Inaccurate platform clocks

[Llanigraham]

If it causes a passenger to miss a connection that they should have (just) been able to get, then yes it does matter.

But this is JUST a 24 second discrepancy!! And NO trains are timed that accurately, so I fail to see how anyone can miss a connection on that timing.

Sorry but all I see here is a few people finding any excuse to blame the railway for what is really nothing! :roll::roll::roll::roll::roll:

 

[mallard]

But this is JUST a 24 second discrepancy!! And NO trains are timed that accurately, so I fail to see how anyone can miss a connection on that timing.

You have a 5 minute connection between trains. The first train arrives 90 seconds late, takes another minute to get the doors open (the guard was busy selling a ticket as the train arrived). The second train closes and locks its doors 40 seconds before departure by station clock that the guard is looking at.

That leaves you with just 110 seconds to get from the first train to the second. If you take away another 24 seconds, you've only got 86. Tough to make it to another platform even for a reasonably fit person with no luggage. Impossible for a good proportion of passengers, including anybody with mobility problems, those with children, etc.

 

[ComUtoR]

You have a 5 minute connection between trains. The first train arrives 90 seconds late, takes another minute to get the doors open (the guard was busy selling a ticket as the train arrived). The second train closes and locks its doors 40 seconds before departure by station clock that the guard is looking at.

That leaves you with just 110 seconds to get from the first train to the second. If you take away another 24 seconds, you've only got 86. Tough to make it to another platform even for a reasonably fit person with no luggage. Impossible for a good proportion of passengers, including anybody with mobility problems, those with children, etc.

If the first train fives 280 seconds late then your totally screwed but that's another dumb analogy for you.

Another one is that the Guard is looking at the Platform Clock and not his company issued watch or is being dispatched by a member of staff using their company issued watch. Or the Driver, me, is looking at their company issued watch.

But that's all kinda irrelevant because when I get to a station. I'm looking at my own personal watch/phone/timepiece/app to ensure I arrive on time.

*edit*
Just a weird thought. If a train departs platform A at the right time, as displayed on the platform clock and a train departs platform B right time as displayed on the platform clock but the platform clocks differ. Are both trains still on time ?

 

[Bletchleyite]

It really shouldn't be hard in 2017 for all clocks to be connected to an NTP server via whatever the PIS uses.

If there are some out of sync with the PIS, remove them. If the PIS is out of sync, fix it!

 

[embers25]

It really shouldn't be hard in 2017 for all clocks to be connected to an NTP server via whatever the PIS uses.

If there are some out of sync with the PIS, remove them. If the PIS is out of sync, fix it!

Exactly, surely this should be simple. Today the maximum difference between en route was 1.5 minutes which I agree isn't massive but it should be 0 as they should all be linked centrally. At least stations from the same TOC. Also trains not appearing at all on PIS on platforms shouldn't happen...particularly when the next train advertised is 90 minutes later. If an advance ticket holder left the station based on that platform PIS to get a snack etc technically they'd be in trouble if they boarded the next train.

 

[WelshBluebird]

But this is JUST a 24 second discrepancy!! And NO trains are timed that accurately, so I fail to see how anyone can miss a connection on that timing.

Sorry but all I see here is a few people finding any excuse to blame the railway for what is really nothing! :roll::roll::roll::roll::roll:

As I have said previously, there have been quite a few times where I have made a connection by literally just a few seconds due to my incoming train being late. If the connection had left 24 seconds early, I would not have made those trains, and in at least a few cases I would have had a fairly long wait (1h+) for the next service and thus been claiming money back from the ToC. So fine, if companies want to open themselves up to avoidable delay repay claims then that is up to them!

 

[smilerish]

Exactly, surely this should be simple. ... At least stations from the same TOC.

Surely NR should be responsible for infrastructure like clocks? If nothing else, they're held accountable for train movements across the network and if the trains are all running to different time pieces, the timetable becomes a shambles.

On the subject of which: from a performance perspective, trains may have a 5 minute window before they're 'late', but the WTT is written to 1/2 minute precision. So surely the clocks by which the network is run should be at least that accurate - even if it's just a clock for passengers to consult?

