Sunday, 20 September 2015

Tools of the Trade

(Reflections on a life in science: #4 instruments and gadgets)

Vacillation and procrastination are amongst the enemies of this series of posts, as well as the unexpectedly long time it takes to choreograph and follow the steps necessary to achieve a relatively ordered transition into retirement. My office is now reasonably clear, although there are several boxes of ‘stuff’ that still need to be transported home and squeezed into our painfully re-organised study; thankfully, this volume has been much reduced by a sustained programme of scanning – computer files are so much easier to handle and store. I also need to summon up the reserves needed in order to release the majority of my beloved collection of books into the hands of students, colleagues and charities: about five metres of them still await their fate. What I am bringing home falls into two broad camps: the useful (what I’ll continue to use in public talks on glass etc.) and the sentimental. There are however several items I’d love to have been able to keep but must simply photograph or scan and leave behind. Amongst these I include a few samples and measurement cells from truly key experiments – I don’t want the headaches associated with their safe disposal and so will gratefully rely on the services of our technical support team – and the more bizarre items like the 1926 Kelvin Double Bridge shown below.

Rescued before it ended up in a skip, this thing of beauty has served as a talking point, a coffee table and a footstool in my office for a great many years. It has ‘broken the ice’ with innumerable visitors, alongside the impressive high voltage rectifying valve from a WW2 radar unit and my home-made ‘sol-gel’ clock. (To give you an idea of dimensions: the Kelvin Double Bridge has an approximate box size of 64x46x25 cm – or 25x18x10 inches in its original context – and weighs a lot. It is going to a good home, but I’ll miss it. The vacuum valve is about 40 cm tall)

However, I mostly wanted to record a few of the ‘artefacts’ associated with my career that I have dragged out of various cupboards or off high shelves. Apart from the inevitable nostalgia, it’s been a sobering process in certain respects. The area of numerical calculation for example, which is intrinsic to the physical sciences, represents a particularly notable progression: my career grew out from the days before personal computers and electronic calculators. Similarly, I began my life as a scientist relying on hand-written documents and carefully drawn diagrams and graphs; even with a typewriter available the equations would remain hand written, and modern-style printers scanners and photocopiers were unheard of. Open access to telephones (landlines all) was reserved for the more senior people, and international calls were very rare indeed. The arrival of fax machines represented a huge step forward in timely communication, with e-mail and the internet following on behind; indeed, I still love the idea of e-mail’s immediacy – even though I have long-since tired of all the dross that gets transmitted. I might add a little context here by mentioning that I was the first regular user of e-mail in the Science Faculty of my university, that I defined the first framework for our presence on the web (and later chaired a working party to do much the same for the university as a whole), and that I had one of the first ‘personal’ desktop word processors in the Faculty; much of this practice I brought with me when I moved from the Rutherford-Appleton Laboratory to take up my new academic position. It’s easy to see why, on reflection, I have been a little taken aback as I review the scale of developments during one modest scientific career. In essence, this post comprises a collection of images with a bit of surrounding prose; it is in that sense even more self-indulgent than earlier posts in this series.
These are some of the artefacts from the more ancient levels of my career. Compasses, protractors and set-squares were ‘universal’, and are still used now; ‘Rotring’ pens and stencils, flexi-curves and Letraset transfer sheets  were arguably a little more specialist in their nature. My detailed calculations were initially performed using four-figure Log Tables, later up-graded to five-figure when more precision was required; the double-sided slide rule enabled fast ‘reasonable’ estimations to be performed. (By the way, slide rule operations were themselves based on the use of logarithms: for example, adding/subtracting lengths to perform multiplication/division.) The first truly ‘consumer-level’ hand-held calculator I saw, in the hands of a fellow undergraduate with rather more cash at his disposal than I had, was the Sinclair Scientific. I didn’t get a calculator myself until I was part way through my PhD; the one shown above is the second one I purchased – it handled all but the most protracted calculations required for my thesis (The Electronic Transport Properties of some Liquid Metals and Alloys, finished in 1978 and awarded the following year).

