Frameparts of classical race bicycles
A classic bicycle frame is a collection of parts: eleven steel tubes (FIG.1) bonded with brazingparts we call lugs (FIG.2).
Around 1900 the shape of the bikeframe is usually a diamond frame with angels of about 70 degrees. The process for making lugs in those days was forging. A white-hot piece of steel was hammered in a mould (often by steam driven sledge hammers) into a rough shape. This rough part was machined by lathes and milling machines (FIG.3) . It involved a lot of labour and it was an expensive way of working; it was used till around 1920. Lugs were made for both internal, (lug in tube) brazing, and external brazing ( tube in lug). The brazing often happened in a hearth ( like in a Blacksmith's workshop). The D-shape frontforkparts were common in England until the fifties
Around 1897, the firm Cresent started making lugs by pressing and welding steel sheets. Due to better foundry techniques in the early 1900's sand-casting is also becoming a fast growing production process for lugs. Both ways of production existed together for decades. Lightweight construction became important when bikes were used for recreation and sports. Until the sixties, bikes were mainly used for transporting people and material. Rugged construction and long lasting parts were preferred. For these utility bicycles the factory chose sand-cast brazingparts; as these were strong, the tube could be cutoff straight (see FIG.4). Building such frames became easier, but the frames were too heavy (for racing).
As the production process of steel tubes evolved in the 20th century, bikeframes could be constructed lighter. But the sandcast lugs are too stiff for thinner tubes and cracks could occur where the tubes entered the lugs. The pressed and welded lugs are thinner; shaping the lugs, made them lighter and distributed the stresses more evenly. Now these thinwalled tubes for sportsbikes, needed to be filed carefully, fitting the other tubes in the lug. Forkcrowns and brackets were often still sandcasted until the seventies. The Belgium based firm EKLA (FIG.5) has made sandcast lugs in racing quality till the end of the sixties. These were milled to exact fitting sizes with thin walls. But since the fifties the stamped and pressed lugs prevailed.
Examples of uncut stamped/ pressed lugs can be seen in FIG.10 (so called blanks), sold by Bozzi in Italy, but probably made at Agrati. This is the way lugs come out of the production process. In FIG.11/12 we see two different cut patterns. The same blanks were used for other types, as we can see the1950 catalog from Emilio Bozzi. This manager was the owner of Wolsit, Frejus en Legnano. These were already well known trademarks before the war. Wolsit disappeared in the 60’s, but Frejus and Legnano were top of the line bikes until the seventies.
The fitting of the pressed lugs was never accurate; it took the builder a lot of time making the right angles and fitting. In the seventies we see castings on the market made by the "lost-wax-model"-method. This proces is called "investmentcasting" in English; in Italy, where Cinelli promoted this production method, they talk about "microfusione". Parts made by this process are expensive, but the fitting and shaping are much better and that saves a lot of time during construction. Investment casting or the "lost-wax-model" has taken over the market of brazing parts during the eighties. Most factories, even Cinelli and Colnago outsourced the production to specialized firms like MFI (Micro-Fusione Italia). From 1990 we see nice investmentcasted lugs from Taiwan; leaflets by Longshen are pure bike-hardcore. Most of them are also available in stainless steel, a bit more expensive though. Rivendell, Henry James, Richard Sachs, and Pacenti are still making very fine lugs. Nostalgic too, because TIG-welding has pushed the brazed frame into a niche-market around 1995.
Thousands of factories produced brazingparts over the years, even if you only look at the racingbike scene. Most parts were not marqued, and the succesfull designs were often copied. So the information on this site limited.
The story of framebuilding, is partly about seizes and dimensions of tubing and threads. From the fifties till the eighties Reynolds and Columbus were market leaders; other tube manufacturers adapted to their standards. The French stuck to their own metric standards for decades. The framebuilders in China work for whole world; even in France we see English sizes (now ISO) in new bikes. The wrong standard won. English sizes are expressed in inches (1inch = 25,4mm). Engilsh threads are expressed in TPI (turns pro inch). In Italian standard threads we can find strange mixtures of metric and Imperial sizes. In the Netherlands we use English (= ISO) standards.
During the rise of the All Terrain Bike fashion in the eighties, TIG-welding became the most popular construction method and we soon saw TIG-welded racers on the streets. Tubingmanufacturers adapted their recipe to improve weldability. Form and seizes are less important for welding. Oversized and aerotubing would be used many times.
Lugs manufacturers produced oversized and aero-lugs too, but in fact they had lost most of the market. Famous manganese steels like Reynolds 531 and 753 disappeared because they needed to be brazed.
Of course we are talking about steel tubing. Postwar classical racers in Europe, can be divided in French, Italian and English bikes. Sizes of tubes and threads in the frames can be different. Dutch bikes will be built according Brittish Standards
In older Italian frames the sizes of the rearforks (20 or 21mm round) can differ from the current standard 22,2mm. Shortly after the war, the best known companies were Falck, Libellula, and Mannesman-Dalmine. In the fifties Cinelli has some of his tubes made by Columbus. He gets the rights for selling those tubes, for the bicycle-industry, worldwide. They become a major player in the market. After 1975 it’s hard to find Italian racers with Reynoldstubing.
The lugs and brazingparts are made Italian factories like Cinelli, Roto, and Silva. French companies making lugs are Nervex, Prugnat and BCM (=Bocama); Vitus (see FIG.1a) made steel tubing in France. All firms produced metric sizes and Imperial sizes. Haden and Davis made brazingparts in England, and of course Reynolds not only produced their Imperial sized tubes, but also metric sizes for the French market.
Because of different threads and sizes French frames are harder to built up. It’s difficult to find parts for classic French frames. If you have a project, search in French E-bay shops.
The development of the stickers from Columbus can be seen in FIG.21.The early dove is in a red background; the drawing shows 2 detailed wings. Sometimes the drawing is on a somewhat lighter background. The text under the drawing is "marca depositata". Around 1970 the dove is white and less detailed; the text is replaced by this sign: ® (registered trademark). After 1978 we see a fat single wing dove. Till mid-eighties tubes were made Chrome-Molybdenium steels; later we can also find Niobium and Vanadium additions, but these tubes were made for welding. In the eighties the classic round tube was challenged by new designs, like the Columbus Aero (see FIG.2e) and Max tubing. Columbus develloped special lugsets for brazing these tubes (see FIG 9c). A lot of framesets were also offered in oversize steel tubing, to compete with carbon and aluminium, and increase stiffness.
Stickers from Reynolds ( FIG.24) of course also changed over the years. We can see the decals used when all tubes were Reynolds 531 High Manganese steel. In 2008 (after 100 years!) the last series Reynolds 531 were made. Manganese steels are difficult to weld. They have been replaced by the Chrome-Molybdenium tubes Reynolds 525 and Reynolds 725, that can be welded and brazed.
Japanese tubes by Miyata were not used by the European bike industry. The Dutch firm Kowallik& Gaastra were importers of Miyata bikes in Europe (1974). The knowhow of Gaastra (Batavus) and the fact that the bikes were built to their specifications (Koga-Miyata), has upgraded the quality of factory; he did the same for Giant (a supplier of Schwinn and Specialized early eighties). The tubes of Ishiwata can be found all over the world, not just Japan, but also Taiwan, U.S., and Europe ( Flandria, Jan Janssen, etc.) The bike industry in the Far East, had a complete supply of tubes and brazingparts, that sometimes looked like the famous European designs. The best Japanese tubesets came from Tange and have been used by European builders, like Ernesto Colnago; this was their top frameset.
Brackets usually come in English, Italian and French sizes. Bracketwidth is 68mm for a French or English bracket, and 70mm for an Italian bracket. French and Italian brackets have a right-hand thread on both sides: Italian 36mm x 24 TPI and French 35mm x 1. The tube sizes for French brackets will be metric (28mm); English and Italian brackets will have Imperial sizes (28,6mm). Every now and then we find brackets with French sizes where the driveside is issued with a M35x1 left-hand threading; this is often called Swiss.
The cup on the drive-side (right) should be fixed; the bearings are adjusted on the left side of the bracket. English brackets always have a lefthanded thread on the driveside 1.370" x 24 TPI; this should be preffered if you have a choice. In the fifties and sixties there were English 3-speed bikes made by Raleigh, Humber etc.: with a 71 or 76mm bracketwidth and a 26 instead of a 24 TPI thread. It can be hard to find parts.
Cheap brackets are made of steel sheets, pressed into the desired shape and welded; they never fit very good. The cheap brackets of FIG.25 have such a welding seam; the bracket left is made for lugless frame construction like welding or bronze-brazing. This bracket has English threads and the builder should watch out to keep the lefthanded trhead on the right side. I have a M5 Blue Glide recumbent bike, where the welder forgot this; very disturbing for a mechanic! When mistakes like that slip through qualitycontrol this is bad. Sandcast brackets of the sixties (FIG.26) often have rough castingseams and surfaces. In those days there would still be an oilport; of course bullshit when the bracket is holed like the one on the right! Left in FIG.26 is probably an Elka; on the right is a De Rosa (circa 1970) according to Hilary Stone.
Around the early seventies the modern Investmentcasted brackets ("lost-wax-model") are appearing; in quality bikes like racers, they are used a lot. In FIG.27 we see two investment-casted brackets on the left that were used in the Gazelle Champion Mondial frames. Left-under is a Cinelli SCM with Gazelle logo en above that a Roto. The holes in this bracket were machined at Gazelle; the bracket on the right is obviously a Colnago.
At Gazelle the framenumbers of the Champion Mondial bikes were punched into the bracket some time before brazing. Sometimes these brackets stayed on the shelf for a longer period, and the rest of the parts maybe out of date with the productiondate that is coupled to the framenumber. The bracket of FIG.28 never became a frame; probably because there should have been seven digits instead of six!
The bracket of FIG.30 has no threading, it was made for pressed-in bearings, the diameter is 38mm and width is 79mm; the rearforks should have an oval shape. In the pictures we see the productionsteps 1,2&3: a steel plate is pressed in two steps into a cartridge shape. At pic 4 the closed side has been cut off, and four bulbes are pressed from the inside; at pic 5 the bulbes have been cut open, and are drawn from the inside to the outside. The RGF and BoCaMa bracket from FIG.29 are made the same way.
Since 1978 Columbus and Cinelli are within the same company; there is a new logo with a winged C. In those days they made 2 types of brackets: SCM and CCM; the difference is the orientation of the brazing of the rearforks ( see FIG.31 ).
The fashion of the routing for the derailleurcables changes in the end of the seventies; instead of going over the bracket they go under it. The CCM has a hole for mounting a plastic runner-part (low friction no paint damage). For a while even steel brazing parts were used; then they started to integrate the cablerouting in the investmentcasted brackets. Especially the Cinelli "spoilerbracket" is a fine piece of engineering ( see FIG.32/33 right). But it was expensive, heavy and not really practical. The cables would damadge the paint anyway. The plastic runners remained popular and replaced these over-the-top parts later. In FIG.33 we see that the brackets have reinforcements on the inside of the forks. The bridge piece near the bracket can now disappear.
The frontfork of FIG.34 (round forkparts) is a Cinelli flatcrown Pista 24mm. The fork is strengthened on the inside with loose tangs. Some forkcrowns have special pockets for this purpose. In FIG.35 on the right it is just an edge no more than 2mm; on the left the tang goes into a 14mm deep pocket. This design gives a gradual stiffening in the direction of the crown.
FIG.36 shows there were different types of forkblades. In FIG.3 we can see a D-shape, much used in England till around 1950. In FIG.37/38, we see two photo's of an other classic design: the double plated crown. These are three machined parts, that form a crown (where the fork-blades form a integral part of the structure). In FIG.39 we see a forkcrown type that was used in English bikes; Raleigh, Humber etc., made hundreds of thousands of them........
In FIG.40 we see a "flat" forkcrown at the bottom , a "semi-sloping" crown in the middle and a "sloping" crown at the top. In the trackfork of FIG.34 are round 24mm Columbus forkblades; in FIG.41 we see different trackforkcrowns for 22mm blades. At the top Cinelli, second Bocama, third Prugnat, fourth is a Georg Fischer (sandcast) and at the bottom a Davis Components. In FIG.43 some more old glory: top-down: Roto, Haden Europa, Vagner Topdiamond, and a Nervex-copie by BCM. Another fashion in forkcrowns was internal brazing; forks would look like one part (FIG.42). Cinelli and Davis made lots of them. We will look at the development of Cinelli crowns in the years 1950/ 1990.
FIG.44 Cinelli forkcrowns
The cyclingcareer of Cino Cinelli stops after the war and he starts designing bikes and bike-parts. His topmodel was called Super Corsa and used an integral sandscast forkcrown that was internally brazed. This was a heavy piece of steel; Cinelli designed new brazingparts, not only for his own bikes, but he also sold them to other manufacturers. Much of the foundery-work was outsourced, among others, to Georg Fischer. Cinelli thinks the standard Reynolds forkblades are not stiff enough sideways; he seeks an Italian supplier and finds Columbus. He builts his bikes with Columbus forkparts 27,5x20mm, and Reynolds frametubes. Cinelli starts coöperate with Columbus and they begin a complete line of cycletubing. In the early seventies Cinelli starts to desing investment casted parts (Microfusione). His semisloping forkcrown MR is made for Reynolds tubing and the MC for Columbus blades. The MC is somewhat wider and does not need tangs, like the deep pockets of the MR do. Reynolds starts to market a new forkblade in the seventies in a size almost identical to Columbus (Continental New Oval). Now they can use the new SC and CC crowns too. The microfusione parts fit very well and mean less work for building a frame. In the early eighties the aero fashion, of which Cinelli is one of the first, makes him redesign his SC and CC crowns; the SCA and CCA are lighter and more aerodynamic then their predecessors.
A crown for a French steeringtube will have a 25mm fit; do not use those for English crowns (these should have a 25,4mm fit and 25,4 x24TPI thread). Older Raleighs, and tandems with a 1 1/8" steeringtube, may have 26 instead of 24 TPI thread. It is hard to find parts for 1" headsets for these forks.
In older racers, lugs can be complicated; certainly in English racingbikes ornated lugs symbolised the craftsmanship of the builder. They used blanks, uncut lugs, and cut these by hand into ornated lugs! Fashion, it had nothing to do with construction. French manufacturers like Oscar Egg and Nervex, also made complicated lug desings in their factories. They went bankrupt around 1960, but Peugeot enough stacked to use Nervex lugs far into the seventies. We can find many types of cutouts in Nervex lugs. They all had numbers in catalogs, and could be made to order. Sometimes a design was altered. and that number got an extra code: bis, or altered again, new code: ter. The middle lug from FIG.46, has de code 86 bis; later this is replaced by the code 86 ter with two extra windows,. The cutouts on the downtube side also had a code; in this case 158. I am the proud ownwer of a Nervex 86 bis/ 158; alas I've no complete set.
In FIG.46 on the left we see a raw lug ( blank), the factory and the framebuilder can use their fantasy to shape lugs; on the right we see a minimalistic type (Prugnat S). The Bocama lugs from FIG.45 were extensively worked over, before going to the framebuilder. You can find them in Gazelle Champion Mondial frames.
In the seventies investment casted lugs apeared like the Cinelli's of FIG.52. These lugs have sharp edges and the overlap for brazing is small (see FIG.51). Pressed and welded steelsheet lugs can easily be reformed upto 3 or 4 graden degrees: for cast lugs this only a mere one degree. The craftsmanship of the framebuilder is for a large part fitting the lugs into the desired angels! It is surprizing how bad standard fittings can be; even the more expensive ones. Sometimes there is a lot of reaming and filing to do, before we get to the brazing in the right angels and measurements. In FIG.53 in the middle we see a Nervex Professionel lug from the early fifties; from 1956 the two little horns are replaced by a fishtail as we see on the left. The lug on the right is a crude Nervexcopy made by Bozzi.
In cheap bikes the drop-outs are no more then a sheet of steel of three millimeters thickness. The better drop-outs are made of forged or cast steel, often machined for excact tolerances. In FIG.54 we see five classic Campagnolo rear drop-outs. On the right we see three horizontal dropouts as used in many steel racers. Short, middle or long; the older the bike - the longer the dropouts usually are. Vertical dropouts should be brazed very accurately, because otherwise wheelalignment is impossible.
The trackdropout left-under is simple piece of steel. The framebuilder may file and cut half of it away. The romantic magic of framebuilding disappears as soon as you start; it looks too much like hard work. The thread of the campa derailleur is 10mm x 26 TPI; that is one of the strange mixes between Imperial and metric seizes; almost equivalent to the other standard: M10x1. The wheel alignment screw is M3x0,5 threaded. The eyelet for mudguards has a M5x0,8 thread.
In FIG.56 we see two classic French dropoutsets. Both sets have eyelets for carriers or mud-guards; if you don't want them, you just cut them off. The Vitus reardropouts (right) have an offset to the inside. This makes rearforks somewhat wider and it is easier to change wheels. The threading for the derailleur is M10x1; the wheel align screw for Vitus is M3x0,5; the Huret is not threaded, the hole is over 3mm; you need a special screw.
In FIG.56, at 1 we see brazingparts for cantileverbrakes; at 2 we see, top-down, three braze-ons for sadlepin bolts. Parts 3 and 4 are brakebridges; at 4 we see two fine cast Cinelli products; at 3 two simple tubes; the curved one is usually found in trackframes. Parts 5 are used to strengthen the brakebridges. Bracket-bridges come in different sizes, they can be fat, see 6, small 7 and butterfly-shaped 8 (Gazelle).
In FIG.57, at 1 we see three kinds of tangs for frontforks; at 2 two frontderailleur braze-ons; at 3 we see cable runners for over and under the bracket. At 4 we see some cable stops, at 5 left to right: a pumpboss, a boss for a ridernumberplate in competion (M5x0,8) and a chainhanger for changing the wheel. At 57-6 downtube controlbosses, cable runners 7, reinforcements for brakebridges or bracketbridges 8, and at 9 reinforcements for bottle braze-ons. To braze the seatstays to the frame, we can use steel spearpoints, like 10 or 11 (heavy). The construction with loose sheets (12) is lighter, but more work.
I guess that there will always be a small market for classic steel racers; especially in England and the U.S. there are still many small custum builders. Rivendell, Richard Sachs and Pacenti still make fine lugs in the USA. Real handcut lugs still can be found at English builders like Kevin Sayles, or American builders like Peter Weigle. There can be upto ten hours of work for some lugs in such a frame; this will be expressed in the price. If you want such a frame, you must be willing to pay for it. If you are looking for a bagain don't look look there.
Even in the Netherlands some small shops built classical frames, but we hardly see ornated lugs. It doesn't make a frame better, just more beautiful and also more expensive (this is un-Dutch ). The English firm Hetchins even went as far as brazing ornated steel strips under a lug; if these strips were longer then the butted part of the tube this could mean weekening the construction. The end of the strip will act as a stress-raiser. I like ornated lugs, but not like the over-the-top 1950's Hetchins; now I've retired, I will try building with bi-laminates.
MORE ON THE INTERNET
Richard answers a question: http://www.richardsachs.com/site/how-frames-are-made/
Lugs on Flickr: http://www.flickr.com/groups/1330795@N21/pool/page10/
Nice lugs, brackets etc. http://www.longshen.com.tw/index.asp ; long gone: Nervex ; Vagner
Internetshop (USA) tools,tubes and lugs: http://www.novacycles.com/catalog/
Many old brazingparts can be found here: http://www.velobase.com/ListFrameParts.aspx?Offset=0
A classic collection: http://www.classicrendezvous.com/main.htm
INFO REYNOLDS STICKERS : http://www.classicrendezvous.com/British_isles/Reynolds_gallery.htm
This English site sells complete sets : tubes + lugs + etc.: http://www.ceeway.com/Tubeandpartsbundle.htm and other parts and tools.
THE PROCUCTION OF LUGS: STAMPED AND PRESSED, INVESTMENT- OR SAND CASTED. A video on investmentcasting at LongShen: https://vimeo.com/62879060