Overspeeding at Cambridge Junction (Hitchin) -- 11/12/25

[saismee]

Even easier if you didn’t sign north of Hitchin, and you need a 'feather' to go Cambridge. At least if you signed up to Peterborough too, then you might pay extra attention to where a 'feather' is sending you

Forgive me as I am not a driver and have no operation experience, but wouldn't that still be the driver's fault if they accept a wrong route, similar to accepting a route into a platform which is too short for their train?

 

[bramling]

I doubt there are flashing aspects for that move; more likely the route is approach released from red (AR). From the down fast to slow, or Royston via the flyover, it's more likely a flashing aspect approach applies (AY at the junction signal). Whether the train accelerated towards the junction after slowing for previous cautionary aspects (for various possible reasons) is important in understanding the events, as is visibility, interpretation and possible interruption of route indications on final approach.

The whole Hitchin area is very complex for junction indicators since the flyover opened.

On the down fast, there’s a flashing YY & Y sequence for the move from the down fast to down slow south of Hitchin, which has existed for very many years, and has likewise been a high-speed crossover for many years. The down fast to down slow crossover north of Hitchin was originally a slow speed move, but the crossover was upgraded in preparation for the flyover, and is now also a high speed move both for DF-DS and DF-DRF. For both these moves there is a flashing single yellow sequence only. There’s also a further move onto the down Cambridge branch which is approach released from red.

Meanwhile, on the down slow there’s the diverging route on to the DRF which is a free green aspect, whilst the moves down slow to down fast and down slow to down Cambridge (as is the subject of this incident) are approach released from red, albeit fairly early.

On top of that there’s the PRIs.

Given the complexity, it probably shouldn’t be too much of a surprise that eventually something was going to happen. Indeed I’m surprised the RAIB aren’t taking it a bit further, considering this incident could have had a rather worse outcome.

 

[MarkyT]

The whole Hitchin area is very complex for junction indicators since the flyover opened.

On the down fast, there’s a flashing YY & Y sequence for the move from the down fast to down slow south of Hitchin, which has existed for very many years, and has likewise been a high-speed crossover for many years. The down fast to down slow crossover north of Hitchin was originally a slow speed move, but the crossover was upgraded in preparation for the flyover, and is now also a high speed move both for DF-DS and DF-DRF. For both these moves there is a flashing single yellow sequence only. There’s also a further move onto the down Cambridge branch which is approach released from red.

Meanwhile, on the down slow there’s the diverging route on to the DRF which is a free green aspect, whilst the moves down slow to down fast and down slow to down Cambridge (as is the subject of this incident) are approach released from red, albeit fairly early.

On top of that there’s the PRIs.

Given the complexity, it probably shouldn’t be too much of a surprise that eventually something was going to happen. Indeed I’m surprised the RAIB aren’t taking it a bit further, considering this incident could have had a rather worse outcome.

Thanks for the details. At least with the route set right across the flat junction, blocking everything apart from an up move from the branch, it's unlikely there'd be a secondary collision in any derailment. The damage and clear-up would be very disruptive, however.

 

[WeGoAgain]

From the Slow line, both the right diagonal and the right horizontal would have led over the 25mph crossover. Possible confusion is with the Slow line to flyover (left horizontal)

It's not a Left Horizontal from the slow to the flyover, it's a left 45 degree angle.

RAIB mention the correct signalling sequence being shown on approach. Are flashing aspects used here?

No.

I think it’d be quite easy to think you’re being routed from fast to slow to flyover without checking the junction indicator which unusually showing position 5,

If driver is expecting to go from Fast, to Slow, to Flyover (as is usual), Signal on the Fast would be showing Position 2...?

Forgive me as I am not a driver and have no operation experience, but wouldn't that still be the driver's fault if they accept a wrong route, similar to accepting a route into a platform which is too short for their train?

Partially. - However, in your example, trains stop at Platforms shorter than the length of the Train 'all the time'. Eg. Thameslink 700-1's at Baldock, and 387's/79's at Watlington.

