Auld Reekie 3/03/19

Although I can ill afford it I am going to put my hand down the sofa, look under the car seats and raid the piggy bank to send a few quid to the DPS.

I might even rejoin it - I was a member many, many years ago but let my membership lapse.

 

What, I wonder, was the nature of that OHLE incident and was it the cause of the damage? If it was, who pays the bill?

 

No, the OHLE incident was beyond Dunbar. I should have kept RTT from that day, but when we pulled up at Drem at 1740 it was alongside a 91 set waiting in a queue to proceed, probably the 1620 from Edinburgh. Two separate incidents 20 miles apart.

 

I seem to recall that the incident was shown on travel disruption page of National Rail website as object caught in the OHLE.

Dave B

 

So we're entirely in the dark as to what caused the failure.

 

Bad news that a generator was also damaged, as the generators are unique for the Deltic power units. Unlike the traction motors that are similar/identical to those on the Class 37s. However, this 'incident' still sounds most likely an electrical flashover to me - although a big flashover involving all six traction motors and a generator. Brian Webb states in "The Deltic Locomotives of British Rail" p76, that "Main generator flashovers often occurred simultaneously with traction motor flashovers". The implication here is that a faulty generator could cause the traction motors to flash-over, or visa versa.

Although the weather further south was indeed very wet and windy, as we powered through East Lothian our local weather was still largely dry with light/moderate winds. I'm pretty sure that there was no lightning in the vicinity at the time. If there was lightning then people on the train would surely have seen and heard it. Did anybody see or hear lightning close to that time ? And as mentioned by others, the OHLE 'issues' were much further south.

So the OHLE, lightning and the weather all appear to be unlikely 'red herrings' to me. As an aside, are the DPS now saying that Deltics cannot operate in the rain, or under OHLE or indeed during lightning ?

If the DPS suspect that the OHLE arced onto the body of Deltic 9 then this would potentially be a serious railway incident, and this needs to be reported and investigated. However, the 'Farady cage' effect would mean that any such 25kV arcing would be routed around the outside of the locomotive body (assuming it is all conductive metals), and earth down to the track.

 

You're forgetting earth currents: things get odd with those. Assuming OHLE contact and/or lightning, which I'm inclined to discount.

I'm not familiar with the electrical arrangements of a Deltic but for a flash over to occur, either the generator lost regulation and its output voltage shot up, or somehow a massive back EMF of the motors occurred. This, to me anyway, sort of suggests something went open circuit at an inopportune moment allowing an arc to form.

 

The excellent book "Deltics at Work" by Alan Baker (with photos by Gavin Morrison) has a whole page (p25) covering the subject of Deltic traction motor and generator flashovers. Deltic flashovers were more frequent in the early days, and indeed a flashover occurred on the very first 100mph test run logged by Cecil J Allen in 1961 with D9001. From the book :- "What CJ did not realise was that at one point when speed was approaching 100mph the electrical machines flashed over. No one who was present, including the Running and Maintenance Engineer himself, the late Cyril Palmer - who was on the footplate - had experienced such electrical bangs before, and it is no exaggeration to say that everybody was pretty frightened. However, what later became normal procedure in such circumstances: close the controller, wait a few seconds and then re-apply power again, was followed and all seemed well!".

Interestingly Alan Baker also says that not all EE 538 traction motors are the same. Some of the later traction motors had better armatures and also better field windings that were much less prone to flashovers. British Rail would put all the best traction motor components together and use these traction motors for Deltics. This reduced the incidence of Deltic flashovers.

So did the refurbished bogies fitted to Deltic 9 have EE 538 traction motors with the 'later' style armatures and field windings. Also were these bogies geared for 100mph running (53:18 gear ratio) ? I seem to remember that some classes of 37 (heavyweights?) were latterly geared by BR for 80mph running, with presumably a different gear ratio.

Most of the Deltic books that cover flashovers state that they often occurred during the initial high-speed test runs with ex-works locomotives. Unfortunately the Auld Reekie was No.9's high-speed test run and a flashover occurred.

 

I was chatting to a friend on the phone today about Deltic 9's unfortunate failure, and trying to find reasons to explain why the flashover had apparently been so bad. We both ruled out rain and lightning, as the weather was clearly not too bad, as can be seen from the YouTube footage passing Prestonpans where there's no sign of rain. Prestonpans was just after the location where the flashover occurred. Also the OHLE problems were elsewhere that day, as has been discussed previously.

