Power Bus Wire

[RobjUK]

8 hours ago, Crosland said:

With DCC you want to minimise the voltage drop for consistent running at all parts of the layout but most of the time the loads (the loco) are distributed around the layout and it's unlikely you will be drawing the whole 5 A from the booster to the farthest part of the layout. Under normal running voltage drop is not a great concern.

 

When it becomes critical is if there is a short circuit. If the wiring resistance is too high then the booster will not see a high enough fault current and may sit there slowly cooking itslef and/or other components. Hence the "coin" or "screwdriver" test. A deliberate short at the farthest point from the booster should cause an immediate shutdown.

 

Sorry, but to me that is the wrong approach.

 

Basing it on the 5A controller or booster you mention, if that trips at eg. 7.5, then the cable resistance can drop over 12V at 7,5A and still cause a trip.

 

That same cable resistance would be dropping 8V or more at the rated 5A load of the controller, which is crazy and would seriously mess up the operation of locomotives.

 

 

The wiring needs to be much, much lower resistance than what's needed to simply cause a trip to ensure proper performance at full load - eg. aim for no more that 1 - 2V drop at maximum current, meaning less than a quarter the resistance of the "trip capable" cable, so at least four times the cross sectional area that you "trip test" would require.

 

 

[Edwin_m]

The worst case is when a short circuit causes the current to be just less than the current at which the command station trips out the power.  If the limit is 5 amps and the track voltage is 15 volts then 75 watts is disappearing into heat somewhere on the layout - that's a large-ish old-fashioned light bulb and we know how hot they would get.  Some will be distributed along the wiring but most will probably be at the point of the short-circuit especially if it's not a perfect short but has a bit of resistance, such as a pickup in a loco as someone mentioned above (that's happened to me).  If you have diode block detectors it also makes things worse, because they contribute about 1.5 volts to voltage drop. 

 

I'd repeat again the good advice about the coin test.  A could of workarounds that might help if sections are a bit marginal are to or reduce the command station trip current if you can or fit circuit breakers with a lower setting, or counter-intuitively to increase the track voltage.  The latter causes more current to flow in the short circuit situation so making it more likely to trip out (though if it doesn't it will produce more heat than otherwise), and the former makes the system trip out at a lower current. 

[Crosland]

5 hours ago, RobjUK said:

 

Sorry, but to me that is the wrong approach.

 

Basing it on the 5A controller or booster you mention, if that trips at e.g. 7.5, then the cable resistance can drop over 12V at 7,5A and still cause a trip.

 

 

5 hours ago, RobjUK said:

 

That same cable resistance would be dropping 8V or more at the rated 5A load of the controller, which is crazy and would seriously mess up the operation of locomotives.

 

Agreed, but as I said, it's very unlikely that any significant length of the bus is ever carrying anything the full 5 A, except under overload or fault conditions. It would be a strange layout that grouped every single loco at the farthest extreme of the layout. As you get further from the booster the current in the bus reduces as there is less load downstream.
 

5 hours ago, RobjUK said:

The wiring needs to be much, much lower resistance than what's needed to simply cause a trip to ensure proper performance at full load - eg. aim for no more that 1 - 2V drop at maximum current, meaning less than a quarter the resistance of the "trip capable" cable, so at least four times the cross sectional area that you "trip test" would require.

 

Did you read the bit where I said "minimise the voltage drop"?

 

I was giving an extreme example as an illustration of the issues.

[Crosland]

3 hours ago, Edwin_m said:

The worst case is when a short circuit causes the current to be just less than the current at which the command station trips out the power.  If the limit is 5 amps and the track voltage is 15 volts then 75 watts is disappearing into heat somewhere on the layout - that's a large-ish old-fashioned light bulb and we know how hot they would get.  Some will be distributed along the wiring but most will probably be at the point of the short-circuit

 

A short circuit, by definition, has no resistance and cannot dissipate any power.

 

3 hours ago, Edwin_m said:

especially if it's not a perfect short but has a bit of resistance, such as a pickup in a loco as someone mentioned above 

 

It's should still be a very low resistance and will not dissipate significant power. If the resistance is high enough to prevent the booster from tripping and dissipates enough power to melt a loco (yes I have seen the pictures) then there was something very wrong with the loco in the first place. You can't design around this, no matter how fat your bus wires are.

 

To be fit for purpose on a DCC layout a loco (any accessory in fact) needs to be able to pass the full booster trip current through it's internal wiring, pickups, etc., without damage, and allow the booster to trip in the case of this kind of internal fault. This point is completely ignored by all users and manufacturers as far as I can tell. It comes down to diminishing returns. How likely is such a fault in the first place?

