How CAD CAM Engine Turning started in 1992

Pledge & Aldworth Engine Turners have been innovators in Engine Turning since 1978. We have built our reputation both on our willingness to accept technical challenges and then successfully surmount them, and also on our pioneering attitude towards Engine Turning as a tremendously versatile design medium for decorating metal products.

We continue to accept challenges as they are presented to us. In 2002 we were asked by the Crown Jewellers to engine turn an ice bucket for 24 champagne bottles, the largest we have ever seen. But the greatest challenge of all turned up in early 1992. The detailed story follows, quoted in italics.

" We were approached in February 1992 by an Italian Pen company that we had never heard of and was hardly known anywhere, and asked if we could engine turn a pen to make it look like the hand engraving on a sample that was sent to us. The engraving was in a stylised floral design.

I immediately realized that we could not do exactly what was being asked for, but I could also see that there was an alternative that the customer might accept.

I had an old straight line low relief engine turning machine dating from about 1925, and I used it to produce some samples of leaves and flowers, and sent it to Italy. A few days later I received a fax saying that one of the samples was accepted and that they wished to order 1,912 octagonal pen caps and bodies to be engine turned in silver. They also stated that it was their intention to form a long term business relationship with our company, and that these items were for the 80th anniversary of the founding of their company, Montegrappa, in 1912, to be marketed for Christmas 1992.

To do this with our old hand operated machine would have taken years! And the price would have been hundreds of pounds per pen. I needed a quick fix in order to secure a very valuable order!

I had seen a six headed machine that appeared to cut low relief or something like it, in one of my old Güdel brochures from the early 1920s, so I assumed that I could buy such a machine and told the customer that if they were prepared to place a binding order for the full edition, I would buy the machine to produce them at a reasonable price. They accepted and I began to call the various manufacturers in Germany and Switzerland.

It quickly became evident that no such machine existed, and I was in serious danger of losing the order. The only advantage I had was that the customer had no idea what size of company we had and was more than a thousand miles away, dealing with us exclusively by faxed messages, since they did not speak very good English at that time.

So I took a gamble and invested about £300,000 UKP of my own money in it. I told the customer that we would build a machine to produce the pens, and that we would have it ready later in the year to produce them. As it turned out, my lack of experience in engineering practice turned out to be to my considerable advantage. It was already late May.

The fact that the pens were to be octagonal meant that the machine would have to be quite complex, and I also wanted to make it "future resistant", versatile enough to do different types of work, since I was unlikely to be able to afford to build more than one machine.

Since CAD was the new way to design, I naturally assumed that using the latest technology would speed up the development process. At every point I was looking for ways to use technology to make things simple. I had already decided on a modular design so that I could add more options later.

The first issue was to design the machine and obtain the necessary electronic hardware to control it. I found a company that was willing to work with me to develop the electronic systems and dived in. It immediately became clear how the machine would be controlled by stepper motors and compressed air cylinders. We used rule of thumb methods to select the power of the motors and just over engineered everything. No risky maths unless absolutely necessary.

I am a pilot, and when you fly, you always use rule of thumb against your instruments and calculations, to ensure that you don't make a silly mistake. If your calculations come out similar to your rule of thumb estimates, then you won't go far wrong. This is important because mistakes tend to have fatal consequences. So I do engineering like I fly.

I purchased some CAD software and began designing the machine. Luckily for me, nobody had told me how people use CAD, so when I got it out of the box and installed it on my PC, I just looked at what it could do, and decided that the easiest way to work would be to begin designing directly in 3D. Since I already had a three dimensional idea in my head of how the machine would work, this seemed to me to be the obvious modus operandi. From 3D thoughts straight to 3D data structures that I could see on my screen and turn around and examine.

Some years later, when I tried to upgrade my old CAD system and investigated all the top industry standard proprietary CAD systems, I discovered that they all worked a completely different way from how I design. Even today, I have still not found a CAD system that lets me easily work the way I want to, the way I always will, that saves so much time.

