Return to Archive -by date - by topic - Archive 2002.
Sabotage can't be ruled out, but there's more to the failure of Points 2182A than missing nuts
A month on we know a lot more about the condition of Points 2182A at Potters Bar on 10 May, immediately before they failed under the Class 365 forming the WAGN 12.45 service to Kings Lynn. We also nknow that they had been inspected regularly and, according to the paperwork and debriefing of the workers responsible, nothing had been found amiss up to the most recent visual inspection by a patrolman 24 hours earlier.
And while several commuters had reported rough riding when travelling through the points at Potters Bar during the weeks before the accident, a high speed recording car trace of a run made on 30 April showed no ‘exceedences' through 2182A. However, on the evening of 9 May a homeward bound commuter was sufficiently concerned about the ride through Potters Bar that he reported it to the booking office at Stevenage .
Media comment favoured faulty maintenance as the likely cause of the accident, causing the share price of Jarvis, the maintenance contractor responsible for the East Coast Main Line, to slump. To stem the fall, a week after the accident Jarvis briefed the national press transport correspondents that the failure could have been the result of sabotage.
On 21 May Jarvis finally deigned to brief the railway and civil engineering press on their sabotage theory. The evidence came from colour photographs of the components of 2182A taken immediately after the accident and the next day. Jarvis would not give us copies, but the photographs have now been widely circulated to permanent way and signalling engineers so that they could check switches and crossings for similar conditions.
My diagram shows the detail of 2182A immediately before the accident. I must apologise to my purist chums in the signalling profession for referring to ‘tie bars' instead of ‘p-way' stretchers. This convention is to simplify discussion.
Jarvis' claims that sabotage should not be ruled out were based on the photographs, of the right hand end of tie bar 2.
Each tie bar is a square section steel bar, with the ends turned to a circular section and threaded. The threaded ends pass through holes in brackets attached to the switch rails. To maintain electrical isolation, where the tie bars pass through the brackets there is a rubber bush encapsulating a steel sleeve.
To secure the tie bars, nuts on the threaded section each side of the bracket are tightened up so that they clamp the rubber bush. To identify these nuts, I will refer the side of the bracket next to the switch rail as the outside and the side visible from the four foot as the inside.
These nuts are used to set the gauge between the switch rails. If you turn the nuts on the insides of the brackets away from the centre of the tie bar, the gauge increases. To do this you would also have to slacken the nuts on the outside of the bracket. Conversely, if you want to reduce the gauge you slacken the nuts on the inside of the bracket towards the centre line of the tie bar and tighten those on the outside until the correct gauge is achieved.
Clearly, these nuts are safety critical components, so each nut is paired with a thinner lock-nut.
To apply a lock nut correctly, you hold the main nut with one spanner and use a second spanner to tighten the lock nut up against the main nut. If you don't restrain the main nut, you can find that both turn when the lock nut makes contact and the locking action, which depends on the two nuts being tightened against each other, is lost.
My technique, when applying lock nut to a bolt or stud, is to apply an anti clockwise force to the spanner on the main nut, while tightening the lock nut clockwise.
Jarvis say that the nuts in question are tightened up by an 18 inch adjustable spanner. Now I know that points look like pretty crude engineering, but when you are setting them up, feeler gauges are used to set the distance between the tip of the switch blade and the stock rail to between a minimum of 1.5mm and a maximum of 3.5mm.
Contrast this relative precision with the forces from, say, a Class 91 wheel with a load of 10 tonnes passing over the points at 100mile/h. This suggests that tightness of nuts is quite important.
And yet the torque applied to the bolts is neither specified nor measured. So if big Oliver is the man on the end of the 18 inch adjustable spanner, the nuts will be tighter than if little Stanley had been on the job.
In my book, you don't use adjustable spanners for critical tightening work anyway. And if it matters that something is tightened properly you use a torque spanner so that the tightness is always the same.
nd of digression.
Photos of the right hand end of tie bar 2 indicate that the nuts on the inside of the bracket had been turned outwards, increasing the gauge. This means, of course, that the nut and lock nut on the outside of the bracket had also been slackened off.
What is the evidence for the inside nuts being moved? Well, the threaded ends of tie bars will probably have been greased on assembly. They then sit there subjected to the elements, atmospheric pollution, toilet effluent and ballast dust.
Over time the threads acquire a black patina. If you then unlock a pair of nuts and turn them, their threads will remove some of this accumulated surface layer, but it is unlikely that the pristine metal of the tie bar thread would be exposed, since nuts have a clearance on their thread.