 

[b0b]

I know route knowledge should mean you know that signal has a feather, but as others have said, it does look like it could be easily missed as station lighting. I can understand missing it.

Should also consider if using bars inside of emulating 5 round white lights is also an issue. Could have stood out more maybe.

 

[saismee]

Partially. - However, in your example, trains stop at Platforms shorter than the length of the Train 'all the time'. Eg. Thameslink 700-1's at Baldock, and 378's/9's at Watlington.

I meant more specifically at terminals such as KGX. I know there was an incident posted about that here a while ago.

 

[edwin_m]

It's not as Left Horizontal from the slow to the flyover, it's a left 45 degree angle.

Corrected - thanks for pointing out.

 

[Magdalia]

The BBC now has a report with a picture, though still not very clear, of the relevant signal.

Passengers fell out of seats as train sped at Hitchin junction

The service went through a junction in Hitchin at more than twice the speed limit.

www.bbc.co.uk

Several passengers fell out of their seats when a train "lurched sideways" after passing a set of points at more than twice the speed limit, investigators have said.

The 10:12 GMT London Kings Cross to Ely service on December 11, operated by Great Northern, travelled through the points at Cambridge Junction in Hitchin, Hertfordshire at 56mph (90km/h).

The Rail Accident Investigation Branch (RAIB) said the permanent speed restriction was 25mph (40km/h) and the associated signals displayed the expected indications.

One possible factor, that I think has not been mentioned so far, is low winter sunlight. Contrasts between light and shade, or a reflection, could have affected signal sighting.

 

[zwk500]

Do we know when ETCS will be authorised for Class 378 through Hitchin?

It's obviously still a relevant incident for the rest of the network, but the rollout of ETCS (even as an overlay) will reduce the potential for these kinds of overspeeds.

 

[Peter Mugridge]

I know route knowledge should mean you know that signal has a feather, but as others have said, it does look like it could be easily missed as station lighting. I can understand missing it.

Should also consider if using bars inside of emulating 5 round white lights is also an issue. Could have stood out more maybe.

I've seen these bar type feathers at several location in the past year or two; are they a new standard? It shouldn't be too difficult to make them look like traditional feathers with a template face on them should it?

 

[whoosh]

I know route knowledge should mean you know that signal has a feather, but as others have said, it does look like it could be easily missed as station lighting. I can understand missing it.

Should also consider if using bars inside of emulating 5 round white lights is also an issue. Could have stood out more maybe.

The Up Slow line (towards London) on the other platform at Hitchin, has an old style 5 light feather. I think it mixes in with station lighting quite considerably at certain points on approach.
That's just a 5mph drop in linespeed though, and no wrong route risk.

 

[Royston Vasey]

It's not as Left Horizontal from the slow to the flyover, it's a left 45 degree angle.

Yes, though the driver of that particular service would almost always be on the fast line, and see a horizontal bar for the flyover. He or she, unexpectedly on the slow line, saw a horizontal bar which means the exact opposite, so it's a risk factor for disorientation.

The problems with horizontal bars are that they have two ends, and they are also common lighting features. 45 degree diagonals have only one clear "upwards" direction where it's understood that it acts as an arrow. The horizontal bar "emanates" from the signal head to show direction of course, but 45 degree feathers are less easy to misinterpret IMHO.

These feathers on the platform end signal are an unusual design of bars vs "bulbs", unusually have a horizontal route position, are unusually positioned low down and against station clutter, and are also unusually attempting to indicate three diverging routes from one signal.

You could ask if that route indicator would be better with only diagonal feathers, with the one to the right accompanied by a "CB" or "DF" indicated rather than doing it with a horizontal orientation of white light. It would be a 25 mph PSR regardless.

Or you could program horizontal feathers to strobe directionally like runway approach lights, for additional clarity.

 

[andy33gmail]

I wonder whether you could add variable speed limit signs (like on motorways) to complex junctions. I.e. show 25 when the points are set vs 100 when they’re not.

I guess differential MU/Loco limits possibly mess this up though

 

[Sheridan]

I wonder whether you could add variable speed limit signs (like on motorways) to complex junctions. I.e. show 25 when the points are set vs 100 when they’re not.