By the time I started designing the machine it was mid June and I needed to work very fast. My ignorance of conventional teaching was my saviour and by December, with the help of an exceptional toolmaker who also thinks in 3D, I had assembled a complete machine and was testing it. Amazingly, the customer stayed on board in spite of the delays.

I still had another big hill to climb and had started work on this in about September. I had purchased my first computer three years previously in December 1989 and was already beginning to program my simple DOS based relational database to process statements. I found that writing simple batch files and programming the database was easy enough but until 1992 I had not even learned BASIC. I had to work fast so I bypassed "hello world". My first actual program was a 5,000 line compiler.

By Early December I had not only finished building the machine but also managed to create the program to cut a panel for one side of the cap, which we tested on the brass plate photographed below. Now I was using CAD in a wholly new way and was in completely uncharted territory. I still insisted on working from the start in 3D and developed a way of creating 3D surfaces in immense detail, with very direct control over every aspect of the surface.

The fact that CAD drawings are, in essence, nothing more than a mathematical table of three dimensional points in "virtual" space, makes zooming in to the finest detail simply a matter of ensuring that the data structure you use has enough room to store large enough numbers to get the required detail. Our CAD CAM machine has an accuracy of 0.0005mm, which represents one motor step. So I simply designed the system to make one step the smallest integer value and made the variables big enough to cope with about 100 times the largest vector value (x, y, z value) we might ever need.

Of course, with any system as complex as this there were bound to be teething problems. I nearly killed myself while we were assembling the machine in late November, when I ran a speed and direction test on the motor that moves the 100kg sliderest in the X axis.

The motor on that axis has an acceleration capability that is nothing short of "formula one" for a piece of kit not a lot bigger than a fat beer can! We had it wired up and I had programmed the keyboard to press F1 to start and F2 to stop. In case I had made a positive negative error (something I often do as I am numerically dyslexic) I had set the slide in the middle so it could go either way. My worst nightmare in my famously misspent youth was to be the subject of ridicule and insults while holding the chalk at the pub dartboard, as I can't do arithmetic to save my life! Trigonometry is easy though, provided I have a calculator. I set the acceleration and deceleration ramp factors and the maximum speed and hovered two fingers above the buttons just in case I had made a mistake.

I had. By a factor of 10; very easy for me to do. I pressed F1, and without even having time to turn my head and look, the sound of the motor winding up was enough to get my other finger down on the F2 button to stop it within little more than half a second. In that time 100kg of steel and aluminium alloy worth about £80,000 UKP had moved a whole foot (300mm) and hit the end of the slide! Luckily no serious damage!

I checked the figures again and quickly saw that the motor was capable of even greater acceleration; sufficient to have sent the 100kg mass off the end of the slide at about 45km per hour. In the direction it was headed, it would have exited the building through a brick wall and landed on my friend's Porsche parked next to my car in the yard a floor below.

While everyone was laughing at that, I realised that not only would the recoil of the half ton machine sliding back across the floor have probably crushed me against the desk and concrete wall, had it gone in the other direction, which was only by chance since until I actually ran the motor I hadn't been able to work out which way it would go because the wiring was so complicated, you would definitely not be reading this today, as I was sitting less than a foot from the end of the slide. The next day, a huge piece of iron gutter fell 50 feet from the roof and landed on the bonnet of the Porsche! Lucky the engine is in the back!

We had many electrical problems in the beginning with weird moments when the system would suddenly halt or run wildly out of control wrecking the workpieces. Eventually we discovered a ground loop in the cable shielding that was acting like a radio antenna and picking up all sorts of signals from around the building. Once we fixed that the electronics were reliable.

From Christmas 1992, it took about another 6 months to get the machine into reliable production and get most of the bugs out of the software. There were times when nothing would work and I remember one case when out of half a million lines of code, I just happened to notice, while talking to the toolmaker and still looking at the screen, that a dot had been tryped where there should have been a comma, in a function that was only called on rare occasions! It was a complete fluke that it was even showing on the screen!

