Missed opportunity: 1500V DC fourth rail

[stanley T]

Alternative time lines are a bit of a waste of time, and in the case of the UK rail business, often a bit depressing. They can be quite fun and instructive, however.

In the general mess that was the post-1955 modernisation programme, one decision is generally regarded as far sighted: to go for 25kV a.c as the standard for future electrification. Globally, perhaps, but under British conditions, one wonders. With the insulators of the time, this involved major engineering works rebuilding bridges and lowering trackbeds with the restricted British loading gauge. The WCML, only as far as Weaver Junction, took an age to electrify – and the early a.c locos, with mercury arc rectifiers, were none too hot either. Those problems lessened with d.c overhead (1.5 or 3kV) but all that heavy gantry, numerous substations, and heavy wiring to blow over in British gales.

There was an alternative. As far back as 1910 the South Eastern and Chatham planned to electrify its suburban lines at 1500V DC fourth rail- the latter more sensible with such a high voltage than third rail, to prevent earth leakage. Bottom contact, already proven in the USA, and safer (the top and sides of the live rail can be insulated) and far less prone to icing. That voltage is about as high as you can get with electrified rail, to avoid arcing, but it’s doable – it’s used on the newly built Guangzhou metro. Manchester to Bury worked perfectly well on 1200v DC side contact for years. Unfortunately WW1 intervened, the SE&CR electrification never happened, and the Southern got the underpowered and far more hazardous 600-750v top contact system.

The point is that even today costs of electrifying on 750V third rail per mile are about the same as at 25kV – and that’s with better insulators for a.c not requiring such high headspans. The bridge and tunnel costs cancels out the numerous substations. With substations 3-4 times further apart, installation costs would have been substantially lower than for 25kV a.c in the 50s and 60s, structure more reliable (no dewirings) and the trains simpler and cheaper as well.

High speed? Well until the TGV came along 1.5kV held the world speed record, and even today TGVs speed along at up to 200km/hour between the end of the LGV at Tours to Bordeaux and beyond. Ah, but that is overhead: what about you can’t get shoegear that works at over 100mph? Well a class 442 got to 108mph and was then limited by the lack of SR juice; nobody has really tried to make a 125mph pickup shoe and it should not be much harder than a high speed pantograph. Freight? Electro diesels as with class 73 and the more powerful class 74, with the added advantage of being able to use unelectrified sidings.

Today a.c delivers superior performance , but in this world the WCML, ECML and GWML would all have been electrified by 1970 as steam carried on for an additional 5 years or so. The zoo of unreliable diesel locos that emerged after 1955 would not have been built. The diesel HST would not have happened, but an electric one would, as would fourth rail Pendolinos; as for non-electrified main lines, a 16 or even 18 cylinder Paxman Valenta would have delivered a 3500-4000 hp loco with a lower axle load than a class 67.

 

[WatcherZero]

Do you know what the energy lost to frictions like on fourth rail? Its bad enough on third rail.

 

[Old Timer]

Voltages above 750 DC are far harder to transmit, especially down steel rails, and the costs of the feeder stations would have been exorbitant. That is why no conductor rail systems beyond 750V DC have ever been progressed with.

1500v DC OHL was far too expensive for main line use and requires much heavier pantograph equipment than 25kV AC systems.

 

[stanley T]

As for friction - what, even with graphitised carbon as a collector head? It's a good lubricant as well as electrical conductor.

Energy losses with carbon steel rail are reduced with modern conductor rails, aluminium core with a stainless steel outer layer, but that would not have been available in the 50s, I suppose.

The high voltage losses problem could be minimised by having one rail at +750V and the other at -750V ( I think that was the SE&CR's intention). How does that work - does that imply two live rails? Doesn't the LU system have a negative potential on the fourth rail?

 

[Philip Elliott]

Surely you would do a 1,500V 4-rail system with one rail at +750V and the other at -750V?

 

[Peter Mugridge]

RER Line C in Paris has a 1500v DC overhead conductor rail fitted in parts of the central area where tunnel roof clearances are tight.

