I got a couple heads from a 1989 R100GS for the salt racer. These are the ones to have since they have the so-called ‘D ports’ and have larger intake spigots since they take 40mm Bings. Got them cheap because one of them was busted up: 4 broken fins, a cracked 6mm hole that takes the rocker cover stud and a stripped spark plug thread. Other then that, they were OK!
Aside from the broken fins, you can see where I welded up the rocker cover hole ready to face and re-drill:
The process involves TIG welding a bead on top of the broken fin edge, then another bead on top, then another, then another….until a thin wall is built. Then into the mill to have the sides shaved to the correct thickness then another couple beads and repeat.
And after quite a few hours of welding, milling and filing:
TThey say you can’t be brave unless you’re scared. Well, at least I’m half way there! The thought of tearing across a granuley surface with ‘soft spots’ at 220+ km/h is giving pause for thought so I want to make the bike as stable as I can.
The up-side-down GSXR750 forks will help at the front end. They are considerably shorter than the stock forks which will lower the bike and reduce the front projected area and therfore drag, but the wheelbase will also be reduced. Straight line stability is improved with wheelbase length but there can be a trade off: a swing arm with wheel hanging on the end is effectively a ‘pendulum’. The longer the pendulum, the more sensitive the system is to tank slapping harmonics. The swing arm pivot points must be stiff enough to resist this tendency.
At first I looked at extending and stiffening the stock monoshock swing arm but after lots of measuring, doodling and thinking, I decided against it for the following reasons:
- Lots of work – for the later R80 monshock bikes, to extend the swing arm, you can’t just add an extension block between the stock arm and bevel drive because the wheel would move away from the scalloped section of the arm causing a clash. So you need to extend the tube within the scalloped part and would need to build a good welding jig first to get things right.
- More work – extending the stock shaft. The shaft is hardened carbon steel which, while not impossible, presents welding issues.
- Gear ratio problems. The longest bevel drive BMW made for the Mono’s is 33/11 or 3:1. Even with a 5% longer 5th gear, I’d most likely run out of gearing even at 8500RPM. I’ll cover gearing in greater depth in a later post.
- I found a couple stock BMW gadgets that solve all of the above: a R100RT swing arm coupled with a R100S bevel drive coupled with K100 wheels!
The R1100RT swing arm is a very stiff diecast alloy section and is 115mm longer than the stock mono item. It is a bit narrower at the pivot than stock but bolts straight on to the airhead frame with its 17mm pins – bless BMW. After considering the lost wheel base due to the new front end, the final wheelbase will be about 90mm longer than stock. The R1100S bevel drive has a very long 2.75 ;1 gear ratio which will be about right for the speeds I’m looking for on the salt.
However, while solving my biggest headaches, this arrangement has issues of it’s own, the cheif one being rear wheel centerline offset. The stock R1100 rear wheel is quite wide (160) and presents a huge offset while the earlier K100 17″ wheel is narrower (140) and bolts straight on – bless BMW again. But with swing arm pushed all the way over to the right, offset is still almost 11mm.
Of course, with the swing arm pushed over, the gearbox output shaft and drive shaft don’t line up so I will have to move the engine to the right as well! This shouldn’t present any balance issues since the gyroscopic effect of the wheels more than make up. The torque reation of the transverse flywheel has a far greater effect and that doesn’t effect things much.
I plan to take some meat off the K100 hub which will reduce the offset to around 5mm which is where it will stay. A longer pin and a spacer will have to be made to look after the large gap at the left hand pivot.
TThere’s a certain satisfaction in hacking tons of detritus and redundant bits off a frame – a bare canvas on which to plonk your own detritus and redundant bits!
Effectionally known as the ‘rubber cow’ the airhead frame is well known for it’s elastic nature. BMW, in their wisdom, took the lengendary Norton Featherbed concept but rather than continuing each loop over the top and under the tank to the head stem, the BMW loops converge at a single spine that starts at the rear end of the tank and continues to the head stem.
This is where a good deal of the elasticity lurks…. and waits to be exited by the various pendulum effects from the front end and rear swing arm sections amongst other things. I love a simplistic look but this can be hard to achieve. I could have gone the way that so many do and add tie rods that run from the head stem down across each side of the motor to the swing arm pivot area, but to my eye, this really stuffs up the simple lines of the bike.
I stood and stared and orbited around the frame for hours before settling on a stiffening plan and opted to horizontally stiffen the spine area and tie the top loops through to the head stem in a way that replicates the original Featherbed idea. Some gusseting of the head stem area was also added. Measuring and drawing in CAD for laser cutting makes things a whole lot easier.
The frame for the roller module gets a look-in now. The frame is made from 75x75x6 box section and will be eventually bolted to the floor. A bracket for the brake calliper was fabricated and the pillow block bearings are standard 60mm UC212 units. The deck is 5mm aluminium chequerplate.
The bearings are induction heated and slipped over each shaft end, then the assembly is lowered into the completed frame.
Now that the shaft has been shrunk into the roller, the assembly goes into a CNC lathe to take a lick off the O.D. and to machine cross grooves using the live tooling of the machine.
Not many dyno rollers have this feature (usually heavy knurl instead) and I’m exceedingly lucky to have access to these facilities.
The dyno roller starts off with a raw hunk of K1045 steel, 460mm in diameter and weighing 400kg. Some people turn down each end to form stub axles, but in my case, the shaft is quite long so this is not feasible. Instead, the shaft will be made separately and shrunk into the lump.
Next, the shaft was turned up. By design there’s almost 0.1mm (4 thou) interference between the shaft and the bore!! How the hell are we going to get this puppy in? With extreme heat and cold of course. As the ‘hot’ shaft set a bath of liquid nitrogen furiously boiling, I started heating the lump with a very large blow torch. After 20 minutes, the boiling stopped, signalling that both liquid and metal were at the same temperature: -196C or -385F.
But the lump still needed lots more heating before there would be enough clearance between the shrunken shaft and the expanded lump to safely insert the shaft. The time came to insert the shaft – a tense moment: if we hadn’t got it right, the shaft could sieze part we down creating a very big headache. After shaft and lump were united and allowed to equalize with ambient temperature, they were as good as one piece of metal.