Chapter 13. In the Frame

In the last Chapter I left Marge, or at least the engine and main frame, teetering on a sack barrow but have made some tangible progress this last week. I realise it was way back in Chapter 9 that I dismantled the functioning BSA in as little as half a day. Painting has gone far better than I expected although of course now I’m getting comments along the lines that “If you’re so good at painting bikes…..” and you can guess the rest. Anyway, with a sense of excitement I have started the process of putting it all back together. I remain convinced that it’s better to do a dry build with the frame unpainted, then dismantle paint and reassemble. Even now I have managed to get a couple of chips, easily fixed, but at least I know everything fits together and I shouldn’t have to cut, file or worst of all weld anything.

I had marked all the engine mounting studs and trimmed them to length as I slowly started putting the frame back around the engine. Fortunately Jane was able to help me lift the bike down onto a more stable level before continuing.  I’ve now got as far as the frame fully assembled, gearbox and clutch in place and thankfully chains back in alignment.

As you can see the forks are re-assembled with the damper discs and look pretty close to Webb style girders. I’ll need to change the double lock nut on the lower front link for a hand adjuster but I can get the damping set-up fairly well as it is.

 

The new Dunlop K70 rear tyre and a few other parts arrived yesterday, but I’m having trouble with the seat cover. The seat frame that came with the bike measured up as a small Lycett, and this was half confirmed in that it has 14 springs, whereas a Terry saddle has 15. Having ordered what was described as the correct seat cover on eBay I’ve now discovered it won’t fit. I don’t know whether the cover is wrong or the seat frame is not quite what I thought. In the following photo you’ll see that the spring arrangement at the front is slightly different to most picture of a Lycett, but if anyone has any suggestions please let me know – I’m just about to return the cover via eBay. 

Frame designs

I don’t know if all pre-war bikes are like this but, the frame only becomes stable once the engine and gearbox plates are bolted in place. I have done my best to make all the fits as close as possible and followed the recommendations in Tuning for Speed with respect to the bearing area across the plates – i.e. at least 2/3 of the plate width. It helps that I have 5/16” thick plates. Even so it has made me think about the subject of frame stiffness and how much it affects handling and even if this matters for a pre-war bike.

Ian Dabney sent me a good article from January 1921, entitled MotorCycle Frame Design. In this the frame is thought of more like a girder bridge structure, the weight of the engine and rider, supported by the tubing between the two wheels – i.e. really only considering vertical loads. There is some mention of twisting (of the gearbox) due to chain loads, but little about the twisting of the frame along its fore and aft axis. The conclusion is essentially providing triangulation of the frame sections and reinforcing the top tube with an additional tube.

For us amateurs we are usually dealing with a bike that has already been designed and built, with the only variables being suspension settings and tyres. As we are mainly riding on the road we don’t have the no doubt painful option of pushing a bike to the limits and then that little bit further, as you might on a race track. Judging whether a bike handles well or not is therefore a matter of how it feels on the road. Likewise we probably have a better range of experience than many bike riders, but who can say they have tried and tested many frame designs? For my part the worst handling bike I’ve ever ridden was my Honda CB250, as mentioned earlier in the blog, but have had the pleasure of one of the best handling bikes, my Ducati 748. In a comprehensive road test I read, admittedly after I had bought the bike, it actually scored 100% for handling, even better than the 916. The reviewers put this down to having a slightly smaller width rear tyre than the larger bike, 180 instead of 190 mm. Any shortfall in cornering speed would therefore be down to the rider.

Between the Honda and the 748 most of my riding has been on my Commando, which is a strange one in many ways. The engine, gearbox and swing arm assembly is supported in the frame by what’s known as the Isolastic system, with large rubber bush mounting assemblies at the front and rear of the engine-gearbox plates and a pair of rubber “bobbins” at the head steady. The swing arm spindle is mounted in tube welded between the engine plates. This provides the benefit of reduced engine vibration being transmitted to the rider, but of course introduces flexibility between the rear swing arm and steering head. With the standard set-up sideways play in the Isolastic system is adjusted with shims at the front and rear bush assemblies and requires regular maintenance.

In the early to mid ‘70’s a Commando held the production bike lap record at Brands so it can’t have been a bad bike, but even then it had some quirks, particularly the potential for a tank-slapper at just over the ton, and a tendency to wallow (that’s the best way I can describe it) in a fast sweeping bend. One important fix for the former problem was to have a front TT100 tyre in good condition. It also apparently helped to go down from a 4.10 tyre width to 3.90. I also fitted a large diameter swing arm spindle by reaming out the engine plate tube. 