 

[455driver]

Considering that it sometimes takes a little while (I've experienced more than 2 minutes) for doors to be released after a train arrives,

How often does that really happen? :roll:

 

[Shaw S Hunter]

It really shouldn't be hard in 2017 for all clocks to be connected to an NTP server via whatever the PIS uses.

If there are some out of sync with the PIS, remove them. If the PIS is out of sync, fix it!

All well and good but you try finding senior management in a TOC or Network Rail that actually cares very much about PIS! Those that have any genuine interest in customer service are much more switched on to the "touchy-feely" part of the experience and technical stuff involving hardware seems to be completely beyond their understanding so they wash their hands of it. And it's nothing new, it was just the same in BR days.

 

[Tio Terry]

Automated railway clocks were originally driven by half minute impulse systems and their accuracy was largely down to the local telecomms staff checking the BT closk against their own impulse system. The former BR Southern Region tried to improve this by using Waterloo as the governing time signal that was distributed across the Region via the telecomms transmission systems, it wasn't that successful.

Later systems relied on the Rugby Time Signal. A radio signal transmitted from the national Atomic time clock and were very reliable. It's a 60khz radio signal that can be received by anyone within the UK and other parts of Europe.

In 2007 the so called "Rugby Clock" was relocated to Anthorn in Cumbria, that did cause some problems with signal reception at some locations. When most clocks lose their MSF signal they rely on an internal time clock to provide time information. If they receive a radio signal they will update their time, but if they don't they are reliant on the accuracy of their internal source. This may explain the discrepancy between clocks, some are still receiving MSF signals, others are running on internal timing clocks.

More modern systems will derive their timing from either the internet or from satellite navigation systems. But the same applies, it they lose their external time sources they will rely on their internal systems which may not be terribly accurate so it will lead to differences to time displays across a given site.

 

[fowler9]

But this is JUST a 24 second discrepancy!! And NO trains are timed that accurately, so I fail to see how anyone can miss a connection on that timing.

Sorry but all I see here is a few people finding any excuse to blame the railway for what is really nothing! :roll::roll::roll::roll::roll:

With all due respect I think I mentioned that the clocks on platform 2 at Liverpool South Parkway are about half an hour out and have been for a while.

 

[Tim R-T-C]

Manchester Airport clocks have been wrong for well over 12 months and no two clocks on the station show the same time. Some are hours out.

Yes, I was there last month and one clock was correct, one was an hour slow and the other was about 6 hours out.

An hour wrong might not be an issue in many places, but at an airport where people are 'zonked' after a long flight and have a booked advance ticket, it could lead to them getting on the wrong train.

Surely there should be a regulation that clocks out by more than a minute, that cannot be readily fixed, should be switched off. Better no time than the wrong time.

 

[SpacePhoenix]

Could the GSM-R system have a time signal piggybacked onto it? There would probably be some stations with poor coverage but it would (hopefully) cover most stations and locations

 

[Elecman]

In theory there should be no poor coverage within GSMR unless a couple of adjacent masts are down. Isn't GSMR now classed as Safety Critical equipment?

 

[3141]

I'd like to thank Shaw S. Hunter #23 and Tio Terry #39 for their interesting and informative posts.

I agree with Llanigraham that 24 seconds probably won't make much difference, though other have pointed out how it might in certain circumstances. Such circumstances are rare and unlikely, but it's interesting that when a safety issue comes up , and someone says that's rare and unlikely to happen, others immediately answer that one day there'll be a combination of factors and it will. Obviously safety and accurate time-keeping aren't the same (mostly), but someone could miss a train they ought to have been able to catch.

I think a very important factor is attitude. If 24 seconds difference doesn't matter much, what about 48....or 96? When a public service is being provided there ought not to be any discrepancies. When clocks are just one second apart, as I saw them at Vauxhall earlier this week, it's not worth making a fuss about, but if the difference builds up then somebody should be concerned enough about such things to want to put it right. The posts I've mentioned show that that can be surprisingly difficult.

I certainly didn't start the thread in order to put blame on the railway. But the railway justifiably expects me to be on board the train at least 30 seconds before it's due to depart, and not to think I have a right to make a connection in less then five minutes at most stations, and I think the public can justifiably expect published information to aim at 100% accuracy. That's something that human beings cannot always achieve, but when they don't they should try to correct the situation in a "timely" way.

