Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
Well it was suggested in the Manchester - Liverpool Electrification thread that we have a seperate Transpennine thread. Now I know that Manchester - Stalybridge is actually part of the Northern Hub project , but to me, haulage bashing Peaks in the 1970s-1980s - Liverpool-Newcastle was classed as...
Indeed and have been absolutely transformational in DC areas eliminating huge amounts of cabling and impedance bonds although never got why Thameslink didn't adopt them
Indeed and have been absolutely transformational in DC areas eliminating huge amounts of cabling and impedance bonds although never got why Thameslink didn't adopt them
It’s because the TC sections are so short, and the service so frequent, the probability of a train stopping with a wheel on a counter triggering a miscount is quite high.
It’s because the TC sections are so short, and the service so frequent, the probability of a train stopping with a wheel on a counter triggering a miscount is quite high.
I would have thought that issue is a risk wherever axle counters are installed. ie you can't control where a train stops albeit i get the probability would be much increased with short sections needed for ATO to provide close headways. Anyhow stick them everywhere else is now largely the default option although I believe across complex S&C with lots of permutations siting the counters can be issue that means track circuits remain necessary
I would have thought that issue is a risk wherever axle counters are installed. ie you can't control where a train stops albeit i get the probability would be much increased with short sections needed for ATO to provide close headways. Anyhow stick them everywhere else is now largely the default option although I believe across complex S&C with lots of permutations siting the counters can be issue that means track circuits remain necessary
When Nottingham was re-signalled in about 2013 track circuits were used in the platforms and axle counters elsewhere. Someone pointed out on another thread that there is now a solution to the issue of spurious occupations, where both sections remain occupied but can now be cleared when the whole train is proved to have passed the next axle counter down the line. However, perhaps the heavy use of the Thameslink core means even that imposes too much operating restriction?
The issue would have been the need to split the detection within the platform. Multiple platform occupancy doesn't require this unless standards have changed recently, but it's often provided. I think the solution I mentioned would have been employed had Nottingham been re-signalled more recently, and this would largely have avoided the problem.
The issue would have been the need to split the detection within the platform. Multiple platform occupancy doesn't require this unless standards have changed recently, but it's often provided. I think the solution I mentioned would have been employed had Nottingham been re-signalled more recently, and this would largely have avoided the problem.
IIRC Bournemouth was pretty much the pilot for axle counters in the UK, done under Railtrack. and was about as successful as you might expect it to have been.
...apart from re-starting from a shut-down of some sort (any sort) , when the railway has absolutely no idea of what is on the line or where it is. (Unlike track circuits.)
The first train has to run through (at caution, presumably, slow enough to be able to stop short if another vehicle or a blockage is encountered.) Bad luck on the driver of that first train, or all the pax and others on trains queued up following the first one...
I know Merseyrail moved to axle counters in the Loop because it was so wet and salty that conductivity across the sleepers meant that track circuits didn't work.
Axle counters dont:-
Detect broken rails
Detect metal obstructions, notably wreckage from a train smash which occured on the other track.
Dont detect track curcuit clips.
Dont need lines and lines of computer code and shedloads of configuration all input by humans so liable to wrong side failure because of bugs or config errors.
But they are cheap avoiding all the stuff on the track to make track circuits failing.(insulating the 2 rails electrically,,coping with interferance from electric trains etc)
Axle counters dont:-
Detect broken rails
Detect metal obstructions, notably wreckage from a train smash which occured on the other track.
Dont detect track curcuit clips.
Dont need lines and lines of computer code and shedloads of configuration all input by humans so liable to wrong side failure because of bugs or config errors.
But they are cheap avoiding all the stuff on the track to make track circuits failing.(insulating the 2 rails electrically,,coping with interferance from electric trains etc)
Track circuits don't always detect broken rails either, depending on how the bonding is done.
Axle counters don't let train disappear if there is poor wheel-rail conductivity.