I guess differential MU/Loco limits possibly mess this up though

There are some locations where a warning board for a speed reduction is before a junction, and the associated AWS warning is only active if the route has been set towards the lower speed, or if no route has been set at all. If a route has been set that continues on the higher speed route then the warning won’t sound. This isn’t a perfect solution but it minimises unnecessary warnings as a minimum.

 

[edwin_m]

Yes, though the driver of that particular service would almost always be on the fast line, and see a horizontal bar for the flyover. He or she, unexpectedly on the slow line, saw a horizontal bar which means the exact opposite, so it's a risk factor for disorientation.

The problems with horizontal bars are that they have two ends, and they are also common lighting features. 45 degree diagonals have only one clear "upwards" direction where it's understood that it acts as an arrow. The horizontal bar "emanates" from the signal head to show direction of course, but 45 degree feathers are less easy to misinterpret IMHO.

These feathers on the platform end signal are an unusual design of bars vs "bulbs", unusually have a horizontal route position, are unusually positioned low down and against station clutter, and are also unusually attempting to indicate three diverging routes from one signal.

You could ask if that route indicator would be better with only diagonal feathers, with the one to the right accompanied by a "CB" or "DF" indicated rather than doing it with a horizontal orientation of white light. It would be a 25 mph PSR regardless.

Or you could program horizontal feathers to strobe directionally like runway approach lights, for additional clarity.

I believe it's not permitted to have two diagonal indications sloping the same way on the same signal (so not position 1 and 6 or 3 or 4) though some installations pre-date this. In these cases the orientation of the bar or line of lights would be the same and it's only the position relative to the aspect that allows a driver to identify the route. But I don't think there is the same prohibition on the same signal having left and right horizontal indications.

 

[WeGoAgain]

Yes, though the driver of that particular service would almost always be on the fast line, and see a horizontal bar for the flyover. He or she, unexpectedly on the slow line, saw a horizontal bar which means the exact opposite, so it's a risk factor for disorientation.

The problems with horizontal bars are that they have two ends, and they are also common lighting features. 45 degree diagonals have only one clear "upwards" direction where it's understood that it acts as an arrow. The horizontal bar "emanates" from the signal head to show direction of course, but 45 degree feathers are less easy to misinterpret IMHO.

These feathers on the platform end signal are an unusual design of bars vs "bulbs", unusually have a horizontal route position, are unusually positioned low down and against station clutter, and are also unusually attempting to indicate three diverging routes from one signal.

You could ask if that route indicator would be better with only diagonal feathers, with the one to the right accompanied by a "CB" or "DF" indicated rather than doing it with a horizontal orientation of white light. It would be a 25 mph PSR regardless.

Or you could program horizontal feathers to strobe directionally like runway approach lights, for additional clarity.

Agreed on all points. - My post was more about accuracy, but the issues you've raised are all valid.

 

[Royston Vasey]

Agreed on all points. - My post was more about accuracy, but the issues you've raised are all valid.

Indeed, and just shows how objectively poor the design is.

I believe it's not permitted to have two diagonal indications sloping the same way on the same signal (so not position 1 and 6 or 3 or 4) though some installations pre-date this. In these cases the orientation of the bar or line of lights would be the same and it's only the position relative to the aspect that allows a driver to identify the route. But I don't think there is the same prohibition on the same signal having left and right horizontal indications.

From that signal, I don't think that would be needed. You're either diverging at the immediate set of 25 mph points ( / ), or you're continuing on the slow ( ) or ( \ ).

You could then just supplement the route indicator feather with a theatre indicator, something like those detailed here on railsigns.uk. Closest I can find is this Indian example, where the trackwork does vaguely resemble that at Cambridge Junction, although the signal is probably for the line on our right.

 

[Efini92]

I've seen these bar type feathers at several location in the past year or two; are they a new standard? It shouldn't be too difficult to make them look like traditional feathers with a template face on them should it?

I don’t think they’re that new, earlestown has had them for a number of years.