 

[edwin_m]

In theory there should be no poor coverage within GSMR unless a couple of adjacent masts are down. Isn't GSMR now classed as Safety Critical equipment?

But how many approval hoops would someone have to jump through to get a new feature added to a Euro-standard safety-critical system? At the end of the day* it would probably have just as much chance of inadequate coverage as other time sources.

*assuming the clocks could agree when the end of the day actually was.

 

[satisnek]

I've always assumed that modern platform clocks were on MSF time and have never looked into this... I've observed that trains can depart up to 30 seconds before 'time', very rarely (if ever) before this and if you miss one (that's not a connection) then it's your own stupid fault.

But it hasn't always been this way... Worcester Foregate Street Platform 1 in Thames Trains days, analogue clock was two minutes fast and we walked onto the platform to see two red tail lights disappearing into the distance... The station staff had despatched the train 'on time' according to the platform clock and were adamant that it was showing the correct time. And I was damned sure it wasn't - my digital watch (which went 'beep' on the hour) was within 5 seconds of the BBC time signal that morning. A right old ding-dong ensued.

Thank goodness those days are over.

 

[notlob.divad]

All this reference to clocks recieving radio signals to syncronise is all well and good and very true, but it surely misses a fundamental point, all the clocks in the same station could very easily be kept to the same time. Each station only needs one master to recieve the radio signal and everything else on the display screens can sync to that clock. Even if the radio signal isn't recieved, all the clocks in the station can stay at the same time.

With npt protocols and the like, this really isn't hard and has existed for years. Even before NPT there was NET TIME a DOS based command that managed to sync computer times over ethernet or serial buses and I know Lynux based systems have a similar function.

I can understand times between different stations and on a train potentially being slightly different, but there is no excuse for displays at the same station showing different times on different platforms.

 

[fowler9]

Checked Platform 2 at Liverpool South Parkway today and it is still roughly half an hour slow. None of the PIS on the platform work either. Platforms 1, 3 and 4 all have about a minute difference as well. The PIS on the bridge over the platforms all keep showing trains overtaking each other as well. If you don't know what the train you are after is meant to look like it is like a lucky dip.

 

[tsr]

The newer of the bridges at East Croydon is completely lacking any sort of synchronised time - most of the CIS screen clocks are totally out of sync with each other. Don't worry, that sort of thing is audited, so one day this failure will be too expensive to ignore...

As for clocks onboard trains, these vary from "nearly correct" to "wildly inaccurate" in my experience, depending on what the TMS needs to function, if there is a TMS. I'd never, ever use them for dispatch.

 

[DaveTM]

From the date in the bronze age at which people started to realise the earth was approximately spherical, they realised astronomical time varied from place to place on Earth. When the railway reached Bristol for instance, a clock was built with two minute hands, one giving the correct (astronomical) time in London, one the correct (astronomical) time in Bristol. Railways (and the telegraph) brought time synchronisation, and the concept of an entire country running on one "mean" time - GMT/BST. We can now syncronise our time pieces with an agreed standard clock, using radio transmitters or UDP packets over the internet. I can buy a £10 clock radio that is correct to better than an eye blink if it is powered up within 500 miles of Cumbria.
So why effing why do my Cumbria controlled wrist watch, my train, and my station, all disagree on the time? My train runs windows XP embedded and synchronises with machines at the depot several times a week. Surely it was not beyond the abilities of the architects to consider putting an NTP daemon on MITRAC and punching the correct holes in the firewall. My train also decides how many doors to open based on a GPS receiver which relies on a clock accurate to the speed of light divided by an accuracy of about 10 m. Surely it is not beyond the abilities...
Personally, I decide when to dispatch based on the clock shown on my MITRAC display, as I hope this will be the time shown on the timestamps should the data recorder on my train ever have to be downloaded. I am assuming however that two computers on my train agree on the time. If this is not true, I really hope it is not beyond the abilities of the investigators to allow for the discrepancy...

 

[PermitToTravel]

Do you really want to know?

The main clock on the concourse we have is set by... the precise method of turning it on at three seconds to midnight and then it turns on and starts counting from 00:00:00.

Only on the railway would you have a clock that you cannot change the time

That is fairly unique though, many get their time from the computer/server in charge which I presume uses NTP to keep the time correct.