The computer code on axle counters is all standard, and I suspect modern track circuits include a fair amount of software too. Either will require configuration data in the interlocking.
Rail insulation is still provided in electrified axle counter areas, to direct return current through the rails and not through the ground. In fact I believe the pad that provides the insulation is even used on non-electrified lines, as it also provides some cushioning underneath the rail.
Axle counters dont:-
Detect broken rails
Detect metal obstructions, notably wreckage from a train smash which occured on the other track.
Dont detect track curcuit clips.
Dont need lines and lines of computer code and shedloads of configuration all input by humans so liable to wrong side failure because of bugs or config errors.
But they are cheap avoiding all the stuff on the track to make track circuits failing.(insulating the 2 rails electrically,,coping with interferance from electric trains etc)
or the type of break. Some types of damage that could leave the rail structurally unsound or with a chunk of head surface missing may not be detected. Bonding design is always a compromise and there are inevitably spurs of parallel rail that cannot be continuity monitored, usually in pointwork where the risk of damage may be higher.
The major benefit with axle counters is that all the rails, on parallel tracks can be bonded together every so often to maximise parallellism in the return path for traction current, reducing overall loop resistance and providing maximum redundancy to guard against faults states diverting current via higher resistance paths through the ground and surrounding metalwork, which, with the very high currents of lower voltage DC systems in particular, can be very damaging.
== Doublepost prevention - post automatically merged: ==
It’s because the TC sections are so short, and the service so frequent, the probability of a train stopping with a wheel on a counter triggering a miscount is quite high.
I think they also did an analysis of total system complexity. With so many short occupancy sections in the final configuration, there would have been very large quantities of sensors and other components. In this specific case, with some individual sections as little as 50 to 70m long, including multiple platform splits (i.e. much shorter than the typical trains), the complexity alone was seen as a reliability/performance issue. The miscount thing would also have been a huge problem on such an intense operation of course. The 'supervisory section' solution, common in Germany, the home of axle counting tech, will premier at Birmingham New Street, which is supposed to be resignalled 'very soon'.
== Doublepost prevention - post automatically merged: ==
IIRC Bournemouth was pretty much the pilot for axle counters in the UK, done under Railtrack. and was about as successful as you might expect it to have been.
ISTR Bournemouth seafront was the pilot for many particular Siemens products used, but axle counters more generically have been used in certain niches in the UK since the late 1980s. The concept actually emerged very early possibly pre-WW1, but mechanical treadles switches and relay-based counters, shift registers etc in early examples were never safe or reliable enough for widespread use before WW2. The tech really took off when inductive sensors and solid-state electronics allowed complexity and cost of implementation to fall dramatically in the 1950s. They have been totally bog standard everywhere in German-speaking central European countries since as far back as the 60s or 70s.
Early UK applications:
- Long single lines where a pair of sensor installations with their modems, an evaluator rack and some long cables can replace many individual maintenance-intensive track circuit sections spread out along the infrastructure (no issue with TC clip protection of adjacent lines either)
- Big metal bridges where it can be difficult to keep TCs working effectively, especially if there's salt spray involved
- Sea wall environments for similar reasons to bridges above
- Other sites where railhead contamination was a problem for TCs
== Doublepost prevention - post automatically merged: ==
And something that can be largely designed out of modern systems by specifying local uninterruptable power supplies wherever concentrations of processor-based equipment are located, such as at axle counter evaluator sites.
IIRC Bournemouth was pretty much the pilot for axle counters in the UK, done under Railtrack. and was about as successful as you might expect it to have been.
...apart from re-starting from a shut-down of some sort (any sort) , when the railway has absolutely no idea of what is on the line or where it is. (Unlike track circuits.)
The first train has to run through (at caution, presumably, slow enough to be able to stop short if another vehicle or a blockage is encountered.) Bad luck on the driver of that first train, or all the pax and others on trains queued up following the first one...
