Third Rail - How does it work?!

[sem34090]

A silly question, I know, but can anyone provide a simple explanation as to how a third rail system works, particularly how the return current works and where the positives and negatives are?

[The Evil Bus Driver]

Usually the conductor rail is the positive and one of the running rails is the return. London Underground is different as there are two separate live rails, one positive and one negative. These lines are usually DC and have power supply stations every couple of miles or so.

[Talltim]

The third rail (in the U.K. nowadays ) is +750ishV and the running rails are 0V. Thus you get a difference of 750V. As TEBD says above, there are other ways of doing it, with a lower positive but also a negative fourth rail such as London Underground uses, with +420V outside rail and a -210V centre rail, giving a difference of 630V. In certain areas third rail mainline trains run on LU tracks using just the outside rail, the running rails being bonded to the return but not energised. This means they have reduced power, only getting 420V.

 

Edited April 10, 2019 by Talltim

[Kelly]

There are areas where there are 4 rails, but nothing to do with LU, but to do with earth leakage protection. Earth leakage can corrode pipes etc, so is reduced as much as possible as a result.

 

When the LSWR first introduced their 3rd rail system, the National Physics Laboratory complained about such leakages, and modifications were needed. The Royal Mail made similar claims later in a different area iirc.

[Nearholmer]

Not quite, Talltim.

 

This sketch should explain things at a basic level, although there are all sorts of variations in the way that the running rails are bonded, according to the type of track circuits used for signalling.

[sem34090]

Thank you all - More or less answers my question. A friend asked me, knowing me to be interested in SR EMU's, and it suddenly occurred to me that I didn't know and really ought to! 

 

I assume that, generally no charge is carried in the running rails on a 750 DC Outside 3rd system though as otherwise how to level crossings work on the Southern?! 

[Nearholmer]

The running rails in such a system are not deliberately earthed, in fact they are deliberately insulated from Earth at each rail fastening, so it is possible for potential differences to exist between the running rails and Earth.

 

However, the magnitude of such differences is kept small by providing very good bonding between all the running rails that are in the return circuit, and by the fact that the thousands of rail fastenings, when taken together, all in parallel, present a relatively low resistance to Earth.

 

 

Edited April 12, 2019 by Nearholmer

[phil-b259]

8 hours ago, sem34090 said:

... as otherwise how to level crossings work on the Southern?! 

 

By having multiple conductor rail pickups (usually at each end of the unit on both sides) and linking them all together by a big cable running the length of the unit* thus meaning there is usually at least one pick up in contact with the 3rd rail at all times.

 

Gaps are also provided at pointwork as the 'side entry' ramps that allow for a continuous 3rd rail can only be traversed at slow speeds.

 

If the gaps are too long then the unit may be able to coast through them - but its not always thus. A 2 or 3 car unit must NOT proceed out of London Victoria station unless it has a green signal - a single yellow (proceed at caution) at the platform end runs the risk of it ending up 'gapped' between conductor rails as the momentum is not there to carry the unit past them. Not a problem with a 4 car unit due to it being long enough to not get gapped.

 

* Said cable is only within the unit - i.e. there is no cable connection / jumpers to pass the 750v between two 4 car units making up an 8 car train.

[phil-b259]

32 minutes ago, Nearholmer said:

The running rails in such a system are not deliberately earthed, in fact they are deliberately insulated from Earth at each rail fastening, so it is possible for potential differences to exist between the running rails and Earth.

 

However, the magnitude of such differences is kept small by providing very good bonding between all the running rails that are in the return circuit, and by the fact that the thousands of rail fastenings, when taken together, all in parallel, present a relatively low resistance to Earth.

 

 

 

If the running rails were not insulated from each other (and earth potential) then track circuits wouldn't  be possible!

 

One of the downsides of the 3rd rail system (as opposed to the 4th rail system used on London Underground) is that track circuit currents and the traction return currents must mix in the same rails.

 

From a traction engineers perspective they want as many rails connected together as possible to provide the best negative return path back to the substation as possible - while the S&T engineer wants to keep each rail electrically separate from every other rail

 

To get round this problem it is necessary to install 'Impedance bonds' (sometimes known as spiders on account of the leads attached to the running rails) in the 4ft at Insulated Block Joints (and mid way through long track circuits) to allow the DC traction current to pass round the IBJs / across to adjacent lines while simultaneous blocking the AC based track circuit current from doing the same.

