combination of two, etc.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf