Artisan describes how to
produce precision pipes
‘Pipes and Pipe Work’ is a subject about which I have a ‘bee in my bonnet’. I have seen many models spoiled by untidy or ill-formed pipe work. It can, of course, be argued that so long as a pipe starts and finishes in the correct place and is of a suitable bore it fulfills its intended function whether it looks pretty or not. If the application is required to be purely functional this argument is credible. If our model is intended to reflect full size practice however, whether scale or freelance, it must be recognized that full size engine builders took a pride in the appearance of their creations. Whether the engine is a stationary engine, a locomotive, steam or internal combustion a full size example would have carefully formed pipes arranged in a tidy and orderly formation and properly supported. There is no reason why our models should not have the same attention given to this feature.
My personal objective is always for pipes to be formed into neat geometric formations with properly formed bends and straight runs that are, in fact, straight. All too often bends are of variable radius and straight runs are in fact ‘wobbly’.
It is very difficult to achieve such result by ‘freehand’ bending, particularly where tight bends are concerned and with large diameter pipes. Unless something is done to prevent it doing so the pipe will tend to collapse as it is bent due to the need to stretch the wall on the outside of the bend and compress the wall on the inside. The smaller the radius of the bend the worse the situation, with a kink forming instead of a bend if too tight a radius is attempted.
The classic way of preventing this is to fill the pipe before bending using sand or a low melting point alloy that can be easily removed when forming is complete. The alloy known as Woods Metal melts at a temperature of less than 100 deg. C and it can be removed from the pipe by immersion in boiling water. Another technique often used on larger pipes is to insert a pipe bending spring. For very small pipes the spring may be fitted outside the pipe. Personally I have never had much success with filling pipes and if a spring is used the need to control the shape of the bend calls for some mechanical aid if accurate formation is to be achieved. Furthermore the minimum radius of the bend is limited by the need to be able to remove the spring after forming. The technique works fine for plumbing the central heating system but leaves much to be desired for model plumbing requirements.
The technique I use is to employ a purpose made pipe bender as shown in the photographs. The bender employs formers for each size of pipe and for each bend radius. The formers consist of a disc with a semicircular groove in the edge of diameter equal to the pipe diameter. The pipe is pulled round the former using a plate with a similar groove formed along one face. The pipe is trapped between the former and plate such that at the point of contact with the former the pipe is constrained in a circular cavity and cannot distort into an oval shape. The result is that a bend radius of as small as two pipe diameters can be produced without any ovality developing. The picture shows a 3/16in. diameter pipe being pulled round a 3/8in. radius former, the radius being defined at the centre line of the pipe.
My own bender is based on a design published in Model Engineer magazine many years ago and is provided with formers for pipes up to ¼in. diameter. There are, of course, numerous proprietary pipe benders on the market but these do not generally have the versatility of the device I have described. It will be appreciated that I am talking generally about copper pipe. The same comments apply to steel pipe except that much greater force is required to pull the pipe round the former and the minimum bend radius attempted needs to be a little larger than might be used with copper, particularly for the larger diameter pipes.
I always try to manipulate copper pipe in the ‘as drawn’ condition. In this state the pipe usually starts off straight and free of any local irregularities. Bending is no problem with the material in this condition and the resulting formed pipe will be far more rigid than if it were annealed before bending. It will also be less likely to suffer damage or distortion due to the odd knock in service. Before leaving the subject of pipe manipulation the question of how to position the various bends in order to achieve the correct route arises.
My own technique is to make a template by manually manipulate a piece of soft iron wire of about 16 gauge into the shape required. If a bend is formed in the wrong position the wire is easily straightened out ready for another attempt. All that is required is to ensure that the bends are formed to the same radius as those proposed for the pipe. The wire template can then be used as a guide for forming the pipe. A little time spent getting the template right and in following it accurately is well worthwhile. If a bend is formed in the wrong place in the actual pipe it is almost impossible to rectify without producing unsightly kinks or irregularities
The joints in the full size pipes, either between individual pipes or between pipes and fittings, are made in a manner appropriate to the size of the pipe. i.e. small pipes up to about 1in. diameter will usually have screwed unions of some sort whilst larger pipes will have bolted flanged joints. Even the smaller pipes may be fitted with flange joints. All of the pipes on our models are of small size and lend themselves to the use of screwed unions. While this is a very convenient means of making joints it is not always visually correct.
A classic example is the pipe connection to the boiler feed clacks on the side of locomotive boilers. It is common practice for this connection to be made using a union nut of about 3/8in. A/F on models. I have been unable to find a single example of this form of connection in full size – four bolt flange connections are invariably used. (No doubt there will be an email to the editor quoting an example of a large union nut in this application!) If you are trying to be prototypical and fit a flange joint remember that flanges invariably have four or multiples of four bolts. It is very rare for six bolts to be used.
The majority of connections on our models will, however, be made using conventional unions. The female part of the union is normally formed with a conical seating having a 60 degree internal angle which is easily machined using an ordinary centre drill. The matching end for the pipe may take the form of a union nipple or an olive as shown in the sketch. Personally, I never use nipples for my pipe connections. They take up far too much space and require an unnecessarily large union nut. My method of making the olives is one I developed over sixty years ago and I am still using one of the cutters I made then. The method involves making some simple tooling as shown in the picture.
The tool is formed by drilling a deep countersink in the end of a piece of silver steel with a centre drill, followed by milling or filing away a quadrant from one side. The tool is then hardened and tempered in the usual way. A piece of brass bar stock of appropriate size is machined down to a size slightly smaller than the inside diameter of the thread in the union nut and a hole of pipe diameter drilled into the end. The form tool is then held in the tailstock chuck and applied to the end of the bar until a chamfer is formed to leave a knife-edge with the bore. The olive is then parted off and the form tool applied to the end of the bar for the next olive.
The process is repeated until the end of the prepared section of bar stock is reached, when the whole process can be repeated until the required quantity of olives has been produced. I usually make a batch of 20 or so at a time and keep a stock of various sizes. Someone will now write to the editor and say that the tool I have described will not work because the cutting edge has no clearance. This is, of course, theoretically correct, but try it – it works! I specified that the chamfer should be machined to knife-edge with the bore of the olive. This has the effect of swaging over the end slightly so that when it is applied to the pipe it locates precisely where it is required. If the pipe is trimmed so that the end sits comfortably in the conical seating of the female part of the union the olive will be in exactly the right place when it is silver soldered in place.
Don’t forget to put the nut on (the right way round) before soldering the olive!! Be very sparing with the silver solder. It must not be allowed to run onto the conical face of the olive and too big a fillet between olive and pipe will prevent the nut seating properly. Having smeared a little Easyflo flux on the joint I use a small flame to bring the olive and end of the pipe to the required temperature (a dull red heat) and just a touch with a piece of 0.8 mm diameter silver solder wire at the back of the olive does the job. NEVER try to get away with soft solder for this application. The process of silver soldering the olive to the pipe anneals the brass and it only requires a very modest tightening of the union nut to ensure a sound joint, the olive bedding into the female part of the union to form a good seal.
The end of the pipe should, of course, be cut square and thoroughly de-burred before fitting the olive. I always cut my pipes with a junior hacksaw and clean up with a fine file and a scraper. Roller type cutters are available for small pipes and produce a square end, but beware – they tend to produce large burrs in the inside of the pipe which restrict the bore. This restriction is proportionately greater the smaller the pipe.
I can honestly say that I have never had a leak from a joint made in this way.
Finally, if you would like to use a flange joint but do not want the hassle of making tiny gaskets (which will probably be a source of leakage) the arrangement shown in the sketch uses an olive as the sealing medium and the flanges to clamp the joint together. The bolting is still a fiddly job, however, and will be explained in a future article.