 

(To understand the specifics you need to have an understanding of AC theory which governs how Capacitance, Inductance and Reactance / Resistance interact to provide a short circuit or open circuit depending on the electrical frequency present).

 

At pointwork putting bonds by every IBJ is impossible so only a single running rail is used as part of the traction return path - but it needs to be continuous from end to end and needs to be changed to go through both running rails at either end.

 

 

Edited April 12, 2019 by phil-b259

[Southernman46]

16 hours ago, phil-b259 said:

 

Gaps are also provided at pointwork as the 'side entry' ramps that allow for a continuous 3rd rail can only be traversed at slow speeds.

 

* Said cable is only within the unit - i.e. there is no cable connection / jumpers to pass the 750v between two 4 car units making up an 8 car train.

Side ramps - 20mph in the facing "running-on" direction.

 

Traction jumper cables - (I believe) The last EMU's built with such were the 1957 BR Tyneside 2-EPB's (subsequently removed) and (definitely) the last to carry them being the 4-SUB's (the RH jumper looking at the front) as the control & lighting circuits were 750v.

[WessexEclectic]

On 11/04/2019 at 02:39, Kelly said:

When the LSWR first introduced their 3rd rail system, the National Physics Laboratory complained about such leakages, and modifications were needed. The Royal Mail made similar claims later in a different area iirc.

 

You wouldn’t happen to know any further details or sources for the NPL story? Thanks.

[Compound2632]

On 11/04/2019 at 02:39, Kelly said:

When the LSWR first introduced their 3rd rail system, the National Physics Laboratory complained about such leakages, and modifications were needed. The Royal Mail made similar claims later in a different area iirc.

 

The National Physical Laboratory in Teddington, Middlesex, is between the Kingston loop and the Hampton Court branch. The story as I heard it when working there was that the earth currents between the two lines were sufficient to disturb precision magnetic measurements. I never had any direct evidence of that but the microwave and optical frequency standards with which we were working are highly sensitive to fluctuations in magnetic field, which we were at pains to screen against. A colleague had a half-hourly feature in optical frequency standard data that was found to be correlated with trains passing the level crossing at Strawberry Hill - where there is a substantial gap in the third rail. Half-hourly as the clockwise and anti-clockwise Kingston loop trains are timetabled to cross there, to reduce the frequency with which the crossing is closed to road traffic.

 

The Paris Metro is, IIRC, a four-rail system, so in principle there should be significantly less ground leakage of current. Nevertheless when I was working on microwave frequency standards at the Observatoire de Paris, before my time at NPL, the best data was obtained during the (frequent) Metro strikes. This was down to two factors: reduced magnetic field perturbation; and the difficulty of getting home, encouraging long working hours on the experiments!

[Nearholmer]

The NPL problem was recognised, and part of the "introductory lore", when I worked with BR(S) traction power supply department from the mid-70s to late-80s, and there was also lore relating to something similar, although I can't remember exactly what, that was observed at Greenwich (Declination of the earth's magnetic field? Variation of magnetic north? which might be the same thing. Magnetic inclination? Which I know isn't the same thing. Something like that.).

 

I have a feeling that Greenwich Observatory and The general post Office were the two bodies that caused regulations to be introduced to limit the potential between traction return circuit and the general mass of earth to (IIRC) 6V. If I've got that figure right, 6V is actually a very high potential in this context, sufficient to drive very high currents under the right conditions, so they weren't being needlessly punitive in the restriction.

Edited May 6, 2019 by Nearholmer

[Compound2632]

Do you know the nature of the GPO's problem?

[Nearholmer]

Interference with telegraph, and later telephone, circuits that used earth as the return conductor.

 

I’m not sure of the exact mechanism of interference, but I do know that long-distance telegraphy used astonishingly low currents, detected by highly sensitive galvanometers, and my guess would be that localised changes in earth potential due to the flow of relatively high traction leakage currents was sufficient to ‘swamp’ the tiny changes in potential involved in telegraphy.

 

The 6V limit was probably sufficient to ensure that, in the event of accidental connection of traction return to Earth it didn’t pull the earth up above whatever potential the GPO was using (12V maybe?).

 

EDIT: I misremembered the limits, see detail in subsequent posting.

