Brainiacs...How Do Axle Counters Work ?

[iands]

Hi,

 

So, real life track pins eh? Who'd have thought it.....

 

So, getting back to the axle counters then, these are totally different 'heads' to those used for HABD's etc? Again, I'm just curious as to how they count axles and know the difference between an axle and a part of the bogie/undergubbins.

 

Many Thanks,

 

Kindest Regards,

 

Shed.

Hi Shed,

 

The HABD and Axle Counter heads are totally different. In the simplest of terms, the HABD is effectively a "heat sensor" and an Axle Counter is effectively a "mass detector". Both very clever pieces of technology, but totally different aspects of the laws of physics.

 

Regards,Ian.

[Edwin_m]

Hi Shed,

 

The HABD and Axle Counter heads are totally different. In the simplest of terms, the HABD is effectively a "heat sensor" and an Axle Counter is effectively a "mass detector". Both very clever pieces of technology, but totally different aspects of the laws of physics.

 

Regards,Ian.

However a HABD is likely to have an axle counter included, though perhaps not as sophisticated as the ones used for signalling purposes.  This is needed so it can report back which axle has the fault.  A WILD can presumably count axles on its own without extra hardware. 

[Ncarter2]

But something of a pain to install?

 

Jim

Oh yes, not a job any of us enjoy.

[iands]

However a HABD is likely to have an axle counter included, though perhaps not as sophisticated as the ones used for signalling purposes.  This is needed so it can report back which axle has the fault.  A WILD can presumably count axles on its own without extra hardware. 

Yes, an HABD has the ability/functionality to count axles but not in the way an axle counter does, and the HABD has a "detector" on either side of the 4-foot so it can detect a hot axle on either the left or right, or indeed, both axles of a wheel set.

 

Regards, Ian.

[flyingsignalman]

A photo of axle counter heads at Heswall Hills Stn on the Down Wrexham line, taken 25/11/2013. A S&T man seeing the photo said these were an old type.

 

[flyingsignalman]

Some ancilliary equipment on the Up Wrexham line at Heswall Hills Stn 25/11/2013 for the axle counter there.

 

[Nearholmer]

In a sense, I don’t think that, so far, this thread has really answered how the detector head detects, excep5 calling it a ‘proximity switch’. Is it a Hall effect device? Or, a strain gauge? Or, what?

 

Kevin

Edited April 14, 2018 by Nearholmer

[phil-b259]

A photo of axle counter heads at Heswall Hills Stn on the Down Wrexham line, taken 25/11/2013. A S&T man seeing the photo said these were an old type.

 

100_5387.JPG

 

He is correct. They are of the 'AZLM' type which were commonly installed from the year 2000 and which required techs to learn 28 German words to fault find on it.

 

The prefered type these days however is the 'Fraucher' type which are not only easier to fault find and require less trackside equipment (no yellow 'Mushrooms'), but the sensors clip onto the rail and do not require holes drilled in precisely the right place to fit them (making things much easier for rail renewal). Oh and all the fault codes are in English too....

[phil-b259]

In a sense, I don’t think that, so far, this thread has really answered how the detector head detects, excep5 calling it a ‘proximity switch’. Is it a Hall effect device? Or, a strain gauge? Or, what?

 

Kevin

 

iands did that back in post 5 - where he said that the sensors generate an electromagnetic field that collapses in a very specific way when a train wheel (maintained to the specification laid down the the group standards for rolling stock) wheel (as opposed a p-way shovel) passes through it.

 

If you look at the AZLM type illustrated above you will notice what amounts to a set of radio transmitters on one side of the rail and a set of receivers mounted on the other. This creates two horseshoe shaped beams of electromagnetic force around the rail that get interrupted when the train wheel passes through them. The net result if you plot the received voltage against time is a  'bathtub' style curve as the wheel passes. Unless this curve (and the one produced by the adjacent sensor) do not match the parameters built into the system then a successful count will not be achieved.

 

The Fruscher type, while looking very different and only being fitted to the inside of the rail has components that effectively work in the same way. Again that 'bathtub' curve effect is what you get when a wheel passes through.

 

Maintenance / Setup of the AZLM type can be found in this technical manual written for Indian railways - who employ the AZLM axle counter exactly the same way as NR does in the UK https://www.slideshare.net/srguduru/handbook-on-maintenance-of-digital-axle-counter

[St. Simon]

He is correct. They are of the 'AZLM' type which were commonly installed from the year 2000 and which required techs to learn 28 German words to fault find on it.