I know Merseyrail moved to axle counters in the Loop because it was so wet and salty that conductivity across the sleepers meant that track circuits didn't work.
Axle counters dont:-
Detect broken rails
Detect metal obstructions, notably wreckage from a train smash which occured on the other track.
Dont detect track curcuit clips.
Dont need lines and lines of computer code and shedloads of configuration all input by humans so liable to wrong side failure because of bugs or config errors.
But they are cheap avoiding all the stuff on the track to make track circuits failing.(insulating the 2 rails electrically,,coping with interferance from electric trains etc)
Track circuits don't always detect broken rails either, depending on how the bonding is done.
Axle counters don't let train disappear if there is poor wheel-rail conductivity.
The computer code on axle counters is all standard, and I suspect modern track circuits include a fair amount of software too. Either will require configuration data in the interlocking.
Rail insulation is still provided in electrified axle counter areas, to direct return current through the rails and not through the ground. In fact I believe the pad that provides the insulation is even used on non-electrified lines, as it also provides some cushioning underneath the rail.
You will have to explain more. In my extensive experience, it’s very rare for a track circuit (not including sequential types) to fail to work when power returns. D&P throwing the isolation switches off and on too quickly without pausing will blow fuses. But that can affect anything…
So of the pros and cons of axle counters do depend on the actual technology of that particular type of axle counter AND on the railway operating rules and procedures in force at the time.
The older design of axle counters could not cope with very slowly moving wheels, or wheels being stopped on or very near the axle counter head (sensors).
More advanced technology now enables a work around for this problem most of the time.
Similarly, for 99.9999% of the time, after a total loss of power (black out, power blip, planned switch off, whatever), within 60 seconds (or quicker) track circuits will recover to normal operation (does not include sequential track circuit systems).
Axle counters have to be provided with back up batteries to maintain their operation if you want a fast recovery from a loss of electrical power. If the time without power exceeds the back up battery capacity, then each axle section needs to be reset.
With older designs, that means the S&T attending on site. More modern designs can be reset by the signaller. But aspect restrictions may apply.
Axle counters are typically more expensive for short sections than track circuits. But more competitive for long sections (where on the track, multiple track circuits may be needed, even if the signaller only has one track circuit indication on their panel / display).
Back up batteries obviously further increase the cost.
Broken rails. In non-electrified areas, most track circuits are of the double rail type. The vast majority of broken rails occur in LWR / CWR (long welded rail / continuous welded rail). In the vast majority of cases, the rail breaking leaves a gap of at least 3mm or more. Hence in these cases, the vast majority of the time, the track circuit will fail due to the broken rail.
There are not many broken rails in 60 foot areas, or in or around switches and crossings. In these areas, broken fish plates or star cracks in the rail where the bolt holes for the fish plates are, are more common. And occasionally broken crossings (S&C). Obviously track circuits can’t detect any of these defects. Especially as bond wires are used to bypass fish plate joints.
Axle counters can fall off the rail, and hence in theory, could miss a train axle (modern types are supposed to detect this problem). But no axle counter system can detect when contractors or other track staff cut out or remove a rail with the count head still on it, and then place the rail (complete with count head) by the side of the line.
Axle counter heads can and are affected by metal objects near the count heads. Normally this causes a failure of the axle counter system.
Axle counter heads can, and are damaged by various on track plant and rail grinding machines. Track circuits generally are not affected by on track plant and rail grinding machines.
Both axle counters and track circuits can have their cables cut by rail tamping engineering machines (tampers) or ballast regulating engineering machines (regulators).
Axle counters are great for areas where the ballast is contaminated or very wet. In these conditions, it’s hard to get a track circuit to work reliably.
In summer, on hot days, failure of the insulations in insulated rail joints (IRJ) that electrically separate different track circuit sections often causes track circuit failures. Some of which have to wait until temperatures drop in the early morning before the permanent way are prepared to take the fish plate off in order to renew the insulation pieces.