 

[jopsuk]

25kV AC Overhead rails aren't uncommon where clearances are tight- pretty sure they're used in Edinburgh, and I've found reference that they're used on the continent at up to 250km/h

 

[90019]

25kV AC Overhead rails aren't uncommon where clearances are tight- pretty sure they're used in Edinburgh, and I've found reference that they're used on the continent at up to 250km/h

The Haymarket tunnel that was electrified recently uses them, I think the other one is still cables.

 

[Nonsense]

Alternative time lines are a bit of a waste of time, and in the case of the UK rail business, often a bit depressing. They can be quite fun and instructive, however.

Agree entirely.

 

[TomJ93]

Would a 'upside down' version of OHLE work? basically all the gubbins between the rails and a pantograph?

 

[ainsworth74]

Would a 'upside down' version of OHLE work? basically all the gubbins between the rails and a pantograph?

Once you charged the wire they electricity would arc straight across to the nearest rail and probably melt both the wire and the rail. Basically there are clearances that have to be maintained when it comes to the OHLE otherwise you can get arcing which is bad (especially if it's a person that gets to close to live OHLE).

 

[Aictos]

RER Line C in Paris has a 1500v DC overhead conductor rail fitted in parts of the central area where tunnel roof clearances are tight.

Like in the Haymarket tunnels in the Edinburgh area?

 

[apk55]

There are several problems I can see with the proposed system.

1/ Maximum power that can be supplied. Even with substations every 2 to 4 miles the power that can be supplied is limited to about 4 to 5MW. This is proving a problem in countries with 1500V systems such as Holland and France and in some cases they are looking at converting to 25KV.

2/ Gapping problems are worse with 4th rail Systems. While not a problem with long multiple unit trains, with short trains or locomotives this could be a serious problem, with the possibility of a train stalling or gapped while negotiating a complex junction at a station throat.

3/ Clearance to under frame equipment. Some coaches and MU's have lots of low slung equipment underneath, which is only a few inches above rail hight and may pose problems. This can include disk brakes on axles. I have heard that this can be a problem on joint run lines with LT's system

 

[Holly]

The big advantage of a DC system is that you get more power for a given level of insulation. By a factor of the square root of two (AC gives roughly 30% less).
The big disadvantage of DC systems is that they can't use the more efficient high voltages, such as can be used in OHLE.

AC is inherently safer, but not by enough to be the basis of a decision between the two.

The big disadvantages of using both (some of each) is that they are not terribly good at peaceful co-existence. The root problem being that even fairly small DC stray currents are enough to drive an AC transformer (or autotransformer) circuit into non-linearity resulting in losses and unwanted heating. As a result changeover sections are complicated and expensive to engineer, build and maintain. Changeover at speed even more so.

In the long term, it would be good to see AC on the third rail even if this means some lowering of maximum power. As a result changeover sections would be relatively simple. The two systems could be mixed and matched far more than they are with relatively short sections of each, wherever the civil engineering advantage lies.

 

[stanley T]

There are several problems I can see with the proposed system.

1/ Maximum power that can be supplied. Even with substations every 2 to 4 miles the power that can be supplied is limited to about 4 to 5MW. This is proving a problem in countries with 1500V systems such as Holland and France and in some cases they are looking at converting to 25KV.

2/ Gapping problems are worse with 4th rail Systems. While not a problem with long multiple unit trains, with short trains or locomotives this could be a serious problem, with the possibility of a train stalling or gapped while negotiating a complex junction at a station throat.

3/ Clearance to under frame equipment. Some coaches and MU's have lots of low slung equipment underneath, which is only a few inches above rail hight and may pose problems. This can include disk brakes on axles. I have heard that this can be a problem on joint run lines with LT's system

Re (2) the sensible way round would be electro-diesels for freight, pretty much essential for night working when there is engineering work and the juice is turned off. A 5000HP electric with an 850HP diesel to slowly work across the gaps would do fine. For passengers, it's all EMUs these days anyway, but locos could either have a flywheel (worked fine in the old class 71) or a DVT or DPT at the other end which was powered. Main problem is the minimum length of a short train: either 3 x 23m or 4x 20m, but then on busy lines (why electrify quiet ones) they should not be shorter anyway,

Re (3) a safe fourth rail system with both rails "live" (at +750V and -750V respectively) would require bottom contact with insulated (top and sides) electrified rails on either side of the running rails, not with a central rail. Might make gapping problems worse, though.