 

 

Edited May 6, 2019 by Nearholmer

[Edwin_m]

The Trafford line of Manchester Metrolink has some special measures to reduce electromagnetic interference in the vicinity of some premises containing EM-sensitive equipment. 

[New Haven Neil]

2 hours ago, Compound2632 said:

 

The National Physical Laboratory in Teddington, Middlesex, is between the Kingston loop and the Hampton Court branch. The story as I heard it when working there was that the earth currents between the two lines were sufficient to disturb precision magnetic measurements. I never had any direct evidence of that but the microwave and optical frequency standards with which we were working are highly sensitive to fluctuations in magnetic field, which we were at pains to screen against. A colleague had a half-hourly feature in optical frequency standard data that was found to be correlated with trains passing the level crossing at Strawberry Hill - where there is a substantial gap in the third rail. Half-hourly as the clockwise and anti-clockwise Kingston loop trains are timetabled to cross there, to reduce the frequency with which the crossing is closed to road traffic.

 

The Paris Metro is, IIRC, a four-rail system, so in principle there should be significantly less ground leakage of current. Nevertheless when I was working on microwave frequency standards at the Observatoire de Paris, before my time at NPL, the best data was obtained during the (frequent) Metro strikes. This was down to two factors: reduced magnetic field perturbation; and the difficulty of getting home, encouraging long working hours on the experiments!

 

Stephen, were you at NPL in the John Dunworth years?  His daughter is a friend!  Sorry!

[Compound2632]

1 minute ago, New Haven Neil said:

 

Stephen, were you at NPL in the John Dunworth years?  His daughter is a friend!  Sorry!

 

Not knowingly. 1995 - 2015.

[New Haven Neil]

2 minutes ago, Compound2632 said:

 

Not knowingly. 1995 - 2015.

 

Ah no - he was Director 64-76.  Way before your time!  His daughter is one of the most intelligent people I have even met - her husband is the other!  Dr D died aged 100 just over a year ago.

 

Back on topic the matter of the NPL and the earth return currents is mentioned in 'Southern Electric'. I don't recall which volume though.

[Mike Storey]

2 hours ago, Compound2632 said:

 

The Paris Metro is, IIRC, a four-rail system, so in principle there should be significantly less ground leakage of current. Nevertheless when I was working on microwave frequency standards at the Observatoire de Paris, before my time at NPL, the best data was obtained during the (frequent) Metro strikes. This was down to two factors: reduced magnetic field perturbation; and the difficulty of getting home, encouraging long working hours on the experiments!

 

The vast majority of the Paris Metro uses only a 750v DC third rail, very similar to the BR/SR system. The exception are the five rubber tyred routes, which use a guide rail supply - I am not sure what the technical difference in supply and return on those.

 

[Nearholmer]

I’ve found a copy of the BoT Regulations governing leakage from traction return (not on-line, it’s an Appendix in an old textbook), and the version as amended up to 1912 requires:

 

- potential difference return to Earth not to exceed 3V;

 

- leakage not to exceed 2A per single track mile;

 

- potential difference between any two points in the return circuit not to exceed 7V.

 

It is mostly aimed at street tramways, where the risk of leakage is far greater than on a railway, and the title makes clear that one of the big concerns was electrolytic corrosion of gas and water mains by return currents choosing that path.

Edited May 6, 2019 by Nearholmer

[Compound2632]

Mention of tramways, and also thinking of the RER line close to the Paris Observatory, prompts the observation that overhead systems must also be using the rails for return, or neutral.

[Nearholmer]

They do, except in the case of trolleybus systems, some odd three-phase systems like people-movers at airports, and very early Siemens DC tramway/railway systems which had two overhead slotted-tube conductors.

 

There are/were also probably other oddities that I’m ignorant of.

[Compound2632]

Trollybuses. Now I understood that they were a cheap way of continuing to use tram infrastructure without going to the expense of replacing worn out rails. So if the trams were returning through the rails, how did a trollybus, with its insulated wheels, work? 

 

EDIT: looking at a few photos, it's clear current collection is via a pair of poles from a pair of overhead wires, so conversion from tram to trolley must have involved re-building the overhead infrastructure, albeit using the same masts.

Edited May 6, 2019 by Compound2632

[uax6]

Through a negative return wire strung in the overhead.... or by dragging a skate on one of the rails behind the bus....

 

Andy G