Axle counters however, are based on complex electronics. If the case of the enclosure is not weathertight (for any reason), and water gets in, they fail. This also means that maintenance, repair or installation work is far more difficult if it’s raining. This is much less of a problem with track circuit equipment.
Regarding computer code. Conventional track circuits don’t have any microcontroller or CPU. Only the EBI Track 200 (formerly TI21 Digital) joint less (for use on LWR/CWR track) and similar modem types use microcontrollers/CPUs.
Older designs of axle counters also don’t use microcontrollers/CPUs. But later, modern versions do use microcontrollers/CPUs.
Older designs of interlocking also don’t use microcontrollers/CPUs. But obviously computer based systems (such as SSI - Solid State Interlocking) do.
In my area, each computer based system is isolated and independent as far as the software is concerned. But this is technology that is over five years old. So the latest systems may be more integrated. Having separate systems does have the advantage that the software is less complex and easier to design and easier to test.
Axle counters require dedicated multi pair telecommunications cables are are somewhat sensitive to defects in these cables or the connectors.
Track circuits can use the existing signalling cables. But are subject to the normal standards that apply to these cables.
In terms of maintenance, the modern way is not to have any! Both for axle counters and for EBI Track 200 (formerly TI21 Digital) joint less track circuits. Other conventional track circuits or older designs of axle counters do require maintenance. The amount of time varies depending on the actual equipment and the actual application.
With faults and fault finding/rectification, again, it depends on the equipment and the specifics on site.
Axle counters have a disadvantage that you have to go to the site of the evaluator (often a REB located nowhere near the actual axle counter section) then normally you then have to go to the site of either the count in, or count out head. Or to the many telecom dis.boxes to check the telecommunications cable connections.
If a axle counter head is damaged, or the electronic junction box is suspect, that’s between half an hour to two hours to change these parts (assuming you have suitable spares).
Then you have to go back to the REB to reinstate the isolation links.
With a track circuit that is a single track circuit section, an experienced techinican can locate the cause of the fault in about ten minutes or less. Repair time varies depending on what the problem is.
It’s somewhat different if the track circuit is actually multiple sections and it’s failing intermittently… then it can take hours to locate the problem.
It should however be pointed out that despite the majority of track circuit failures not being due to any track circuit equipment failing (*), the average reliability of track circuits is better than 99.999%. Some track circuits have never failed in their entire working life.
* Note: the majority of track circuit failures are due to cables being damaged, faulty IRJs, track tie bars being incorrectly fitted, metal objects causing short circuits between adjacent rails in S&C/point work, poor ballast conditions (salt, mud, coal, coke, poor drainage/flood water or a combination).
Conventional track circuits are also likely to have a longer service life.
The biggest advantages for axle counters, is that you don’t need IRJs, they can be fitted on electrified lines without additional equipment, they can be fitted on lines where track circuits already exist, making changeover from an existing signalling system to a new one quicker and easier.
The biggest disadvantages are that a failure of an axle counter evaluator can cause an entire area to loose all axle counters for that area. That could be fifteen or more axle counter sections. Similarly loss of one telecommunications cable could affect large numbers of axle counters.
It’s very, very, very rare for that many track circuits to fail due to a single cause.
My conclusion: there is no simple winner here.
== Doublepost prevention - post automatically merged: ==
or the type of break. Some types of damage that could leave the rail structurally unsound or with a chunk of head surface missing may not be detected. Bonding design is always a compromise and there are inevitably spurs of parallel rail that cannot be continuity monitored, usually in pointwork where the risk of damage may be higher.
Always been the case. But are you really saying that there are more broken rails (rather than chunks breaking off) in S&C than in plain line (LWR/CWR)?
Some track circuit interrupters are still used in axle counter areas, I think (can’t check at the moment) because axle counters are not always provided to fully cover sidings.