(1) is the killer objection, added to which regenerative braking does not really work on DC. 5MW is more than adequate for 125mph running, but that only happens on 1.5kV in southern France where the traffic density is low by UK standards. Two 11 car Pendolinos between the same substations wouldn't really work; but then 3 minute headways @125mph is a separation of 6 miles.

The big disadvantages of using both (some of each) is that they are not terribly good at peaceful co-existence. The root problem being that even fairly small DC stray currents are enough to drive an AC transformer (or autotransformer) circuit into non-linearity resulting in losses and unwanted heating. As a result changeover sections are complicated and expensive to engineer, build and maintain. Changeover at speed even more so.

Fourth rail systems are much less of a problem here than third rail, where the return path is through the running rails.

 

[es373]

I would have thought that AC was the chosen method purely for the fact that DC equipment tends to be so much larger, heavier and bulkier than AC equipment.
Just take a look at an AC traction motor and then look at a DC traction motor.

It doesn't just apply to motors either.
--- old post above --- --- new post below ---

25kV AC Overhead rails aren't uncommon where clearances are tight- pretty sure they're used in Edinburgh, and I've found reference that they're used on the continent at up to 250km/h

True.. Next time you goto St Pancras. Take a look at the eurostar platforms and look how low the 25kV line is.

 

[apk55]

Re (2) the sensible way round would be electro-diesels for freight, pretty much essential for night working when there is engineering work and the juice is turned off. A 5000HP electric with an 850HP diesel to slowly work across the gaps would do fine. For passengers, it's all EMUs these days anyway, but locos could either have a flywheel (worked fine in the old class 71) or a DVT or DPT at the other end which was powered. Main problem is the minimum length of a short train: either 3 x 23m or 4x 20m, but then on busy lines (why electrify quiet ones) they should not be shorter anyway,

Re (3) a safe fourth rail system with both rails "live" (at +750V and -750V respectively) would require bottom contact with insulated (top and sides) electrified rails on either side of the running rails, not with a central rail. Might make gapping problems worse, though.

(1) is the killer objection, added to which regenerative braking does not really work on DC. 5MW is more than adequate for 125mph running, but that only happens on 1.5kV in southern France where the traffic density is low by UK standards. Two 11 car Pendolinos between the same substations wouldn't really work; but then 3 minute headways @125mph is a separation of 6 miles.

The big disadvantages of using both (some of each) is that they are not terribly good at peaceful co-existence. The root problem being that even fairly small DC stray currents are enough to drive an AC transformer (or autotransformer) circuit into non-linearity resulting in losses and unwanted heating. As a result changeover sections are complicated and expensive to engineer, build and maintain. Changeover at speed even more so.

Fourth rail systems are much less of a problem here than third rail, where the return path is through the running rails.

How would you cope with a triangular junction or loop. Presumably one side would be positive and the other negative so if a train is turned the polarity would be reversed. A very long section gap would be required which would involve the train coasting and a limit on the length of connected shoes. The train would also need polarity sensing and changing equipment.

 

[apk55]

The Gapping problem would horrendous - even worse the TFL's 4 rail system as pickup on on one side would not be possible for the full length of a point. A high speed point on a ladder would involve a very long gap, followed by yet more.

Secondarily how would you deal with triangle or reversing loops. Assuming one side is positive and the other negative and you turn a train it would be supplied with opposite polarity. This would involve trains fitted with polarity sensing and changeover equipment. And involve coasting sections longer than the maximum length of trains fitted with a bus bar.