I’ve written before (here) of my visit to the working re-build of the very first electronic computer: Colossus, at the National Museum of Computing, next door to Bletchley Park. I’m buoyed by the fact that my own first experience of ‘high-level’ computing was on a machine – the PDP11 should you wish to know – which by modern standards wasn’t fast, but at least that didn’t run on valves. The programming language required, which we learnt in one of the post-exams courses we were obliged to take as first- and second-year students, was the rather inelegant Basic. Even so, as a course exercise, I managed to write a working programme to balance instrument packages aboard an imagined multi-level spinning orbital satellite. Later calculations, necessary for some of the final parts of my PhD research, were based on the much more powerful (although now passé) Fortran-IV language; these had to be run in something called ‘batch mode’, meaning, in essence, ‘overnight’. Current desktop computers and laptops are far, far faster and easier to use.
These are all relics of a bygone age of numerical computing. Although lighter and in many ways more convenient, the paper tape was nevertheless fragile: one tear could wreck several metres of program and/or data. Punched cards were certainly more robust, but a failure to stick to colour-codes for the various parts of a batched computer run (delineating program, data, end-of-section etc.) or to number each card within a given section would be disastrous if the deck of cards, which might be over a thousand deep, were to be dropped! For my early neutron scattering experiments I bought a large rucksack simply to carry the data on the train back to my university. Magnetic tape represented a huge step forward (this one, about 1 km long and in a roll of diameter approximately 27 cm, holds datasets from the 1980s at 3250 ‘bits per inch’ and is essentially unusable now – it’ll be destroyed). The output from each batch run could be on tape/card, but it wasn’t at all uncommon to get results printed onto fan-fold paper; print-outs of programs etc. would be done in the same way.

Let’s move on from numerical calculation. One of the many documents I have flicked through as I decided whether to recycle their paper, scan them or keep the original was my final year undergraduate research project report. The report was the first thing I had authored which made it to type: a friendly typist agreed to do the text part, and I inserted the equations, graphs/figures/diagrams by hand using the tools illustrated above. In order to have a copy of this for myself the typist had to use carbon paper between the top sheet of paper and a second one beneath* – and I had to re-draw everything else! The production of my PhD thesis was a little more advanced. After I had written it all out by hand another typist, in this case a friend who was under-employed and bored in her role in another department, typed up the text and my recently-wedded wife and I spent four long days inserting the mathematical content in the spaces left for the purpose. For this, we had the loan of our friend’s ‘golf ball’ typewriter which I augmented with a ‘ball’ of maths/science-related symbols I purchased. However, all the diagrams and graphs still had to be drawn by hand. I’m not an expert draughtsman, far from it, so I ‘cheated’ a little: everything was done on A3 paper, using my Rotring pens, letraset sheets, flexicurve etc., and then photographically reduced to A4 – that way, all my small mistakes were rendered ‘invisible’. (In passing, one might also reflect on the fact that there were photographic darkrooms aplenty back then and one could make a living doing this sort of thing. The story of photography, however, is well outside the scope of this post.) It took me almost six months to generate the top copy of my PhD thesis. Thankfully, photocopier machines had begun to appear by then and the university had a couple of them in its main library; I gratefully paid to have the requisite copies generated – indeed, I’d have gone without food in order to cover the cost if that’s what it would have taken. Research students have in more recent times asked about the curious formatting rules & regulations they had to follow in order to submit their own theses: many of these were anachronisms associated with the laborious nature of generating a thesis ‘in the old days’. For instance, although self-limiting in some senses, strict page/word limits were imposed in order to render the task tractable#; double-spacing to the lines of text allowed one to incorporate by hand any small corrections or revisions required by the examiners (imagine having the retype an entire section in order to correct a few words!); wide margins enabled the binder to guillotine the pages more than once if new pages needed to be inserted. Word processing, scanning, drawing packages, printers and so on have rendered all this work so much easier. Bliss.

Three pages from my PhD thesis are shown above in order to illustrate the laborious nature of its production – even setting aside the preceding 30 months or so of experimentation, reading, calculation and so on. The diagram on the left is of a bespoke sample measurement cell I made (in fact, I made several of them … more below); drawn on A3 paper with pens and stencils and then reduced to A4 to yield this final result. In like manner, the graph next to it was put together; note that each and every data point and error bar was hand-applied. On the right is a page of equations typed, symbol by symbol, under my guidance by my wife into the space left for it by my volunteer typist.