 

[Crosland]

12 hours ago, Dungrange said:

Does that therefore mean that the wiring should be designed for the "fault current" rather than the capacity of the Booster?  That is, instead of saying "I have a 5 Amp booster" and using 5 Amps to determine suitable wiring, I should seek to find out at what current draw the command station or circuit breaker shuts down.   I'm assuming that a 5 Amp system doesn't necessarily shut down at 5.01 Amps?

 

No, that's the minimum requirement in an extreme example to ensure the booster trips. You need to minimise the voltage drop as much as reasonably possible. One inch square copper bus bars would be taking it a bit far, for example

[Edwin_m]

11 hours ago, Crosland said:

 

A short circuit, by definition, has no resistance and cannot dissipate any power.

 

 

It's should still be a very low resistance and will not dissipate significant power. If the resistance is high enough to prevent the booster from tripping and dissipates enough power to melt a loco (yes I have seen the pictures) then there was something very wrong with the loco in the first place. You can't design around this, no matter how fat your bus wires are.

 

To be fit for purpose on a DCC layout a loco (any accessory in fact) needs to be able to pass the full booster trip current through it's internal wiring, pickups, etc., without damage, and allow the booster to trip in the case of this kind of internal fault. This point is completely ignored by all users and manufacturers as far as I can tell. It comes down to diminishing returns. How likely is such a fault in the first place?

 

A quick search shows definitions of short circuit involving low resistance, not no resistance.  There will always be some resistance. 

 

The probem I had was a loco with a plastic chassis where the pickups on the two bogies were linked to the motor by thinnish wires which could probably take about 1 amp.  The pickups themselves would also have had some resistance (it wasn't clear from the remains where the heat was coming from).  It derailed such that the left hand wheels of one bogie touched the right hand rail while the other bogie stayed on the track.  You can't design the internal wiring of a loco to take maximum DCC fault current - particularly as the faul current depends on the command station - so you need to be sure the cutout will work quickly enough to avoid overheating, which is how cables fail if they are used above their rated current.   DCC command stations cut out very quickly as long as the trip current is reached. 

Edited July 26, 2019 by Edwin_m

[newbryford]

On 24/07/2019 at 12:10, Pete the Elaner said:

 

Please could you explain your reasoning for the comment about solder?

 

Because single core - when stripped can have a coating on it that the solder does not break through properly.

I have seen joints where the stranded dropper wire has been wrapped around the solid core and soldered and all looks well.

Months later, the stranded joint can be rotated around the solid wire, as the solder hasn't penetrated to the solid core...........

Yet others have used this method successfully.

 

As with most DCC wiring, stuff will simply work, even though it isn't the ideal way of doing things.

 

The first DCC layout I was involved with was almost 50' long with reasonably parallel conductors along the main bus and the DCC controller main feed at about 1/4 along the length - with board joints every 4'.

 

It worked, but we had random runaways. Possibly attributed to not using a twisted pair over such a long length.

I'll never find out as it has been sold on.

Others can quite happily use a "ring main" DCC bus on a large layout,but others will have all sorts of problems on a 6x4 with it and breaking the ring makes it work...........

 

There is the argument about using twisted wire for the bus to reduce inductance when compared to ordinary parallel connectors.

Why use twisted, when the rails (that carry the same power/signal) are parallel by their very nature?

But then how many of us have a continuous unbroken pair of rails along the length of a layout? Insulated joiners will break up that parallel conductor into much shorter sections (and therefore inductance)

 

If it works for you - then OK.

 

 

 

 

 

 

 

 

 

[RFS]

10 hours ago, newbryford said:

 

Because single core - when stripped can have a coating on it that the solder does not break through properly.

I have seen joints where the stranded dropper wire has been wrapped around the solid core and soldered and all looks well.

Months later, the stranded joint can be rotated around the solid wire, as the solder hasn't penetrated to the solid core...........

Yet others have used this method successfully.

 

 

Like many other modellers, I have long since given up working under a fixed baseboard with a hot soldering iron. Instead I use Scotchloks to connect a short wire to a choc-bloc connector, and if I were doing it again today I would use Wago connectors. Dropper wires are soldered to the rails on the workbench before the track is laid, and then the under board work just needs a screwdriver (not even that with Wago connectors). You just need to ensure you use the right size Scotchlok for your solid-core bus wire - for my 2.5mm wire that means the blue type. 