 

The prefered type these days however is the 'Fraucher' type which are not only easier to fault find and require less trackside equipment (no yellow 'Mushrooms'), but the sensors clip onto the rail and do not require holes drilled in precisely the right place to fit them (making things much easier for rail renewal). Oh and all the fault codes are in English too....

Hi,

 

It's not quite simple as Frauscher kit is preferred nowadays, it depends on where they are being installed. If the area, or specifically maintenance area, already has AZLM kit, then generally any new axle counters will be AZLM, to decrease the number of different spares, as well as reducing the need to train staff on different systems (as it would a waste for staff to be fully trained just to maintain one head for instance).

 

Sometimes the same is true if an interlocking area already has one type, just to prevent creation of complex, and very expensive, conversion of interlocking data between the two types (as there are differences).

 

If an area was being converted from only track circuits to axle counters, Frauscher are generally being installed rather than AZLM. EDIT, they are also now being used as standard on overlay MSL systems.

 

One thing I learnt on a Modular signalling course about AZLM kit (although Modular specifys Frauscher Equipment, the trainer was an ex-Thales AZLM specialist), is that the computer code in AZLM Axle Counter Evaluator on the lines at Bletchley, or any site in the world, uses code that was developed from the Lorenz Cipher in WW2, the same code that was broken in the neighbouring Bletchley Park. Ironic really.

 

Simon

Edited April 14, 2018 by St. Simon

[St. Simon]

Hi,

 

There are videos from Network Rails Standardised Tasks of the AZLM kit (which are freely available on YouTube):

 

 

This is from Frauscher:

 

 

Simon

[phil-b259]

Hi,

 

It's not quite simple as Frauscher kit is preferred nowadays, it depends on where they are being installed. If the area, or specifically maintenance area, already has AZLM kit, then generally any new axle counters will be AZLM, to decrease the number of different spares, as well as reducing the need to train staff on different systems (as it would a waste for staff to be fully trained just to maintain one head for instance).

 

Sometimes the same is true if an interlocking area already has one type, just to prevent creation of complex, and very expensive, conversion of interlocking data between the two types (as there are differences).

 

 

However the same can be said of point machines and many other bits of equipment. While NR may well have a 'preferred' type, installation of other approved products is permitted for precisely the reasons you state - although that doesn't always work like that as Three Bridges ROC is having two totally separate types of Interlockings (SmartLoc from Alstom and Westloc from Siemens) plus each Companies own type of workstations installed during resignalling schemes.

[St. Simon]

However the same can be said of point machines and many other bits of equipment. While NR may well have a 'preferred' type, installation of other approved products is permitted for precisely the reasons you state - although that doesn't always work like that as Three Bridges ROC is having two totally separate types of Interlockings (SmartLoc from Alstom and Westloc from Siemens) plus each Companies own type of workstations installed during resignalling schemes.

Of course, there are also commercial and competition reasons for using different bits of equipment.

 

Simon

[Nearholmer]

I understand that the disruption of a magnetic field is involved, but,, by what means is the magnetic field detected? That is what I’m asking.

[iands]

I understand that the disruption of a magnetic field is involved, but,, by what means is the magnetic field detected? That is what I’m asking.

Read #34 from phil-b259. A collapsing magnetic field can only be detected in 3 ways: a change in frequency; a change in current; or a change in voltage. The flange of a passing wheel over the axle counter causes a distortion/collapse of the magnetic field, which is detected as a dip in voltage. This is what is "counted".

 

Regards, Ian.

[Edwin_m]

iands did that back in post 5 - where he said that the sensors generate an electromagnetic field that collapses in a very specific way when a train wheel (maintained to the specification laid down the the group standards for rolling stock) wheel (as opposed a p-way shovel) passes through it.

 

If you look at the AZLM type illustrated above you will notice what amounts to a set of radio transmitters on one side of the rail and a set of receivers mounted on the other. This creates two horseshoe shaped beams of electromagnetic force around the rail that get interrupted when the train wheel passes through them. The net result if you plot the received voltage against time is a  'bathtub' style curve as the wheel passes. Unless this curve (and the one produced by the adjacent sensor) do not match the parameters built into the system then a successful count will not be achieved.

 

The Fruscher type, while looking very different and only being fitted to the inside of the rail has components that effectively work in the same way. Again that 'bathtub' curve effect is what you get when a wheel passes through.