I think they also did an analysis of total system complexity. With so many short occupancy sections in the final configuration, there would have been very large quantities of sensors and other components. In this specific case, with some individual sections as little as 50 to 70m long, including multiple platform splits (i.e. much shorter than the typical trains), the complexity alone was seen as a reliability/performance issue. The miscount thing would also have been a huge problem on such an intense operation of course.
And something that can be largely designed out of modern systems by specifying local uninterruptable power supplies wherever concentrations of processor-based equipment are located, such as at axle counter evaluator sites.
Yeah, except that what the railway calls a UPS and what is actually a UPS appear to be rather different…
The AzLM axle counter type in our area are actually fed off a battery charger that includes a number of batteries internally. Hence the DC output voltage is not regulated like it would be for a true UPS. And in the event of a charger failure, there is no automatic bypass, again unlike most good quality UPS systems. Worse, this glorified battery charger design is extremely heavy and awkward to change. And has MCBs mounted on the outside where it’s easy to knock one. Don’t ask how I know… Let’s just say that it was lucky that the area covered by that evaluator (about ten axle counters) had no booked train service when it failed…
Frankly, some of the poor design of modern signalling installations is disappointing, to say the least. The first summer after we had multiple evaluator failures, we were told to remove all the cubicle doors from the evaluator cubicles. As they were overheating….
The use of axle counters in Britain goes back at least as far as the 1930s, the early installations comprising uniselectors linked to electro-mechanical treadles. Electronic axle counters first appeared on BR in the 1970s. Axle counters with 'dekatron' gas counting tubes had been provided at Glasgow Queen Street station in 1967, but were replaced by Az L65 counters in the 1970s when manufacturing of the tubes ceased. Az L70 axle counters were installed on the Forth Bridge in 1979.
Trains generally stop for a reason, such as a signal or a station. Yes sometimes breakdowns or emergency stops happen, but they are the exception not the rule. Given planned train lengths and stopping reasons, you can presumablly work out the range of locations that the trains are likely to stop.
With long sections, I think the axle counters will naturally end up at locations where trains are unlikely to stop. If a train is passing the axle counter to enter a section, then that means it has been cleared to enter said section.
The problem presumably comes when things are too close together, so a train can be cleared to enter a section, but is not able to fully enter it before reaching the next stopping point.
The tech really took off when inductive sensors and solid-state electronics allowed complexity and cost of implementation to fall dramatically in the 1950s. They have been totally bog standard everywhere in German-speaking central European countries since as far back as the 60s or 70s.
As they are all countries with extensive low-frequency AC electrification, do track circuits have particular problems in that situation which led to extensive use of axle-counters as an alternative?
Some track circuit interrupters are still used in axle counter areas, I think (can’t check at the moment) because axle counters are not always provided to fully cover sidings.
Yeah, except that what the railway calls a UPS and what is actually a UPS appear to be rather different…
The AzLM axle counter type in our area are actually fed off a battery charger that includes a number of batteries internally. Hence the DC output voltage is not regulated like it would be for a true UPS. And in the event of a charger failure, there is no automatic bypass, again unlike most good quality UPS systems. Worse, this glorified battery charger design is extremely heavy and awkward to change. And has MCBs mounted on the outside where it’s easy to knock one. Don’t ask how I know… Let’s just say that it was lucky that the area covered by that evaluator (about ten axle counters) had no booked train service when it failed…
There have also been issues where some designs have considered actually wiring up the power failure alarm contacts within in the charger to be optional and not bothered. This then had to be retroactively fitted as it transpired that being entirely unaware that the axle counter was running on limited battery power wasn't an ideal situation.
Frankly, some of the poor design of modern signalling installations is disappointing, to say the least. The first summer after we had multiple evaluator failures, we were told to remove all the cubicle doors from the evaluator cubicles. As they were overheating….
I remember there were also initial problems where the installers were untwisting the twisted pairs to make them easier to terminate. Being unaccustomed to telecom style cables they didn't see the problem with doing this.