[WIMorrison]

I use red scotchlocks on my 2.5mm cable because the 16/02 dropper wire I use to the track doesn’t always connect in the blue scotchlocks - just need a stronger hand (or bigger pliers) to close the red on the 2.5mm cable

[Phil S]

If not using a 'proper' crimp tool, then a mole wrench can be set to provide the appropriate squeeze, and then be released ( I sometimes use IDC connectors in red and blue according to inner/outer... with a wire size that is in limits on both. But I also use wago connectors as these allow change and without tools.

 

Concerning operational voltage drop: Sound decoders can be very revealing of voltage drop around a layout.  Whilst gradually increasing voltage drop with distance may go unnoticed on speed and lighting until it is extreme !

 

Concerning 'power dissipation'  in the distribution ( bus and track combined )...

The max output should be assumed for the feed from the controller until it reaches any protective device such as a psx/breaker.. BEYOND  which a lower setting Could apply.

(Compare to household 240vac supply and 100A feed to board with breakers. of 32A 16A.. 5A )

 

This (max) may be taken continuously without tripping a fault condition. ... the wiring should be capable of the full current without melting etc.etc

When a short occurs (imagine a metal wheel ... it may experience arc welding due to the current, and get hot rapidly, but as stated by others, the heat dissipation will be where the resistance is .... and thin wiring will mean that is spread around the layout (as well as at the wheel)... but a 'heavy bus' will localise it to the fault' (low if not zero value) shorter-than-expected ie 'short' circuit.  

The bus wire should not, theoretically have to sustain the fault current's 'wattage' (Joules of energy per second ) continuously as the controller SHOULD trip .(after a small number of joules).... but disasters happen during minor fault conditions which have not been properly allowed for .... like the 13A fuse left in the '13A' plug used on a low wattage appliance. ....

 

This is why the oft-asked question of what size wire for a layout cannot be answered without knowing the controller maximum output, track lengths/ distances involved and much more such as the loco load lighting and sound requirements. ...

'Fit for purpose ' is the requirement

 

A short at a Hornby motor bogie's motor terminals, running on a ( 1A ?) analogue controller can melt the loco wiring.. when this happens with a dcc decoder installed, the decoder acts as a fuse! ! and protects the wires.

 

(Sent from a swedish lakesidé campsite with trains running nearby) 15kV at 16 2/3 Hz except for 'the bridge' 25kV 50Hz powered from Denmark.

Edited July 27, 2019 by Phil S
Added comment

[Edwin_m]

1 hour ago, Phil S said:

but as stated by others, the heat dissipation will be where the resistance is .... and thin wiring will mean that is spread around the layout (as well as at the wheel)... but a 'heavy bus' will localise it to the fault' (low if not zero value) shorter-than-expected ie 'short' circuit.  

I've never seen that argument put before.  It has some logic to it but because the wiring resistance will vary depending on how far the short is from the command station, it would be difficult to apply this principle to a short near the command station without invalidating the protection against short circuits on more distant parts of the layout.  You'd be better off inserting a resistor of small ohmage and large wattage in the command station feed and making the rest of the wiring as low resistance as possible.  Or just ensuring that the short circuit protection is appropriate to the power demand of the layout, adjusting the command station trip current if possible and using circuit breakers if necessary, and ensuring that the wire is suitable for those currents but more importantly that the whole layout is coin-tested.  It's also worth considering what fault current might flow if a train is bridging two power districts when a short circuit happens somewhere. 

[newbryford]

2 minutes ago, Edwin_m said:

  You'd be better off inserting a resistor of small ohmage and large wattage in the command station feed and making the rest of the wiring as low resistance as possible.  

 

Isn't that the equivalent of using a car light bulb in series?

 

[Pete the Elaner]

6 minutes ago, Edwin_m said:

You'd be better off inserting a resistor of small ohmage and large wattage in the command station feed and making the rest of the wiring as low resistance as possible.

Which is a step shirt of being a fuse. The difference is that a fuse melts, shutting off the circuit.

[Edwin_m]

15 minutes ago, newbryford said:

 

Isn't that the equivalent of using a car light bulb in series?

 

 

2 minutes ago, Pete the Elaner said:

Which is a step shirt of being a fuse. The difference is that a fuse melts, shutting off the circuit.

Absolutely agree.  Keep the bus resistance low and make sure the cutout is effective. 