Meanwhile, on the down slow there’s the diverging route on to the DRF which is a free green aspect, whilst the moves down slow to down fast and down slow to down Cambridge (as is the subject of this incident) are approach released from red, albeit fairly early.

Cheers for clarifying, did the driver missread the signal? Looks like it could be easily done from that still shot.

 

[edwin_m]

From that signal, I don't think that would be needed. You're either diverging at the immediate set of 25 mph points ( / ), or you're continuing on the slow ( ) or ( \ ).

You could then just supplement the route indicator feather with a theatre indicator, something like those detailed here on railsigns.uk. Closest I can find is this Indian example, where the trackwork does vaguely resemble that at Cambridge Junction, although the signal is probably for the line on our right.

I agree it's not applicable to that signal, though it does also have a right horizontal bar for the 25mph divergence followed by the next divergence towards Royston. It was just a general point.

The theatre is an interesting idea. I think there's a maximum approach speed above which theatres are deemed not to be readable, so it would only be useable for routes where some approach control condition guaranteed a speed below this. Another alternative was used at Crewe on re-signalling in the 80s and may still be - they had theatre indicators which displayed a number, but also a diagonal bar like a feather for certain higher-speed divergences.

 

[Magdalia]

I've seen these bar type feathers at several location in the past year or two; are they a new standard? It shouldn't be too difficult to make them look like traditional feathers with a template face on them should it?

The signals in question were installed in 2013, before the flyover opened.

 

[choochoochoo]

Do we know when ETCS will be authorised for Class 378 through Hitchin?

It's obviously still a relevant incident for the rest of the network, but the rollout of ETCS (even as an overlay) will reduce the potential for these kinds of overspeeds.

378s? I've heard they’re struggling to get ETCS working on GN 379s.

== Doublepost prevention - post automatically merged: 15 Jan 2026 ==

If driver is expecting to go from Fast, to Slow, to Flyover (as is usual), Signal on the Fast would be showing Position 2...?

Not if they’ve moved to the slow line prior to the signal at the end of the platform. Then on the slow they’d expect a position 1. Which is why if they got 5, they may subconsciously think 'I’ve got a feather I’m all set for the flyover'

 

[MarkyT]

When the flyover opened a junction route indicator was also installed on the approach to Hitchin so the driver should have prior warning which route the train would be taking before the curve leading into the station.I think this was due to the curve preventing advance sight of the actual signal until the train approaches the station.

Likely a preliminary route indicator (PRI), no doubt justified because of the 'free green' approach aspect sequence for the route from the down slow platform signal to the flyover line, which is allowed because the divergence is no more than 10mph slower than the straight on route towards Peterborough (70/75 in this case).

There may be traction /RA/gauge restrictions for the branch applying to certain traffic, for example, that forbids a particular train taking the flyover line, so the PRI gives drivers time to react and pull up before the junction signal in event of misrouting.

Some early installations of PRIs on the South Eastern division were provided for similar reasons, at Continental Junction near Folkestone to prevent domestic passenger traffic being inadvertently sent to the Channel Tunnel, and at Shortlands to avoid Catford Loop freight being misrouted via Herne Hill.

A PRI was also employed on approach to the down main divergence at Heathrow Airport Junction on the GWML, partly to prevent HSTs being inadvertently sent to the airport, but mainly in that case because there were other high speed divergences from the same line nearby that used the more conventional flashing yellow and splitting distant arrangements, and use of another would have violated application clauses banning 'overlapping' aspect sequences employing a flashing single yellow signal for one junction's approach that could also act as a flashing double yellow for a subsequent junction. Much of the justification for PRIs is in avoiding the delays and risks of sorting the problem out if a train had been misrouted and then needed to be extricated backwards after having come to a stop partway through and blocking the junctions involved.

 

[zwk500]

378s? I've heard they’re struggling to get ETCS working on GN 379s.

Sorry, typo for the class number.