This sustained me until I got the Ducati, when comparisons highlighted shortcomings with the Norton. At that point I went the whole hog and fitted Vernier Isolastic adjustors instead of the shim system, which lets you achieve a tighter mounting and replaced the rubber bobbins with a Norvil head-steady. This is a small isolastic bush assembly which prevents any side-to-side movement at the top of the cylinder head, so in theory the engine-gearbox-swing arm can move up and down to absorb vibration but not sideways, or importantly twist.

The other way we get to understand motorcycle handling is from the experience of others, and in this respect some of the renowned frame designs come to mind. Of course, for us, perhaps the most famous is the Norton Featherbed frame, designed by the McCandless brothers. The other frame designer that shines through is Colin Seeley. His frames are significantly different to the Featherbed, but again he must have had the right philosophy.  The Vincent frame was well ahead of its time, not only with the large section backbone top tube, the engine as a stressed member, but also the cantilever swing arm. I consider the Egli frame to be a development of this, rather than a different concept. And last but by no means least, the Ducati trellis, of which I do have direct experience.

It was interesting to find that my BMW R80 frame looks very similar to a Featherbed, albeit with a single top tube. When I converted it to a Café racer I fabricated a rear-subframe, which I triangulated, and with the help of the technical guys at Hagon who advised on spring and damper rates, set it up to have the same geometry – i.e. trail, seat height, etc. as my Ducati. Its obviously a very different bike but even with its relatively narrow Metzeler tyres handles very well. As and when the Castle Coombe track day goes ahead its my most likely ride, as the Ducati is waiting new cam belts and rear chain.

 

Of course bike design has now moved through alloy twin-spar designs, to what are essentially monocoques carrying the steering head, mounted above the fully stressed engine-gearbox which carries the swing arm. I got to have a close up look at one of Rossi’s Yamahas at Brands one day which was impressive, but the packaging was so compact it was hard to see any detail, although one feature is that the top fork yoke is “relieved” in certain places to provide some flexibility. 

However I got lucky when I dropped off my rental Harley at Riders in Bridgewater, which was just after Martin Jessop’s 2^nd place in the 2017 Lightweight TT. Having shown a geeky like interest I was invited into their Superbike race workshop and got to talk with their race engineer who described their latest swing arm design. This featured torque tubes across the swing arm structure which could be adjusted to change the stiffness – apparently this would be worth one of two seconds according to the track they were on.

So where does this lead me? At this point I have to say these are only my views of what makes a good frame design, and I’m not going into tyres, forks or suspension settings.

• The correct steering geometry, i.e. trail
• A stiffness characteristic between the steering head and swing arm spindle that provides for sufficient rigidity but also flexibility to provide feedback to the rider. I’m pretty sure this is what sets a Featherbed apart from the rest. The “tin-tray” quality of my CB250 was the worst case. The Ducati at the other end of the spectrum.
• Swing arm support and swing arm stiffness – in an article by Dave Degens he said that the rear axle spindle can provide 50% of the swing arm stiffness – which explains why so many specials have features like eccentric adjusters.
• Weight distribution, with a bias towards the front end. Seeley frames provide for this.

The first time I took a pre-war bike around a serious bend was at Montlhery last year. First impressions were that the bikes I rode, Perry’s T80, the Rudge and the Rex Acme, were much better than I expected. I was relieved to see they all had good tyres. In fact, my only negative observation was that the lack of an efficient front brake. Rightly or wrongly I rarely if ever use the rear stopper, having had a couple of lock-ups, which seem more harder to anticipate than the front end going. At Mallory I tried to adapt to using the rear brake more and this helped a lot. I never got to ride any of the challenging Brooklands bikes, such as the Cotton, and will wait until I’m more experienced before “putting a hand up”.

Back to Marge, one feature I have incorporated are some special rear axle blocks, that are a tight sliding fit into the rear subframe “drop-outs”. These take up a slight difference between the drop-out width and the axle diameter and provide a bearing surface for the adjustor bolts. This should help to stiffen up the rear subframe.

 

I was able to get the optimum steering geometry by modifying the top yoke, where a very small difference in the offset has a proportionally larger effect on the trail. My concern, if there is one, is that any looseness in the engine plates will translate into flexibility between the front and rear axles. To this end I’m hoping that the 3/8 alloy plates and stud bolts in EN16T will prove effective. I can’t do much about weight distribution, but will be putting the seat fairly close to the rear of the tank. You may wonder why I even think about this, but I’m pretty sure that the engine power output of my little BSA is going to be fairly marginal and my only hope of keeping you guys in sight is to maintain corner speed.

I’ve just a got a text to say that my rear mudguard is available at the Argos click and collect, probably one of their more unusual deliveries, so I’ll have to dash.

Stay safe and well.

Gareth