[Phil S]

I was not advocating any use of a slow acting car bulb.... just pointing out that resistance is the process by which energy is being converted from electrical to heat .....a super conductor has no resistance and and no energy loss ....  ...wherever there is resistance and current flow there will be heat transfer which needs to be dissipated. ..  open rail may do this relatively well, but a thin wires in an enclosed space will not.

[Crosland]

10 hours ago, Edwin_m said:

just ensuring that the short circuit protection is appropriate to the power demand of the layout, 

 

This ^^^^

 

I have seen some stupidly large boosters being mooted for quite small layouts.

 

If you do have a lrage layout, use multiple smaller boosters or block cut-outs.

[Crosland]

10 hours ago, newbryford said:

 

Isn't that the equivalent of using a car light bulb in series?

 

 

No. A light bulb behaves very differently to a resistor.

[Andymsa]

On 22/07/2019 at 15:02, petejones said:

I was just about to purchase some 24/0.2 cable for my power bus when I remembered I have rolls of household electricity cable in the shed, from lighting to heavy duty mains cable. I'm guessing that would be OK to use, stripped of the grey insulation? Would single core or multi-core cable be best for the power bus?

 

After reading this thread which seems a very emotive subject. So to answer the original post, your mains wire is perfect for you bus wires, stranded is better but not essential if your layout is fixed and isn't portable. Also I would use 2.5mmsq as opposed to 1.5mmsq but again you could get away with 1.5.

 

dont get to wrapped up with resistance drops, having a layout where my buss runs are around 30 feet long with the droppers I don't have issues with resistance. Just do the quarter or as I say 2p test for shutting down with shorts.

Edited July 28, 2019 by Andymsa

[Dungrange]

On 26/07/2019 at 11:36, Crosland said:

Did you read the bit where I said "minimise the voltage drop"?

 

On 26/07/2019 at 11:55, Crosland said:

You need to minimise the voltage drop as much as reasonably possible. One inch square copper bus bars would be taking it a bit far, for example

 

Is it really the "voltage drop" in the DCC bus that we are trying to minimise?  I understand that this is what we want under normal operating conditions, where most of the potential difference from the command station will be dropped across the load (ie the locomotive) and very little should be dropped across the length of the DCC bus to ensure that our 12v motor can in fact be provided with 12v at the far end of a layout.

 

However, if the critical test is that the booster shuts down when we have a short circuit, is it not minimising the resistance of the DCC bus that is the critical design parameter?  Perhaps I don't fully understand the 'coin test' but I think what this does is remove a relatively high resistance load (a locomotive) and replace it with a load of negligible resistance (a coin).  When this short circuit occurs, the command station will still have a potential difference of around 14V across it's output terminals that are connected to the track, but instead of this being dropped across our locomotive, it now has to be dropped across the length of the DCC bus between the command station and wherever we have placed the coin.

 

Looking at the User Manual for my command station, it is designed to trip at 5.5 Amps, which means that for a short circuit to be detected, I would need the total resistance of the circuit between the output terminals on the command station to be less than 2.5 Ohms (14 volts / 5.5 Amps).  If the resistance of the wiring is greater than this, then the current won't reach the critical 5.5 Amps that would be necessary to cause the booster to shutdown.  This therefore implies to me that it is the resistance under a short circuit scenario that matters more than the potential voltage drop under normal operating conditions.

 

On 26/07/2019 at 11:46, Crosland said:

To be fit for purpose on a DCC layout a loco (any accessory in fact) needs to be able to pass the full booster trip current through it's internal wiring, pickups, etc., without damage, and allow the booster to trip in the case of this kind of internal fault. This point is completely ignored by all users and manufacturers as far as I can tell. It comes down to diminishing returns. How likely is such a fault in the first place?

 

This seems to indicate that it is not desirable to have a DCC command station with a high current output (eg 5 Amps) that is not fed through a number of lower rated circuit breakers, since the whole purpose of fuses and circuit breakers is that these would be the 'weak points' in the circuit; breaking the flow of current before damage is done elsewhere.  I guess that this means that there is a trade off between the cost of fitting additional lower current rated power districts relative to the cost of the damage that may be done by a relatively high current flowing, but which is not high enough to cause a shutdown of the DCC system.  Obviously it's time to look at intermediate circuit breakers.

[Crosland]

7 minutes ago, Dungrange said:

Is it really the "voltage drop" in the DCC bus that we are trying to minimise? 

 

Yes.

 

7 minutes ago, Dungrange said:

However, if the critical test is that the booster shuts down when we have a short circuit, is it not minimising the resistance of the DCC bus that is the critical design parameter? 