Now to the equipment – all those things required in order to get the results I needed. Developments have been no less remarkable in this sphere. If you’ve read the opening post in this pre-retirement series (here) you’ll have realised that I gravitate to the experimental side of the sciences; the more hands-on the better. It’s a mild curiosity then that at the heart of my undergraduate final year project was a rather impressive, and presumably expensive, ‘off-the-shelf’ spectrometer (see below). However, my PhD years put all that to rights. Not only did I piece together a project from ‘clues’ left by an absent supervisor, who referred me to the log book of an academic visitor who had by then returned to the his home in the USA, but, almost literally, I had to piece together much of the equipment necessary to make it happen. The sad truth is that there was next-to-no funding for this work, so if I didn’t do it myself it wouldn’t get done. To cut a very long story short, which is lodged forever in my memory, and formative within my evolving research ethos, I built my own furnaces in order to make my chosen molten metal alloys. I designed and built others then to hold these samples at precise temperatures in vacuum. I also had to design and fabricate the calibration/measurement cells to fit inside the furnaces such that I could accurately probe a given sample for its electronic properties (resistivity and thermopower). There was plenty of scope for hands-on experimental science. I have no recollection of any formal risk assessments being carried out at the time; in practice we were all aware of the potential hazards and took reasonable steps to mitigate the possible consequences. Having said that, I recall finding out what happens when one (accidentally) shorts the 12V car battery being used to drive a motor on a screw-jack: did you know that one could spot-weld things together quite effectively that way? I think I'd better stop at this foolish accident and say no more ...

I could go on for quite a while about instruments, equipment design and development and so on. Indeed, there is probably enough material for several posts on the ups and the downs of equipment and instrumentation developments alone. However, I must discipline myself to focus on the goal originally set: a themed skip through the decades, picking out a number of examples along the way. I hope to write a post at some stage on the more major projects I’ve been involved with, but that’s not my intent now. Herewith a few more pictures, and the necessary explanatory words …

Shown above is the EPR spectrometer (electron paramagnetic resonance at the heart of my undergraduate final year project. I’ve mentioned this in an earlier post (here) in the present series; it was the first ‘off-the-shelf’, ‘big-league’ item of equipment I ever used but despite its impact on me at the time you’ll spot that the controls are mostly analogue rather than digital and the output of results comprised a plot from a chart recorder. Note also the hand-written labels to the report’s photograph, which itself was of necessity limited to monochrome.

This painting by my wife illustrates the contrast rather well as I entered the world of an almost completely unfunded PhD project: the inset caption is from my thesis and blandly mentions the need to calibrate each of the measurement cells I made. The calibration setup I put together relied on a modified marmite jar (which had more robust metal lids at the time), rubber sheeting, cling film, masking tape and a contact adhesive. I need say no more.

These are the principal furnaces used: the vertical one on the left, which stood about 2m tall and accessed temperatures in the region 1600°C, and the smaller more ‘flexible’ unit on the right. The latter was made by me from scratch – from winding and cementing the heating wire onto the furnace tube through to the vacuum system. Notice the very early ‘computer’ on the desktop for data collection and basic analysis (details here); this contrasts nicely with the fact that I raised and lowered the furnace using a old car jack. Look carefully and you may also notice, on the shelf to the upper left of this image, the dark glass of the marmite calibration unit mentioned earlier and the plastic bottle of liquid mercury used as a ‘model’ for my high temperature molten alloys.
These are some of the measurement cells I fabricated in order to study the electronic properties of molten metals. For those alloys melting below about 1250°C I used pure silica glass. This meant learning enough glass blowing to fuse a length of wide-bore tubing to a capillary tube (the walls of which I drilled through in places, using an ultrasonic drill and a hand-made drill bit, and ‘plugged’ using hand-finished graphite pegs to allow electrical connection through to the liquid metal). The assembly shown here had a total length of about 40 cm. For the higher temperature work, which went up to almost 1700°C, I had to use multi-bore alumina tubing (the ceramic version of sapphire) and use a diamond wheel to grind the end to provide the right sort of shape to get the resistivity measurements when dipped into the liquid metal. The bottom right-hand image shows the ground tip of the alumina tube next to the end of one of my fingers in order to provide a bit of scale. There’s a lot more to it, as you might expect, but you’d need to read extracts from my PhD thesis for that …

Earlier posts in this series:
1) The Girt Pike – beginnings and transitions.
2) Do Labels Last a Lifetime? – people and other influences.
3) Nomadic Research: random walk or purposeful journey? – a timeline in research

* hence the term ‘carbon copy’: the ‘CC’ used in so many e-mails. If a third, often secret (i.e. ‘blind’) copy were to be made – transmuted to the ‘BCC’ line in e-mail headers – then another sheet of carbon paper and third sheet of plain paper were required; the quality of this additional copy was generally very poor. 

# It is also important to ensure that students are able to communicate their science in an effective manner, and that never requires ‘padding’ the text. Furthermore, one’s examiners are likely to baulk at having to read and assess an overly thesis: I know, I’ve examined a great many since those early days.