 

Maintenance / Setup of the AZLM type can be found in this technical manual written for Indian railways - who employ the AZLM axle counter exactly the same way as NR does in the UK https://www.slideshare.net/srguduru/handbook-on-maintenance-of-digital-axle-counter

Is the level of proving the same for an axle entering the section as for leaving?  I'm thinking there might be a safety argument to say that a borderline mesurement might be registered if it was leaving the section (which would result in a false "occupied" if it was not in fact an axle) rather than entering (which might result in a false "clear").  This discrimination would have to be done by the evaluators, since an axle leaving one section is in general registered by the same detector as entering the next one. 

 

the computer code in AZLM Axle Counter Evaluator on the lines at Bletchley, or any site in the world, uses code that was developed from the Lorenz Cipher in WW2, the same code that was broken in the neighbouring Bletchley Park. Ironic really. 

This is presumably to ensure the integrity of communication in case a cable is disconnected or intercepted.  Irrelevant fact of the week: for the same purposes SSI uses Manchester Coding, apparently invented by for the Manchester Mk1 computer which Alan Turing had a lot to do with. 

[phil-b259]

Is the level of proving the same for an axle entering the section as for leaving?  I'm thinking there might be a safety argument to say that a borderline mesurement might be registered if it was leaving the section (which would result in a false "occupied" if it was not in fact an axle) rather than entering (which might result in a false "clear").  This discrimination would have to be done by the evaluators, since an axle leaving one section is in general registered by the same detector as entering the next one. 

 

This is presumably to ensure the integrity of communication in case a cable is disconnected or intercepted.  Irrelevant fact of the week: for the same purposes SSI uses Manchester Coding, apparently invented by for the Manchester Mk1 computer which Alan Turing had a lot to do with. 

 

As I said earlier, a train wheel has a specifics shape and volume which distorts the electromagnetic field the same way regardless of direction of travel. Moreover BOTH sensors in the count head must see this for the Evaluator to register a a successful count - be that into or out of the section.

 

Any other metallic object (say a shovel) swung over the count heads , although disturbing the electromagnetic field, will not do so in accordance with the prescribed perimeters. As such the Evaluator will not register a successful count in any direction.

 

However as the Railway industry demands a 'fail safe' approach ANY metallic object passing over the count head that is not a train wheel WILL result in the section(s) (sometimes a count head may be shared between 2 track sections and thus 2 Evaluators) being set to occupied by the Evaluator and the system go into 'Fault' mode. Similarly if the count head stops communicating with the Evaluator then the section becomes Occupied and the system goes into 'Fault' mode If there is already a train in section when this occurs then the track section(s) will stay occupied when it leaves.

 

The ONLY way of restoring normal operation once the 'Fault' condition has been triggered is for the signaller to perform a 'Conditional Reset' - which as I explained earlier not only requires the Axle Counter system to be in full working order - but also demands that the last action (and I do mean last action - it must have taken place AFTER the techs have fixed the original problem) is a valid' Count out'. Even then an aspect restriction is applied and the first train through the section will have to be cautioned (this will sometimes be known as a 'Sweep' train).

Edited April 15, 2018 by phil-b259

[Nearholmer]

Iands,

 

Yes, and all I’m curious to know is by what device the magnetic field that is subject to change is detected.

 

Is it a search coil, or a Hall effect device, or etc etc?

 

It’s simple curiosity, that all. The concept that there is a stack of signal processing and logic ‘downstream’ of the transducer isn’t hard to grasp, I’m just nosey about the transducer.

 

Kevin

 

PS: and, wouldn’t phase shift be another way of detecting alteration of a field?

Edited April 15, 2018 by Nearholmer

[phil-b259]

Iands,

 

Yes, and all I’m curious to know is by what device the magnetic field that is subject to change is detected.

 

Is it a search coil, or a Hall effect device, or etc etc?

 

It’s simple curiosity, that all. The concept that there is a stack of signal processing and logic ‘downstream’ of the transducer isn’t hard to grasp, I’m just nosey about the transducer.

 

 

Ahh well if you wish to know the specifics of the sensor then you will have to ask the Manufacturers. https://www.frauscher.com/en/wheel_detection/

 

This page may be of interest however in providing some more info https://www.cst.com/solutions/article/frauscher-sensor-technology-designs-rail-wheel-sensors-with-cst-studio-suite

 

There is absolutely no need for the 1st, 2nd, 3rd line maintainers / designers / installers / faulting personnel to know exactly how the count heads work at the microscopic level - so it will not be shown in any of the documentation they have access to. All such people need to know is how the voltages they can measure will be affected by metalic objects - not why they are affected as such.