 

They are inextricably linked by ohms law. Minimise the resistance and you will minimise the voltage drop for a given current flowing.

 

[RFS]

On 26/07/2019 at 11:46, Crosland said:

 

To be fit for purpose on a DCC layout a loco (any accessory in fact) needs to be able to pass the full booster trip current through it's internal wiring, pickups, etc., without damage, and allow the booster to trip in the case of this kind of internal fault. This point is completely ignored by all users and manufacturers as far as I can tell. It comes down to diminishing returns. How likely is such a fault in the first place?

 

 

Surely the requirement is that, for practical purposes, the internal loco wiring only needs to be able to pass the full booster trip current momentarily. The internal wiring of locos seems to be perfectly capable of doing that just as the thin wiring of solenoid point motors is able to take the momentary surge of power from a CDU. Give the equivalent CDU power to a point motor for a couple of seconds or more and it will burn out. 

 

My layout is approximately 32ftx9ft and has a 5-amp booster,  2.5mm solid core bus wires and droppers on every piece of track. Never had a problem with locos causing shorts - the booster always trips immediately and I've never seen any damage to the internal wiring.

 

Perhaps we're trying to over-think the solution here. Whilst you can under-configure the DCC bus wires you can't over-configure. The difference in cost between, say, 1.5mm and 2.5mm bus wire is trivial so go for the latter. And make sure you don't compromise the solution by not having adequate droppers, especially with turnouts. Whilst copper is an excellent conductor, nickel silver rail, and especially rail joiners, are not. 

[Phil S]

The reason there are  'several answers' is that there are several conditions which it is desirable to meet.

The safety item is the trip current and the loop resistance of whatever dcc bus is used.... it must be low enough to reach the trip by ohms law... it therefore depends on the track dcc voltage and the maximum normal current of the controller ( and by implication/ inference the higher trip value.

The operational voltage drop around the layout affects speed (at a given speed/regulator setting.. (ignoring attempts by the back emf processing in the loco decoder whilst within its range of correction), lighting in coaches... especially noticeable if still using filament bulb lighting .... but my 00 loft layout is taking more than 1 Amp of coach lighting without any trains moving! ... but as I mentitled before. ..Sound-fitted locos are the most obvious when a voltage drop and brown out occurs ... as your can hear the sound restart.  (Obviously stay alive etc can help out here). The NMRA spec says that at 7V all decoders should work.... many will work at a lower voltage.... but if you start with 22V you have more headroom to drop volts. ..which is useful in the garden  ...but our 125m of track there only has about 0.5 ohms resistance thankfully... (and we are now using 20V there because we use 16V on the portable versions so speeds match better). The point being that with many locos running and lights and sound, if the voltage drops to 7V ( only 5v from 12V, but 9v from 16V as per Rocodcc  etc) at the maximum current, then operational problems will start occuring. V=IR   

(14 down to 7 =7) = 3.5A max x 2ohms

(16 down to 7 =9) = 3.5A x 2.57ohms

(22 down to 7 = 15 = 8 amps max x

1.875 ohms

Although I usually calculate this based on an arbitrary 2V drop as being acceptable... or the target...

2 volts drop = 3.5A max x 0.57ohms

2V drop = 8amps max x 0.25 ohms

And my garden line does not meet this standard.... (mine=0.5 ohms)

However I normally only take 3 to 5 Amps... only at a friends open day recentre did we get up to 7.9A with 3 large trains running  (with sound, smoke,  AND bulb-lit coaches) 

[WIMorrison]

Given how few posts are related to problems that cause the poster to rewire a layout bus because he has used wire that is too light or solid and the number of differing views propounded in this thread I suggest that this is largely a willy waving exercise than a real world practical issue.

[Edwin_m]

1 hour ago, Crosland said:

 

Yes.

 

 

They are inextricably linked by ohms law. Minimise the resistance and you will minimise the voltage drop for a given current flowing.

 

Absolutely.  Furthermore, as the voltage drop is the resistance multiplied by the current, it can also be minimised by ensuring that the maximum current possible is no more than what is necessary (power districts, circuit breakers, command station current limit). 

[Edwin_m]

12 minutes ago, WIMorrison said:

Given how few posts are related to problems that cause the poster to rewire a layout bus because he has used wire that is too light or solid and the number of differing views propounded in this thread I suggest that this is largely a willy waving exercise than a real world practical issue.

Tell that to the loco I melted (actually thinking about it, it was a driving trailer with lighting).  There's an important principle here that just using wire with ampage greater than that of the command station may not be enough, and that the coin test is essential.