Some friends have an ex-railway clock that I think is of the same type as the one you're talking about.

It can't be set, and so your method of getting it to show the right time is often used. It does however have a connected antenna - it's incredibly temperamental, but if it's pointing in the right direction at the right moment then the clock will set itself from the MSF signal.

I wonder how many stations have a roof mounted antenna pointing at Rugby? The clocks would have set themselves until April 2007, and now either always show the wrong time or depend on being reset at the right moment.

Sometimes the clock really loses it and makes a good forum avatar

 

[smilerish]

With npt protocols and the like, this really isn't hard and has existed for years.

Indeed - if you're referring to NTP, it was created in 1985. It started on Unix and spread from there. As far as I know, it's the only network time standard - programs like NET TIME just implement the protocol (although according to Wikipedia, Windows didn't reliably implement the protocol until Vista).

I can buy a £10 clock radio that is correct to better than an eye blink if it is powered up within 500 miles of Cumbria.
So why effing why do my Cumbria controlled wrist watch, my train, and my station, all disagree on the time? My train runs windows XP embedded and synchronises with machines at the depot several times a week... My train also decides how many doors to open based on a GPS receiver...

Windows XP's time sync has always been iffy, assuming it's even turned on.

Personally, I decide when to dispatch based on the clock shown on my MITRAC display, as I hope this will be the time shown on the timestamps should the data recorder on my train ever have to be downloaded... If this is not true, I really hope it is not beyond the abilities of the investigators to allow for the discrepancy...

I think this is the most concerning 'discovery' of this thread. If it did come down to an accident, and RAIB are piecing together evidence from various time-coded sources, then what assurance is there that their timeline will be correct when all the clocks show different times? And in such circumstance, fractions of a second really do make a difference.

 

[Llanigraham]

We had a "radio controlled" clock in my signal box, but found that it would only work in one place. As soon as you moved it it lost time.
I've also heard of some boxes where the clock wouldn't get a signal at all.

And just remember that all this technology people are requesting costs money, and the only people who will ending up paying at are the passengers and the tax payer; is that what you want?

 

[700007]

I was at Seven Sisters earlier this week where the Underground and Overground clocks were showing two different times. Underground said 1126 whilst Overground was 1132. I was so confused on which one was correct as I needed an Overground connection at 1133, which luckily I made by running for it. Turns out the Overground time is correct! No idea why the Victoria's clock is so far behind.

 

[edwin_m]

Do any station (non-platform) clocks still adopt the historic practice of being deliberately set a few minutes fast, to encourage people to hurry for their trains?

 

[iknowyeah]

In all these hypothetical scenarios, why do passengers have no way of telling the time until they're 30 seconds away from departure and already standing on the platform?

 

[Shaw S Hunter]

Indeed - if you're referring to NTP, it was created in 1985. It started on Unix and spread from there. As far as I know, it's the only network time standard - programs like NET TIME just implement the protocol (although according to Wikipedia, Windows didn't reliably implement the protocol until Vista).

Windows XP's time sync has always been iffy, assuming it's even turned on.

This looks like a winner. There's still a lot of computers around running XP, remember the NHS cyber-attack?

 

[causton]

Some friends have an ex-railway clock that I think is of the same type as the one you're talking about.

It can't be set, and so your method of getting it to show the right time is often used. It does however have a connected antenna - it's incredibly temperamental, but if it's pointing in the right direction at the right moment then the clock will set itself from the MSF signal.

I wonder how many stations have a roof mounted antenna pointing at Rugby? The clocks would have set themselves until April 2007, and now either always show the wrong time or depend on being reset at the right moment.

Sometimes the clock really loses it and makes a good forum avatar

Yes, the clock is the same style as your avatar.

We can barely get phone signal, the station being under a massive office block, so I can understand the clock not being able to find any signal!

 

[work-traveller]

Looking at the photo of the OP, the clock on the left side is on the Metrolink platform and not the Network Rail platform. That island platform is split use (further away side - Metrolink, nearest side - Network Rail). So, it's possible that the two clocks aren't synchronised because they are controlled by independent systems. Not a great situation all the same. The whole site is owned by Manchester Airport with a split lease between Network Rail and Metrolink.