Running conductor rails with AC would still present problems. The impedance of iron rails increases with frequency, due to the skin effect (the magnetic nature of iron increases the effect) and would be quite significant at 50Hz. (I am an electronic engineer)

 

[stanley T]

I would have thought that AC was the chosen method purely for the fact that DC equipment tends to be so much larger, heavier and bulkier than AC equipment.
Just take a look at an AC traction motor and then look at a DC traction motor.

Big heavy motors, yes, but no transformers or rectifiers. The transformer is usually the heaviest item on an a.c loco.

 

[bangor-toad]

Running conductor rails with AC would still present problems. The impedance of iron rails increases with frequency, due to the skin effect (the magnetic nature of iron increases the effect) and would be quite significant at 50Hz. (I am an electronic engineer)

Hi there,
This highlights the BIG problem of an AC distribution system.

For a DC voltage the current flows down the conductor and it pretty much flows through all of it. ie if youu have a big conductor rail then the current will flow through the whole cross section of the rail.

With an AC voltage you come across the skin effect. This means that the current only flow in the surface of the conductor. For Copper that's about the outer 8mm and for iron or steel it's the outer 1mm at 50 Hz. This means that there is much less material to carry the current than you'd expect and resistance losses can occur.
There is no effect from the voltage - the skin depth for 1500 Volts AC is the same as for 25k Volts AC. That means that the current would have to massively increase in the available conductor - probably leading to all sorts of losses.

By using 1500 Volts AC you will need more current. You cannot just increase the dimension/size from a 20mm cable to a 3rd rail cross section as virtually all of the 3rd rail will not actually any current running through it. The only way to do it would be to start to use some rather more exotic materials and metal - a solution that will be far from cheap!

I think the engineers have got the existing systems right. High current, low(ish) voltage DC 3rd rail works as does high voltage AC systems.

Cheers,
Jason

 

[stanley T]

You have convinced me that the gapping problem with both electrical rails outside the running rails is insuperable - after all, high speed points can be up to 120m long. Central negative rail it would have to be, and thus side contact would be the safest option available.

As for available power and substation spacing, would 3kV DC third rail be feasible (+1.5kV/-1.5kV) or is that getting too hairy? Manchester-Bury managed OK on 1.2kV side contact third rail.

Actually this discussion is convincing me that high voltage, low current OHLE was a far sighted correct decision after all - as long as it is not as crappily built as the ECML.

The electrically savvy among you, why are there no high voltage d.c systems? Motors can't cope with high voltages?

 

[Holly]

With an AC voltage you come across the skin effect. This means that the current only flow in the surface of the conductor. For Copper that's about the outer 8mm and for iron or steel it's the outer 1mm at 50 Hz. ...

It is true that the ferromagnetism of iron conductors dramatically increases skin effect. However, third rails are already made of aluminium with a steel wearing surface. A parallel conduction path would indeed be needed for ground circuits. This is not so very different in principle from the parallel power conductors needed for 25-0-25 autotransformers circuitry used in 140mph OHLE.
Would this be cheaper than a centre fourth rail? I think so (not entirely certain), it would certainly be safer.
Yes, AC on the third rail does create some engineering challenges. A good part is that AC/DC systems third-rail are possible during a phased changeover and in the long term there is a lot more compatibility with existing high voltage systems. The costs are somewhat offset by cheaper power distribution equipment.

As I have written, once it becomes technologically routine and cheap to have short sections of third rail and OHLE then eventual electrification of branch lines (limited to say 80mph) becomes easier and cheaper. Because you can use whichever is cheaper at the locality (OHLE for long featureless runs and level crossings, Third Rail for low over-bridges and tunnels). It is a long term aspiration.

 

[bangor-toad]

However, third rails are already made of aluminium with a steel wearing surface.

.

Hi there,
The same problem with the skin effect takes place with aluminimum 3rd rails. It means that AC power at 750 (ish) in a 3rd rail isn’t going to work.