 

[MotCO]

The theatre is an interesting idea. I think there's a maximum approach speed above which theatres are deemed not to be readable, so it would only be useable for routes where some approach control condition guaranteed a speed below this. Another alternative was used at Crewe on re-signalling in the 80s and may still be - they had theatre indicators which displayed a number, but also a diagonal bar like a feather for certain higher-speed divergences.

Could a theatre box be used to remind drivers of the speed limit for completing that move? It might be confusing to have it as a speed limit rather than a route, but could it be a different colour?

 

[zwk500]

Could a theatre box be used to remind drivers of the speed limit for completing that move? It might be confusing to have it as a speed limit rather than a route, but could it be a different colour?

It would need to be interfaced with the singalling to only display when the route was called. Potentially would need to be proved lit in the interlocking if judged to be safety-critical.

A very expensive bespoke solution that will be rendered obsolete when ETCS is brought in.

 

[Signal Head]

I think it's the horizontal white bar under the station canopy, above and to the right of the signal. The apparent gap in the white bar may be an artifact of the way LEDs and cameras work, like the way some videos show aspects flashing when they are steadily lit.

I presume this is indicating a route towards the Royston line, further to the right than a simple crossing to the Down Fast, and correct for this train if not using the flyover. There might be a risk of expecting the flyover route and getting the one over the crossovers instead, possibly exacerbated by sighting issues under the canopy.

Many years ago this signal had an odd arrangement with the indicator to the right of the aspect itself. Use of a multi-colour LED would appear to have allowed the aspect to be moved to a more conventional position lower down.

These look like VMS (aka Mallatite) LED PLJIs, the gap in the light bar is at the joint between the 'pivot' and 'arm' sections of the housing.

 

[johnnymo]

The signals in question were installed in 2013, before the flyover opened.

in preparation for the flyover, when a left feather was added

== Doublepost prevention - post automatically merged: 15 Jan 2026 ==

My two questions are why was the Ely service on the slow line and is there any indication before Hitchin station of the route which has been set.

 

[Steve Harris]

in preparation for the flyover, when a left feather was added

== Doublepost prevention - post automatically merged: 15 Jan 2026 ==

My two questions are why was the Ely service on the slow line and is there any indication before Hitchin station of the route which has been set.

If I remember correctly, since Kings Cross PSB was recontroled to York the route has gained ARS. It wouldn't surprise me that ARS had set the route onto the Down Slow (as it was no doubt booked to use the flyover) and the signaller saw that there was a gap in Up traffic and he could intervene and send this service "old road" and overtake the stopping service into the bargain.

I've watched Signalmaps and seen the old road used to regulate trains on a few occasions* (normally when a Brighton- Cambridge service is running late and will hinder the Ely/Kings Lynn express/Cambridge Cruiser service).

* Of course this depends on there being a suitable gap in up services to get the down service across the old flat junction without delaying any up service.

 

[MarkyT]

in preparation for the flyover, when a left feather was added

== Doublepost prevention - post automatically merged: 15 Jan 2026 ==

My two questions are why was the Ely service on the slow line and is there any indication before Hitchin station of the route which has been set.

IB82, in front of 1T20 on the down fast, stopped as scheduled at Stevenage. On leaving it then came to a stop at the next signal for reasons unknown. 9S14 departed Stevenage soon after on the slow and overtook IB82. 1T20, was switched into platform 4 behind 9S14 and followed it to Hitchin. 9S14 took the flyover line to allow 1T20 to overtake it via the flat junction. Soon after 1T20 had cleared the junction (rather more quickly than normal clearly!), 1B82 approached from the south, so whatever problem had occurred at Stevenage had been rectified.

The events are more easily understood by watching the rewind map for the day in question at Vail Data:

Maps - Vail Data

vaildata.uk

I suggest setting the replay speed to x10 or more.

 

[Taunton]

but could it be a different colour?

Isn't the easier "different colour" approach to have the signal with the junction indicator show single yellow instead of green?

It's the same as the Peterborough incident, a sharp and unexpected 25mph turnout and the signal showing full green. In semaphore days, if such a starter had a distant on the post as well, would the distant be pulled off for a route over the twisting crossovers ...?