However, the photographs reveal a shiny section of thread, about 25mm long, beside the inner nuts at the right hand end of tie bar 2. The contrast with the other threaded tie bar ends is most noticeable. Even on the photographs taken the day after the incident, when surface corrosion of the exposed steel had started, the brown rust and the black patina on the rest of the end of tie bar 2 were clearly delineated.
Jarvis suggest, and I agree, that this 25mm or so of exposed metal had been caused by someone winding the nuts outward without unlocking them. It would take a lot of force, because with the two sets of threads still locked together they would interfere with the thread on the tie bar end, deforming the metal and stripping off the patina.
Jarvis also claim that examination of enlargements of the photographs have enabled their experts to observe damage to the thread in this shiny section consistent with locked nuts being forced along. Perhaps, assuming the photographs are of forensic quality, but that is icing on the cake.
Based on the corrosion build-up visible on the ‘shiny threads' when photographs taken on the day and the day after the collision are compared, Jarvis' metallurgists estimate that the threads had been stripped to the bare metal not more than 24 hours before the accident. This suggests, say Jarvis that the nuts must have been moved after the visual inspection at 12.00 on 9 May.
Now this is pretty precise. But it has to be said that the protection of the points against the elements by the police and accident investigators was not of the best. There was some delay on the day of the accident before a blue sheet was put over the point mechanism and, as far as I know, nothing was done to protect the shiny threads from condensation under the tarpaulin during the first night.
Also, according to informed sources, metallurgists who have had access to the actual components at the HSE's Labs at Buxton think that 24 hours is a bit short. Jarvis' metallurgists estimate that after 10 days exposure the rust would be indistinguishable from the rest of the tie bar.
Jarvis' metallurgists also believe that the photographs of the fractured end of the lock stretcher provide a clue to the cause of the failure.
With nuts missing from the ends of both tie bars, the point motor could push the right hand switch rail against its stock rail to set the diverging route onto the down slow. But when the drive rod pulled the left hand switch rail back to set the through route, the only force on the right hand switch rail came through the lock stretcher.
My initial assumption was that taking all the forces, caused the lock stretcher to fail. However, if the moved nuts had pushed the right hand switch rail across by 25mm at tie bar 2 it would have halved the width of the flangeway – the gap between switch rail and stock rail.
In this case, there might not have been enough clearance for the wheel flanges and the back of the flanges on the right hand wheels of a train running on the down fast could have pushed the switch rail sideways in rapid succession. This would have applied a succession of forces to the already over-stressed lock stretcher, creating a classic fatigue situation.
According to the metallurgists the surface of the fracture in the lock stretcher is typical of just such low frequency fatigue. Examination of the switch rail and the wheel flanges on the three cars which did not derail should have proved or disproved this theory. Examination of the wheel flanges of all traction and rolling stock that had passed through 2182A in the previous 24 hours could identify the time when the nuts were moved.
So there you have the case for sabotage. Person or persons unknown gained access to the track some time after 12.00 on 9 May, removed two sets of nuts from the tie bar ends and forced one set of lock nuts outward, closing up the flangeway. We know it was after the inspection because an engineer on Potters Bar station say the track man, standing in the four foot, kicking the stretchers and tie bars on the points.
But do we know? Would kicking the tie bars reveal that nuts were off? Not necessarily.
First of all, the tie bars run in rubber bushes. Second, both tie bars were still securely bolted at the opposite ends to the missing nuts. Third, the wound out nuts meant that tie bar 2, at least, was in compression so that the moved nuts would be hard up against the rubber bush.
Now I haven't kicked a tie bar, but I do know that rubber bushes take away the feel in things mechanical. I am not sure that even a calibrated boot hitting the tie bars would have shown anything amiss. Nor would the missing nuts be immediately obvious even to a conscientious visual inspection.
Which brings us to the missing nuts. Back to tie bar 1 and the diagram.
You will note that at the right hand end the nuts on the inner side of the bracket are right up against the shoulder where the threaded end meets the main square section. But at the left hand end there is a considerable length of threaded rod between the end of square section and the nuts on the inner side of the bracket.
As a result, the remaining thread on tie bar 1 was greatly reduced. This length of thread had to be long enough go though the bush in the bracket and then have sufficient thread protruding on the other side to take a nut and lock nut with a few threads to spare.
In fact, I estimate that there would have been barely enough thread protruding to take the nut and certainly not enough for the lock nut too. And there is no sign that other nuts on this tie bar have been tampered with.