Why?
Let’s assume that the 3rd rail is 7cm wide, 10 cm tall and has a 0.5cm thick steel wear plate on the top.
With DC power, the current flows throughout the whole rail. Assume it’s an even current density (even though there is a bit of difference with the steel it doesn’t really make any difference). So, the cross section of the rail is 7 x (0.5 +10) or 73.5cm2.
The current density required is clearly OK as this system is used.

With AC power, the skin effect needs to be considered. I don’t know quite how to calculate it with the effect of the wear layer so lets assume that you can treat each metal separately and add them up.
At 50 Hz the skin depth of steel is 1mm. That means the current flows through the outer 1mm of the steel. In total that’d be top + the bottom + the sides. Numerically that’s 0.1 x7 + 0.1 x7 + 0.1 x.3 + 0.1 x3 which is 2 cm2.
At 50 Hz the skin depth of aluminimum is 10mm. Do the same calculation of the top, bottom and sides (1x7 + 1x7 + 1 x8 + 1 x8) gives a total area of 30cm2.
In total AC current will flow through 32cm2 of the 73cm2 conductor rail.
(OK, this assumes an absolute cut-off effect but it’s a good enough approximation!)

Very roughly, electrical losses are proportional to the square of the current and directly with the resistance of the conductor.
As with AC the current is compressed into a smaller area, the current density will increase by a factor of 2.3 (73.5 divided by 32) and the effective resistance will also increase by the same amount.
To estimate the electrical losses you multiply it all together, ie 2.3 x 2.3.x 2.3.
This means that to deliver the same amount of power down a 3rd rail, the electrical loss due to the rail itself is 12 times greater for AC than DC.

This is too much and would need to be minimized. The only way to do that is to lower the current and to do that you need to increase the voltage. TO make it worthwhile you need to increase the voltage to the kilovolt range. It just isn’t viably safe to have kilovolt conductors mounted some 15 cm off the ground.

AC 3rd rail will never happen.
Using an AC “rail” or bar where the OHLE clearance is tight is a tried and tested method abroad. I am sure we’ll see bits here.

I also think that we will see more dual voltage trains able to operate from 25kV(ac) OHLE and from 750 V(dc) 3rd rail. There is no reason to create any other type.

Cheers,
Jason

 

[Old Timer]

Actually this discussion is convincing me that high voltage, low current OHLE was a far sighted correct decision after all - as long as it is not as crappily built as the ECML.

This myth about the ECML has been debunked several times on here, and I am led to believe that this has already been pointed out to you, however I (and a number of OHL engineers) would be very grateful if we could benefit from your obvious wealth of OHLE experience and knowledge as to why the ECML is is "crappily built". :roll: :roll: :roll: :roll:

Unfortunately I am not in the UK right now so I personally will not be able to reply straight away to any response, but I am sure that I can withstand the excitement of your interpretation of distribution power loadings and critical analysis of the design of Mk 3 contact systems for a week or so.

I do know that several senior Balfour Beatty OHL engineering managers who built the system will be eager to hear more, bearing in mind the obvious inference that can be drawn from your remarks about their competence and integrity. No doubt they will wish to pursue this aspect with you separately in due course.

 

[starrymarkb]

108mph is not the fastest speed achieved on Third Rail.

Back when the TMST was being designed there was a high speed run to demonstrate the third rail to some skeptical French engineers. They used a solo 4-REP which was radar clocked at 115mph. The ride was apparently terrifyingly awful so the results were not a total success from a PR view point (plus the lack on onboard speed indication beyond "off the scale"), but it proved the third rail worked

 

[es373]

Personally I prefer OHL. We all know that AC will travel very good distances not losing a whole lot of its oomph through pure copper wire. Unlike DC which seems to need more section gaps, sub stations etc..( I think anyway)

3rd rail tends to create more problems - I've seen quite a lot of axle boxes that have arc damage which again can prove to be fatal.
I dont really see a lot of problems with OHLE.

This is a good discussion though.. Im liking the read!

