A wheel lathe cuts a wheel back to a true profile, so a wheel that goes through it always comes out smaller, and the diameter is measured before and after to the tenth of a millimetre. Unit 314's wheels come out larger.
I am a depot engineer and the wheel lathe is mine.
How a turn works, because there is only one direction it can go.
A wheel wears and it picks up flats and hollowing, and when the profile is out of limits the unit comes over the underfloor lathe. The lathe cuts the tread back to the design profile, and to do that it takes metal off. Always. There is no way to put metal on a wheel with a cutting tool.
Every wheel is measured before with a calibrated gauge and again after, and the two figures go into the wheelset record, because a wheel has a scrapping diameter and the record is how you know when you reach it.
A typical turn takes between two and six millimetres off the diameter.
Unit 314 has been over the lathe four times in three years.
Every time, the measured diameter after is larger than before, on all eight wheels, by between three tenths and half a millimetre.
I want the size of that to be clear. Half a millimetre is not a rounding error on a gauge that reads to a tenth, and it is on eight wheels in the same direction.
And the lathe cut. I watched it. Swarf came off those wheels and went in the skip like it does on any unit, and the profile after the cut was right.
What has been checked.
The gauges. Three of them, checked against a reference ring, and the measurements repeated by two people.
The lathe's own measuring system, which reads the wheel independently as part of the cut. It agrees with the hand gauges: the wheel was smaller when it went on the lathe than it was when it came off.
Temperature. A hot wheel is a bigger wheel and a wheel that has just been cut is warm. So we let it stand for six hours and measured again, and then overnight and measured again.
Larger. And the other units on the same day, treated the same way, are smaller, as they should be.
The before measurement. If the before figure were wrong, everything follows. So on the fourth turn, in 2024, the wheels were measured before by me, by the lathe operator, and by a colleague from the standards team, independently, and the three sets of figures agree to a tenth.
Then the lathe took about three millimetres off. Then the wheels measured half a millimetre larger than before.
Where the metal is, because that is a fair question and I have asked it.
The tread after a turn is fresh steel, bright, with the tool marks on it. On 314 the tread after a turn looks like that for about a week and then it does not.
It looks like a tread that has run for a few months, on a unit that has run for a week, and it measures the same as it did the day after the cut.
I have had a sample taken. That is a metallurgical test on a wheel in service, which takes a permit and a reason, and my reason was that the wheel material appeared inconsistent with the wheelset record.
The laboratory says it is the correct grade of wheel steel with a microstructure consistent with the manufacturer's specification, and that there is no evidence of any deposit or overlay.
So it is wheel steel, all the way through, and there is more of it than the lathe left.
What is done about it.
The wheelset record for 314 is annotated, and the diameters are recorded as measured, which means the record shows the wheels growing and the annotation says that the readings are verified.
The unit is in service. Its wheels are within every limit there is, its profile is right, and the only irregularity is that its scrapping date, worked out from the wear rate, is in about two hundred years.
What I do.
I measure it on the schedule, and I keep the figures.
The one change is that 314 is turned in the day shift now, not the night.
The lathe runs round the clock and the night shift is the quiet one and I used to put the difficult units through then, because you can take your time.
In March 2024 I was in the pit at about two in the morning with 314 over the lathe and the cut finished, taking the after measurements with the gauge on a wheel, and the wheel was warm from the cut the way they are.
A gauge on a wheel gives you a number and you write it down and move to the next. On the third wheel, with the gauge sitting on the tread, the number went up while I was reading it. Two tenths, in the time it takes to read a dial, with the wheel standing still and nothing touching it but me.