In 1984 Grubbs closed.  This was not due to lack of business, indeed, we were receiving enquiries up to the last week.  No, it was because Parsons was retrenching into its core business.  Their management did not see a viable future in optics (an interesting coincidence, back in the 1940’s they didn’t see a future for the fledgling gas turbine either!)  I was actually the last employee at Grubbs, I had my own keys—but it was an eerie experience working in a large, yet empty factory.

 

Fortunately I was able to transfer to the parent company (Now called Northern Engineering Industries or NEI as everyone called it.

OK I’m cheating here!  Although I toyed with the idea of getting a MG, I never actually owned one.  At this time I drove a Ford Cortina but the MG looks better!

A replica of Babbage’s difference engine.  Such vision all those years ago.

From 1980 to 1984, I was involved in the testing of the large optics manufactured at Grubbs.  This entailed a lot of effort as the following description makes clear.

 

Testing a large telescope mirror is a delicate and lengthy process.  The common method is to place the mirror on a hydraulic support system at the bottom of a tall tower.  At the top of the tower, at the focus of the mirror, an optical system illuminates the mirror with monochromatic light.  The illumination possesses a wavefront which conforms to the required surface shape of the mirror.  Light reflects from the mirror back to the optical system which contains a wavefront shearing interferometer.  This instrument allows light from one section of the mirror to interfere with that from an adjacent section.  The result is a set of fringes which, ideally, should be straight, even and horizontal.  Deviations from this state can be measured to show errors in the surface figure of the mirror.

 

Sounds relatively straightforward, doesn’t it?  Well there are complications.  Even though the mirror is supported by dozens of hydraulic supports, all interconnected so that the weight is evenly distributed, there are still distortions in the mirror shape, caused by the support system.  To eliminate these, the mirror is rotated 360 degrees in four stages with tests being made in each position.  Each test would consist of photographing the interferogram in the horizontal and vertical planes.  Of course, the photography wasn’t easy, we used glass plates and the exposures could be several minutes long!  Once the photographs were taken you had to descend twenty flights of stairs to the darkroom, develop the plates and check that the images were useful.  Only then could the mirror be turned on its support for the next test.  While this was taking place, prints were made and the fringe spacing accurately measured with the results being fed into a computer.  Even this stage was cumbersome, remember this was the 1980’s and personal computers were unheard of.  We used a mainframe computer which was located in Edinburgh!  To measure the prints, a grid was drawn on each print, the print was taped to the platen of a measuring machine.  Using a sharp needle connected to the machine, each intersection of a fringe and a gridline was pricked so that the machine could record the coordinates on a ticker tape.  The finished tape had to be fed into a device which sent the information to the mainframe computer in Edinburgh for amalysis.  This computer would work out the shape of the mirror surface, the errors and the shape of the polishing lap to correct them.

 

This process could take up to two days and had to be repeated after every figuring run.  Now you see why the construction of large telescope mirrors takes years.  And in that time I reckon I climbed enough flights of stairs to reach the summit of Everest!

 

Of course we had to move with the times.  The supply of glass plates was limited (although out technical director stocked up on them) and, eventually, we had to convert to film.  This was progress, we could use a 35mm SLR (a Pentax MX as I recall) and be rid of those infernal plates.  Now I’m not against plates per se, they are stable and last really well.  The problem was loading them into the plateholders.  This had to be done in complete darkness because the emulsion was very sensitive.  Now it’s virtually impossible to tell by feel which is the sensitive side of the plate.  It was not unusual to develop plates with no image because they had been loaded back to front.  This wasn’t amusing when you had spent perhaps an hour setting up and exposing them!

 

The last mirror I tested was the 4.2m for the William Herschel Telescope  This mirror I remember for the drama which occurred in 1982.  The mirror was almost finished after some two years work when the Foreman of the optical shop arrived in the office looking ashen.  “Where’s the boss?” he said, “I’ve got to tell him that the mirror is scratched!”  We all rushed down and to our horror, the beautifully polished surface was disfigured by several deep scratches across the whole surface.  An investigation came to the conclusion that some grit from the surface of the outside yard had somehow got into the pitch vat where pitch for the polishing laps was prepared.  When the pitch was poured and rolled before it was stuck to the lap the grit went unnoticed until the polishing began.  Although the technical director scowled and muttered about “sabotage” for weeks afterwards, nothing suggested anything but carelessness caused the accident.  This incident set completion back several months as the mirror had to be reground to remove the scratches before figuring could recommence.

 

I remember the first time I went “up the tower”  I was standing on the top floor talking to a colleague, I asked what happened if there was a fire lower down the tower.  “I’ll show you” he replied.  He led me to a trapdoor which opened on to the roof.  On the roof was a wooden box, he opened it to reveal a large drum wrapped with rope, the drum was anchored to the tower with a chain.  “You climb into the loop” he said, “Then you throw yourself off the tower!”  I thought he must be joking but he assured me the drum would lower me slowly to the ground.  “Has anyone tried it?” I asked.  “No, but the makers guarantee it” he replied.  I have to say that I was relieved that no fire ever occurred!

 

The story of Grubb Parsons would not be complete without mention of DS Brown the technical director.  David was, without doubt, an expert on optics, however, he managed to combine this with a gift for dreaming up absurd ideas on the practicalities of optic testing!  As an example, one of the problems we had was attaching metal supports to the glass surfaces.  The adhesive was a special silicone and the surfaces had to be thoroughly cleaned beforehand.  All sorts of chemicals were used but DSB insisted that the final cleaning be done with tap water!  Now it could be that he knew something that no one else did, but this was contrary to all known techniques, yet it worked; there were no adhesive failures that I was aware of.  Another of his eccentric ideas concerned a novel indexing table.  He came up with the idea of a ring of ball bearings in the upper and lower sections.  The theory was sound, each rotation moved the table exactly the same angle as the errors were averaged out (and ball bearings are exceedingly accurate anyway).  The problem he ignored was the incredible accuracy needed in machining the grooves which the bearings sat in.  He imagined that the ring of balls could be located accurately enough to compress each ball without exceeding the material’s elastic limit.  What happened in reality was that the balls indented the walls of the grooves in an unpredictable manner, a good idea……...in theory.