I had tried to commission one of the best defence and radar software companies in the world, run by a friend of mine, to help with the software, but when they saw the complexity of the task and how little money I had they declined to attempt the compiler and control system. But out of friendship, they eventually agreed to write a couple of complex mathematical 3D intersection functions that I was having problems with and an optimiser module that seemed to have very complex parameters.

The intersection functions were a life saver as they enabled me to speed up the processing of CAD drawings into machine programs by a factor of about 100 times.

The optimiser proved more difficult. After joint work between Dave Knight and myself, working until 2am for about 3 weeks every day, we made a 3,000 line program that sort of worked. But every time we threw a new drawing at it, it would find a new exception and hang, requiring many new bug fixes and more late nights. After about a year or so more, we had got it to work reasonably well and it had shrunk to only about 300 lines of code. Then we hit a big problem.

I had been designing ever more ambitious creations and the drawing sizes had increased by a factor of about 50 or more. Suddenly, we started to hit hard coded problems in the optimiser that required a full rewrite, as the arrays and memory handling were insufficient to cope with the new larger drawings. I struggled to understand the code Dave Knight had written, but I couldn't see a way to fix it.

I returned to Dave, hoping for a fix, but he had just set up a new company and was choked up with defence and civilian aerospace contracts so couldn't help. Luckily, by now my programming skills had developed considerably and I had found a new way to solve complex programming problems. I told Dave about it and he called it "pseudo-code". I had used a large flow chart system to explain what the program had to do and how it would make decisions on various criteria. I worked for a week or two simplifying the flow chart until I was sure it covered every aspect and worked as simply as possible.

When I finally coded the flow chart into a new optimiser module, the actual bit that does the work was now down to only 10 lines and the whole module was less than 30 lines of code! And it works "in one pass", so fast you wouldn't believe it was doing more than just copy the data!

We have been running the system now with no further major modifications to the software since 1996 and none at all since 1997. The machine has not had a single mechanical breakdown in 13 years continuous daily use and the only faults have been electrical ones which have required replacement of drive control modules and other minor problems with the pneumatics, all components that were bought in and easily replaced. We once replaced the main computer with a new one after it suddenly failed in less than 10 hours from failure to restored production! My PC supplier gets paid extra to give us a priority service.

As we worked with Montegrappa, both they and we benefited hugely from our co-operation. Whilst we cannot of course claim all the credit by any means, it remains true that today Montegrappa is a well known international brand, selling large volumes of very highly priced pens in more than 60 countries throughout the World. Such a meteoric rise from nowhere was unlikely to go unnoticed.

In about 1996, I was told by the UK distributor, a friend of mine, that certain people were already asking how this company had been able to take such a large share of the top of the pen market in only four years from almost nothing. In 2000, the Aquila family sold the company to Richemont. "

Pledge & Aldworth continue to supply Richemont, both as Montegrappa, Dunhill, Cartier and other brands. We have widened our international customer base now to include other pen and luxury goods brands including Grand Masters®, Conway Stewart and others in America, Japan and Italy.

The Very First piece of CAD CAM Engine turning ever cut.

This historic piece of brass was the first test piece, cut when the machine was first assembled, in December 1992. It took us about six months more to resolve some very complex software problems and also some tricky electronic issues and get the system into full and reliable production.
 
The machine has probably more than 100km of wiring in it's circuits and contains many thousands of components. It was designed in 3D using the same CAD software that we use to create low relief surfaces.
 
This engine turned design is one panel from the Montegrappa 80th Anniversary limited edition pen of 1,912 pieces, representing the year of Montegrappa's establishment.
 
The design was taken from an original mechanical low relief pattern roller, redrawn visually (like a still life drawing) in 3D in the CAD system by David Pledge.

And now it is possible to create almost any image from a logo to a picture.
 
This sample tube was cut with our logo wrapped around it in a design with leaves and branches to demonstrate some of the capabilities of this unique process.