 

[stanley T]

Old timer,

I am not impugning the competence of the OHLE engineers that wired the ECML, but you yourself admitted that 125mph trains were put to run over a system designed for 100mph.. The fault, as ever, can be put down to HM Treasury and a political culture that does not care about long term performance, because that is after the next election.

In this purely "what if" thread the idea was to see if a robust, cheaper system could have led to more electrification and if it would have been more politician proof. Sadly, I think that the latter would not have been the case ; the battleground would have shifted to not funding enough extra substations to provide the juice to accommodate extra traffic and 125mph running.

Substations are fairly cheap, it seems, around £2m a throw; the system would have met the needs of the ECML (and whether electrification Newcastle -Edinburgh was ever justified is another matter...) GWML and MML, but the WCML south of Crewe and especially south of Rugby is another matter... let alone the power to get freight over Shap and Beattock.

With 4th rail Pendolinos not achieving their design speed of 125mph as they run out of juice, we would now be talking about a new line with this funny continental system of 25kV a.c overhead called... HS2:? Unless of course 3kV fourth rail, (+1.5kV/-1.5kV on each rail) had proved practicable, that provides a lot more juice.

 

[Old Timer]

Stanley T

It was you who said it was "crappily built". It was not. It was constructed to the then current Standards using "spare" equipment from the Midland and East Anglia Schemes. The extension to Peterborough was done simply to accommodate EMUs as that was all that would operate.

The ECML electrification scheme going forward from Peterborough only really occurred because of the great work done by Norman Howard and Tony Goldring who brought the extension from Hitchin in against a lot of reluctance. They had a dream and saw an opportunity and took it.

Until UK1, there never was a Catenary system in the UK that was designed for running above 100 mph.

The equipment on the CTRL is "constant height" equipment, e.g. every Structure is the same, and as such a monkey with basic training could build it. The UK situation is far more complex because it does not benefit from a constant height clearance and thus the whole system needs to be properly engineered.

The fact that the original Mk3 has been able to be modified several times over and modified to 125 mph running is testament to the work of the designers and construction staff (Balfour Beatty and British Rail). It is a sound design, and in the case of the ECML has more than adequately performed well above its original specification.

That in itself is something that is recognised by OHL Engineers, who hold the original construction teams in high regard. To say that it was "crappily built" is offensive to those who worked so hard to deliver what was then a Project without great support (especially from the enthusiast fraternity as I recall).

Talk of cheaper alternatives is a nonesense, as even those suggesting the idea should be able to extrapolate the results that have arisen from the original work.

DC is NOT the solution and has never factored in any high speed/long distance electrrification anywhere. The limitations of this were learnt from the 1940s onwards.

Similarly, third rail contact systems do not work efficiently above 100 mph which appears to be above the optimal speed in any case.

I really would like to think we can debunk both that as well as the ECML OHL myth at some point.

 

[NightatLaira]

I was told once that aluminium catenaries were used on the electrification of the ECML because copper was too expensive. Is this true? Or is this another bogus myth?

 

[Barrett M95]

The ECML is a mixture of MkIIIA and MkIIIB design (with some modified MkIIIB between Peterborough and Grantham as I understand).

The original fit catenery wire in all such systems is steel-reinforced aluminium - this was standard practice. Where problems were seen with this type of wire, some was retro-fitted with copper alloy catenary wire.

However that applies only to the catenary wire, the contact wire on all MkIII systems is copper with a cross section of 107 sq mm.

UK1 has slightly heavier wire at 120 sq mm and CTRL (HS1) uses heavy duty high speed spec stuff at 150 sq mm - again, all made from hard drawn copper.
--- old post above --- --- new post below ---
On the topic of a 4th rail DC system - it has long been known this is unsuitable for long distance high speed traction, no matter from what angle you approach it (current collection, DC losses, traction bonding, impact of shoes on 3rd rail ramps at high speed... etc...etc...) , it is a poor option and will never happen.

Class 442 holds the world speed record at 108mph and while you might be able to get a little more out of it, it will always be optimal in the 100mph range. I can't see where else such high speeds would be attained on a 3rd rail system, especially in regular passenger service.