My friend reminded me of a modification that was written-up in the Deltic Preservation Society's newsletter around 2005 (Deltic Deadline 169 page 5 ). The modification was to put a 3 second delay onto the 2400 amp over-current cutout relays. These relays trip the individual traction motor contactors M1/2/3/4/5/6.

Any such modifications that reduced the fast-acting response time of the overload relays, could have a big effect on flashover protection. I believe that the overload relays provide a method of quickly breaking the critical traction motor/generator circuit, in the event of a DC electrical machine flashover. Introducing a 3 second delay on the overload relays removes an important part the Deltic's flashover protection and could lead to severe damage occurring.

Was this change ever implemented on Deltic 9 and could this 3-second delay have been a factor in causing Deltic 9's flashover to be so bad ?

My friend also told me that one of the changes done to the Deltics when ETH was being fitted, was the make the overload relays OL1 and OL2 open the traction motor contactors M1/2/3/4/5/6 directly, rather than using the PIR relay as an intermediary. This change was reportedly done to reduce the response time to flashovers, and open the traction motor contactors as quickly as possible. Every millisecond helps in the 'fight' against flashovers.

<edit - Checked the newer schematics. Earth fault and overload detection are now done independently using different relays, since the major ETH rewiring>.

 

Well here's my log of D9009 Alycidon from Edinburgh to Berwick, with the rescue loco 67004 as far as Newcastle thrown in as well just to show how tight the 90 min schedule was.

A nice smooth start out of Waverley, with speed up to 77 mph at Portobello Jn and 77 sustained on the 1:400 past Joppa, worth say 1,639 edhp. If the 1:400 extends as far as Musselburgh as per NR route diagram (I've not seen a five mile diagram) then the acceleration from 77 at Portobello Jn to 80 Mussleburgh would be worth 1,842 edhp.

My notes show an easing of the loco at the overbridge at 6m 53c, just after the former Inveresk station. It would appear we simply coasted to Drem, where the crew would be aware other services could get past. Note that on the 1:300 after Longniddry speed fell from 55 mph to 43 at Fountainhead bridge (former Aberlady Jn) but picked up a tad on the 1:300 down the other side.

As previously recounted, we ran past Drem station to stop on the up main to find a 91 on the 16.20 ex Edinburgh sat parallel in the loop. After the pull forward and set back and a few abortive re-starts we eventually got away nearly an hour after the initial halt, with apparently 2 traction motors isolated.

The big test in this direction is of course Cockburnspath bank, which starts with just under 2 miles of 1:210 after Innerwick, steepening to 1:96 over the Dunglass viaduct (36m 10c), continuing past the former Cockburnspath station (36m 46c) and on for another 3 1/2 miles or so to the A1 overbridge, south of the former Penmanshiel tunnel. I gather the gradients vary a bit on the deviation but taking 1:96 as an average for the 3.5 miles from Cockburnspath to the A1 bridge gives 1,728 edhp. Alycidon was allowed to run up to 83 mph downhill to Restonhill, but clearly all was not well and the loco was failed at Berwick, with what a relief driver was reporting as 4 out of the 6 traction motors 'gone'.

Hope the problems can be fixed soon.

 

Many thanks for publishing the epilogue for the day's events, with what could have been a record of a triumphant Deltic performance that instead was dogged with the misfortune that you describe. Just to reflect that this outcome now makes two "fast-start-followed-by-epic-failure" runs logged from either end of the ECML within twelve months - an unmatched pair, northbound and southbound, steam and diesel.

But just one footnote by way of an observation on the log that you have provided for the onward journey with 67004. At just under 52 minutes for the 67 miles, I can't find any examples of such a fast non-stop run from Berwick to Newcastle with conventional locomotive haulage - let alone on with a trailing load just exceeding 500 tons. The best times that the RPS archives are indicating are around 54 minutes (with Deltic haulage) on 9 coach loads of around 330 tons gross.

I doubt whether many would have cared one jot at that stage of the day. And I doubt that anything would have felt like much cause for celebration, given the prevailing circumstances. But if there was one morsel of a takeaway to be had from the return journey, then that would be it. So thanks once again for sticking with the task at hand and recording it for posterity.

 

When 55019 failed at Edinburgh on a charter, well over 10 years ago, again, a 67 took us back south, we bailed at Newcastle.

On that occasion she was limited by stock to 100mph, but put in a fine performance, think it was 67013, from memory.