 

This illustrates what I have always seen as a weakness in some intellectual’s thinking.  They have mastered the theory but have no “feel” for the problem.  They assume that things in the material world behave exactly as predicted by theory.  Sadly this isn’t so as any real engineer will tell you!

 

Two events dominated 1984.  The first was the birth of my son, Jonathan.  The second was the closure of Grubbs.  Once the decision had been taken to close the works, staff levels reduced dramatically.  My colleagues found alternative work but I was persuaded to continue working until all the active projects had been finished.  Towards the end of this period I found myself alone in the factory.  Everyone else had left and I was given the key to the door!  This was an eerie experience, the factory was a large place and, sometimes a strange noise would alarm me, “probably just rats” I would say to myself and get on with the job.

 

Eventually the door was locked for the last time and I left to start my new job in the Mechanical Research dept. at Parsons.

 

I began my new career in the Mechanical Research Dept. of Northern Engineering Industries (as Parsons had become).  The department was situated on the second floor of the research building.  This change was quite a culture shock to me.  Instead of working alone in an empty building I was a member of a forty strong team in a huge office.  Even the journey to the second floor was novel as I had to use the paternosters.   There cannot be many employees who do not remember the Paternoster lifts.  The paternoster was an endless loop of open-fronted cabins moving slowly enough for all but the faint-hearted to hop on and off.   There were many chilling tales of what befell passengers who remained in the cabins at the top or bottom of its travel but, in fact, it was entirely safe.  There were three of these installed so there was never a wait for a lift in the morning rush.

 

The work of the Mechanical Research Dept. was both broad and varied in scope.  There were individual sections for Balance & Vibration, Steam flows, Condenser design as well as general turbine and generator design.  MR also worked in close collaboration with the other departments such as Applied Physics, Metallurgy, Electrical Research and Blading.  There is no space hear to describe the workings of a steam turbine generator set.  Interested readers might like to look here or here I’m afraid I have not been able to source a really good site covering the subject.

 

The geography of the department was a bewildering arrangement at first.  Often I would have to ask a colleague where someone was.  The reply, when decoded, would lead me to a small office situated in the heart of one of the many “bays” which made up the fifty acre site.  I think I spent the first month just finding out where things were.

 

 Coming from the optical works I was paired with an engineer named Roger Tankerville to develop a method of non-invasive flow measurement.  Checking the speed and direction of the steam flow through the blading of a turbine is essential to verify the efficiency of the design.  Historically, investigating the speed of the steam flowing through various parts of the turbine had been a difficult business.  Various devices had been tried with the hot-wire anemometer being the preferred tool.  The real problem was that the insertion of any device interfered with the flow and so could not be relied on.  In addition to this, the turbine designers were not keen to drill holes through the casing to allow the insertion of such probes.

 

Roger and I were developing flow measurement techniques using a Laser Doppler Anemometer.  This instrument used two intersecting laser beams to create a volume containing an interference pattern which acted like a series of planes through which particles, entrained in the flow, passed.  As each particle passed through an illuminated plane it caused a flash of light which could be detected.  The clever part of the process was an electronic integrator which timed the flashes and worked out the speed and direction of the particle.  Considering the high speeds and huge numbers of particles this was not an easy task!  After much effort we did manage to produce some results which caused great excitement in the blade design group.  At last they had some real data to work with.

 

Another project I became involved with was investigating blade vibration and twisting.  The turbine blades ranged from short, stubby types in the high pressure turbine to long slender versions in the low pressure turbine.  These LP blades were prone to twisting which destroyed their aerodynamic efficiency.  To counter this, they were stayed, near their tips by a metal wire.  The trick was to find out if this was successful.  The solution was to build a tank big enough to take a complete disc of these blades, spin the disk and take photographs of the blade tips as they passed.  This was no easy feat.  At 3000 RPM the blade tips were moving at supersonic speed.  The camera had to capture a sharp picture and the only way to do this was to use a flash of light of very short duration.  Of course, this short duration also meant the flash had to be extremely powerful.

 

Now this was the problem, there were no commercially available flashlights of sufficient power.  We had to make our own.  With the help of our friends from Electrical research a box the size of a large suitcase was constructed to provide the power and, after a lot of research a device was manufactured to provide the spark.  My problems started when I was assigned to help the engineer in charge.  Old George Blaylock was an engineer of the old school.  With the haircut and beard of a boffin he was tremendously knowledgeable but he was also quite timid.  George immediately put me in charge of the “sparks” department as he declared he too afraid to be within ten feet of the power unit.  Now I never did find out if he was pulling my leg, certainly I survived with no lasting injuries, but I still do not know if it was quite as dangerous as he made out!

 

Once this project was complete I was offered a transfer to the Applied Physics dept.  The AP dept. did the leading edge research for the group and were turning their attention to optical sensors, my previous experience with optics was proving to be useful.

 

 

 

 

The William Herschel Telescope

The 4.2m primary mirror under test at Grubb Parsons works.  The worker is Derek Airey a friend of mine.

David Scatcherd Brown (the technical director) and Fred ? (Test manager) looking at the mirror support system

The 4.2m mount nearing completion in Grubb’s erection bay

Next decade

A typical paternoster, ours did not have the grab handles!

A typical LP blade set.  Notice the wire stays on the largest disc.

Go to technical page