 

Obviously being a safety critical component, the sensor will have had to be certified by the relevant authorities - but given the firms behind the technology are German / Austrian and that is where the initial product trials / acceptance took place, you may struggle if you are not fluent in Technical German.

 

If you have the right personal connections then there is a meeting held every year on the subject of wheel sensors http://www.wheeldetectionforum.com/en/home/

Edited April 15, 2018 by phil-b259

[St. Simon]

The ONLY way of restoring normal operation once the 'Fault' condition has been triggered is for the signaller to perform a 'Conditional Reset' - which as I explained earlier not only requires the Axle Counter system to be in full working order - but also demands that the last action (and I do mean last action - it must have taken place AFTER the techs have fixed the original problem) is a valid' Count out'. Even then an aspect restriction is applied and the first train through the section will have to be cautioned (this will sometimes be known as a 'Sweep' train).

Hi Phil,

 

There are unconditional resets, although I don't know the specific differences, but they can only be done twice in 24 hours.

 

If an axle counter section is installed between a pair of signal boxes, one box will control the resetting of the section, whilst the other box will probably have a normally off slot on its section signal to allow the reset to happen.

 

Simon

[phil-b259]

Hi Phil,

 

There are unconditional resets, although I don't know the specific differences, but they can only be done twice in 24 hours.

 

If an axle counter section is installed between a pair of signal boxes, one box will control the resetting of the section, whilst the other box will probably have a normally off slot on its section signal to allow the reset to happen.

 

Simon

 

Down this way the ONLY way an unconditional reset can be carried out is if an EPR has been applied before the counter has been disturbed (i.e. we are basically saying a T3 where all signals must be on / replaced to danger anyway).

 

This is enforced by the Interlocking and cannot be overridden - and was largely a response to an event a decade or so ago where such an 'unconditional' reset in the Severn Tunnel almost resulted in a crash between two trains.

 

There is no way an 'unconditional' reset may be carried out under normal operation of the line concerned or when a failure of the system occurs - any reset during normal running HAS to be conditional, with all the safeguards I described earlier.

 

Things are slightly different where 'dual detection' (both axle counters and conventional track circuits are installed) is in place such as Balcombe tunnel - but even here there are some pretty restrictive conditions placed on when an unconditional reset  may be attempted.

 

As such I would be very surprised if 'unconditional' resets can be attempted in the manor you describe - even if they are in mechanically signed areas or are shared between two signal boxes. It such a situation does exist then it would be informative to have some examples.

Edited April 15, 2018 by phil-b259

[Nearholmer]

Thanks Phil ..... coils. I think that a single coil is used (probably the whole arrangement is duplicated), and that the presence of the wheel in the field of that coil changes the signature of the current in that coil by changing its effective impedance, possibly simply by ‘shunting’ some of the field.

 

A quick glimpse suggests that it doesn’t exactly say how the signal in the coils is ‘read’, which is the ‘stock in trade’, I guess.

 

Kevin

Edited April 15, 2018 by Nearholmer

[St. Simon]

Down this way the ONLY way an unconditional reset can be carried out is if an EPR has been applied before the counter has been disturbed (i.e. we are basically saying a T3 where all signals must be on / replaced to danger anyway).

 

This is enforced by the Interlocking and cannot be overridden - and was largely a response to an event a decade or so ago where such an 'unconditional' reset in the Severn Tunnel almost resulted in a crash between two trains.

 

There is no way an 'unconditional' reset may be carried out under normal operation of the line concerned or when a failure of the system occurs - any reset during normal running HAS to be conditional, with all the safeguards I described earlier.

 

Things are slightly different where 'dual detection' (both axle counters and conventional track circuits are installed) is in place such as Balcombe tunnel - but even here there are some pretty restrictive conditions placed on when an unconditional reset  may be attempted.

 

As such I would be very surprised if 'unconditional' resets can be attempted in the manor you describe - even if they are in mechanically signed areas or are shared between two signal boxes. It such a situation does exist then it would be informative to have some examples.

Hi Phil,

 

You could well be right, I've only seen it written in scheme plans and just formed my reply on my own understanding. I'll ask about it on Monday. After all, in my day to day work, we don't get involved in the nitty gritty of Axle Counter reset.

 

Simon