That said, our old Train Describer cabinet suffered similar problems, although that might have been an age thing. We figured out it tended to fail most Tuesday evenings. Tuesday mornings being when the manager would generally come over, have a look round, and tidy the place up by shutting the TD cabinet door.
One other benefit of axle counters is that often once they have failed they will accept a reset from the signaller and this gets things moving again reasonably quickly. Once the techs arrive they can carry out a download and interpret the error messages to determine what happened and fix the underlying cause. With track circuits they will remain failed unless someone attends and rectifies it, or the fault self-rectifies. A self-rectifying track circuit failure can be an absolute nightmare to find at times.
As they are all countries with extensive low-frequency AC electrification, do track circuits have particular problems in that situation which led to extensive use of axle-counters as an alternative?
Yes they do as there is a smaller gap between 16.67 and either 0 or 50Hz and impedance bonds struggle much more with separating low frequency AC from DC.
Axle counters are very much becoming the primary method of rail vehicle detection down here in Melbourne. Just about every new level crossing removal and track upgrade sees them.
Yes they do as there is a smaller gap between 16.67 and either 0 or 50Hz and impedance bonds struggle much more with separating low frequency AC from DC.
With the trend towards solid state everything and IP networks all over the place, would the endgame be to put IP connections to each counter and do all the evaluation in the solid state interlocking?
Additionally, in the GSM-R era I suppose theoretically you could dispense with all the data cables and just use the radio to talk to the trackside equipment.....
As they are all countries with extensive low-frequency AC electrification, do track circuits have particular problems in that situation which led to extensive use of axle-counters as an alternative?
Yes they do as there is a smaller gap between 16.67 and either 0 or 50Hz and impedance bonds struggle much more with separating low frequency AC from DC.
Plus, at lower frequencies, the physical size (and cost) of impedance bonds increases.
Compare the size and weight of a traditional ‘mains adapter’/‘power adapter’/‘brick’/PSU that uses a conventional mains (50Hz) transformer to that of a ‘modern’ type that uses switch mode technology (like a USB ‘charger’…
One other benefit of axle counters is that often once they have failed they will accept a reset from the signaller and this gets things moving again reasonably quickly. Once the techs arrive they can carry out a download and interpret the error messages to determine what happened and fix the underlying cause. With track circuits they will remain failed unless someone attends and rectifies it, or the fault self-rectifies. A self-rectifying track circuit failure can be an absolute nightmare to find at times.
Alas, not all axle counter faults are like that. But yes, track circuits that fail and then ‘self rectify’, then the signaller goes back to normal running are a right pain. But then, a lot of the time this is a symptom of poor ballast conditions or water in the track formation. Although occasionally it can be due to a component in a joint less TC having degraded, and the system now having become temperature sensitive. Or more sensitive to the ballast conditions.
Alas, not all axle counter faults are like that. But yes, track circuits that fail and then ‘self rectify’, then the signaller goes back to normal running are a right pain. But then, a lot of the time this is a symptom of poor ballast conditions or water in the track formation. Although occasionally it can be due to a component in a joint less TC having degraded, and the system now having become temperature sensitive. Or more sensitive to the ballast conditions.
Ah yes, of course, having only ever worked on AZLM-K axle counters I forget the older varients are a bit less sophisticated.
I suppose to be fair to track circuits I should add that remote condition monitoring can go a little way to giving a history of the state/health of the track circuit, plus EBItrack/TI21 digital receivers can give some hints as to why it has failed of its a hardware problem. I'm trying to push for RCM to be fitted to all multi-section track circuits around here as knowing which section has failed cuts down faulting quite considerably, especially when they self-rectify.
Ahh, one of pet hate subjects: RCM equipment. But I will try to put a sock in it, as that’s going off topic, and if I start, it may be a bit of a long rant…
RailUK was launched on 6th June 2005 - so we've hit 20 years being the UK's most popular railway community! Read more and celebrate this milestone with us in this thread!
www.railforums.co.uk