Now, normally, a tie bar would be installed symmetrically, with equal lengths of threaded rod at each end – and before the nuts were moved this was the case on tie bar 2. Moving the nuts had had shortened the length of thread protruding from the outside of the back. Once again, its looks as if there was not enough thread for a lock nut plus the main nut.
Next, note the escapement gaps in the drive rods. All mechanisms need play to work properly. Thus the drive rod from the point motor is not clamped tightly to the bracket on the switch rail.
Instead, two pairs of locked nuts are positioned on either side of the bracket to leave an ‘escapement' gap. This escapement also provides some slack for adjustment when setting up the points drive and accommodates wear.
To appreciate the function of the escapement gap assume the motor has pushed the switch blades into one position. When the points are reset, the motor starts up to pull the switchblades across, the drive rod moves, but nothing happens until the escapement gap is taken up.
There are two escapement gaps on points 2182A. One is at the connection between the drive rod from the point motor and the bracket on the left hand switch rail. The second is where the drive rod for the back drive is connected to the bracket on the right hand switch rail.
On the motor drive rod, the escapement looks in good condition. Obviously, you don't want the threaded section of rod between the nuts rubbing on the bracket, since it would cause wear and burr the threads. So there is a steel sleeve – a length of tube, over the threads.
But the escapement in the back-drive rod not only lacks a sleeve, but there is about 10-15mm of what could be shiny or burred thread. This could indicate that the escapement gap has been increased.
Next there is the back-drive itself which transfers the drive from the point machine, through tie bar 1 to tie bar 2. This is achieved through the drive rod already mentioned off the bracket for the right hand end of tie bar 1 1.
When the switch blade moves, this drive rod pushes or pulls one arm of a bell crank. A rod, parallel to the rails, is connected to the other arm and transmits the movement to a second, identical bell crank at tie bar 2 which has a drive rod to the bracket on the switch rail.
Once again there is scope for adjustment. The rod linking the cranks has a sleeve at each end which can slide along the bell crank arm. It is secured in place by two clamping screws.
If you slide the sleeve out towards the end of the arm, the movement of the switch rail is amplified and vice versa. On the bell crank at tie bar 1 the sleeve was almost at the limit of movement towards the fulcrum – that is the position which would minimise the movement transmitted to tie bar 2. The surface of the arm beside the sleeve is a lighter colour which could suggest that the sleeve had been moved inward to this effective position relatively recently.
So, two sets of tie bar nuts removed or vibrated loose, one set of locked nuts forced outwards, one tie bar as assymetric as it could be, possibly preventing the use of a lock nut at the end where the nut had come off, a question mark over the escapement in the back-drive drive rod with a sleeve missing and what could be an adjustment reducing the efficiency of the back drive.
Now, enter our saboteur. Motive uncertain, but let's say a disgruntled former employee out to discredit Jarvis. With luck a spanner on the outside lock nut at the right hand end of tie bar 2 will free it without the bar turning. The main nut then comes off fairly easily.
But freeing the lock nut on the inside of the bracket is more difficult. And time is at a premium. So with a bit of muscle the inside nuts are wound out by 25mm. Since sabotage is the aim, there is no point in replacing the nuts which are left on the ballast.
Just to make sure the nuts, or nut, at the left hand end of tie bar 1 are removed. If the saboteur overlooked the effect of the lock stretcher, this would have left the points failed the first time the diverging route was set. Nuts on the ballast again.
This fits the evidence. Had the nuts fallen off they would have fallen inside the brackets.
Was it a second attempt? The facing point check carried out on 1 May by a team of three technicians (experience: 10 years/3years/6months) found the outside nuts from the right hand end of tie bar 2 sitting on the ballast. The team replaced the nuts and checked the gauge, but failed to mention this unusual phenomenon in their report. They did record that a rubber bush was worn on tie bar 1.
Why was the replacement of the nuts not recorded in the report. Jarvis say it is because the report records what needs to be done, not what was done.
Yes, it sounds plausible, but it doesn't explain the asymmetry of the tie bar, the back drive escapement and the movement of the back drive sleeve – which could have been caused during the accident. An alternative theory is that the tie bars were removed when the down fast was being ground to remove rolling contact fatigue (RCF).
These points were identified as an RCF location and there was a spare half set on site in case replacement was necessary. The rail was ground and the Temporary speed Restriction removed last December. Jarvis tells me it has not been reground since and you don't dismantle the points anyway.
We are expecting the Hse to publish its interim report as this issue is published.