Square taper bottom brackets

Common on better bikes from the 1970s to the 1990s, these provide a bigger crank to spindle interface and easier maintenance than the older cottered cranks.

There is some confusion and misinformation about the types of square taper you may find. There are generally two types: JIS (Japanese Industrial Standard) and ISO. ISO is far less common and was used mainly on European brands like Campagnolo and, at one point, TA. It survived on Brompton bikes until fairly recently, causing many headaches for owners of older Bromptons who can no longer buy a 119mm ISO bottom bracket.

Square taper has largely been eliminated on new bikes by splined interfaces, or those where the cranks are permanently pressed to the spindle (Campagnolo and the RH side of Shimano cranksets). There is also a proliferation of bottom bracket standards in different sized shells.

All square tapers have a 2 degree angle but ISO tapers start thinner than JIS tapers. To tell the difference, measure across the flats at the tip of the spindle. JIS is 12.7mm and ISO is smaller (one day I will get one and measure it but, as I said, they are unusual).

A JIS crank, if it doesn’t bottom out, will slide 4.5mm further onto an ISO spindle than would an ISO crank. Likewise, an ISO crank will slide 4.5mm less far onto a JIS spindle than would a JIS crank. This means that Brompton owners can use a 110mm JIS bottom bracket instead of a 119mm (119 – 4.5 -4.5 = 110) ISO bottom bracket provided there is sufficient engagement. I have tried it and it works perfectly well.

Dura-Ace square taper bottom brackets (up to 7400, after which DA changed to a splined connection) have slightly shorter tapers than most other JIS spindles. In fact, taper length does not seem to be totally standardised and I found different length tapers on Shimano bottom brackets of different eras, the longest being on a recent UN55 cartridge unit. All are still 12.7mm at the narrow end. Some people suggest that 7400 was an ISO taper BB but this is not true.

I have fitted Andel track cranks, Sugino XD2 cranks and (yes) Dura-Ace 7402 cranks to the DA 7400 bottom bracket, and none of them bottom out despite the slightly shorter taper. All give the expected chainline based on the 7400 BB’s spindle length of 112mm (English) or 113mm (Italian).

Square taper cranks are subject to wear from fitting and refitting, and the presence of grease on the taper markedly affects how far the crank will slide onto the spindle for a given torque on the crank bolt. My practice for fitting cranks is as follows:

• Grease the taper. I have known ungreased cranks seize and have to be hacksawn off.
• Grease the threads of the crankbolt.
• Fit to the recommended torque.
• Check chainline. if it is much narrower than expected, you may need to pull the cranks, degrease and refit. I will only degrease as a last resort, since the risk of seizure is very great on an all-weather bike.
• Ride for 100 miles.
• Re-torque the crank bolts.
• Remove the crankbolts, degrease the threads in the spindle and on the bolt, and refit to the recommended torque using medium threadlock.
• Do not retighten after this. Repeated retightening will enlarge the crank hole and eventually it will bottom out on the spindle, or even crack.

Strengthening Moulton series 1 rear forks

The series 1 Moulton fork/swingarm has a reputation for cracking or twisting, and it’s generally agreed that the design was not up to the job when combined with indifferent manufacture. The series 2 rear fork is extremely strong but is also heavier and raises the rear end of the bike (and the saddle) to such an extent that getting a toe down at traffic lights can be impossible. Most people also fine the series 1 fork to be more elegant in appearance.

I believe a well-brazed series 1 rear fork will give perfectly good service as long as the rear pivot is free to move and the bike isn’t abused in any way (some aggressive mounting techniques are associated with rear fork twist). However, a common approach is to have the rear fork strengthened by adding a second shaped steel plate underneath the existing brake mounting plate.

These photos show a typical repair before repainting. I don’t know who did the brazing as I bought the fork ready-strengthened. In engineering terms, strengthening a part by adding pieces doesn’t always work as it can create new stress rises, but this particular repair has a good track record on Moultons. Its main benefit is to create a box section at the root of the blades to reduce flex and therefore reduce the risk of fatigue, rather than increasing ultimate strength.

The repairer has also “topped up” with brass where joints were previously welded, probably at the Kirkby factory. The new metal has been neatly formed but the brazing is crude, and I will improve it by filing the brass smooth and filling the holes. To be fair, the factory brazing and welding on an F-frame is usually pretty appalling too.

The repair is more or less invisible when the fork is fitted to a bike, being on the underside. I have seen some cruder jobs where the extra plate is flat (as it is on a series 2 fork) and is visible from the side of the bike. The extra plate here has been shaped, which is not best for stiffness but does give a neater appearance.

I’m going to have this fork painted but it will be kept as a spare in case the unmodified rear forks on one of my F-frames cracks or twists. The unmodified forks I have are free from cracks and quite well brazed, so I don’t expect them to cause problems. They have also survived many decades in their unmodified form.

Was Alex Moulton right (60 years later)?

Warning: may contain iconoclasm.

The reasons that Alex Moulton came up with the original Moulton design are well-known. But do they actually stack up? Most came from an evaluation of the conventional diamond-framed racing bike (a Hetchins) that he liked, but felt could be substantially improved upon. His rationale was as follows:

All other vehicles show a trend toward smaller wheels, so why not the bicycle?

The convergence of bicycle wheels around a size of 700c (or “29er”, which is the same rim size with an off-road tyre) is no accident, even if the precise size is that of an obsolete French utility bicycle. Larger wheels roll more easily over all but the smoothest surfaces and allow sensibly-sized gearing components to be used. The trend of other vehicles towards smaller wheels is just not true any more. It’s true that railways moved from enormous 6’8″ driving wheels on steam locomotives to much smaller wheels on diesels and electrics, but this was simply because steam locomotives, especially express passenger ones, require large wheels to limit the speed of the directly-coupled heavy reciprocating components. Car wheels have become larger and larger over the decades since the Moulton bicycle was first made and, although tyres have become lower-profile, the overall size has still increased. As an example, a 1962 Ford Cortina used 145R13 tyres with an outside diameter of 13″ plus 2 x (145 x 82%) = 22″. A current Ford Mondeo, the closest equivalent, uses 215/60R16 tyres with an outside diameter of 16″ plus 2 x (215 x 60%) = 26″. Roads aren’t getting any smoother.

A conventional bicycle is sized to fit the rider, so it cannot be easily lent to another person

This is true, and can cause headaches for manufacturers, retailers and anyone trying to sell a secondhand bike, where very large frames are almost worthless! To a large extent, the Moulton addresses the problem by being “one size fits all”. Unfortunately, it only addresses the standover and seat height concerns; although most people can mount and ride it, it is only optimised for a small subset of people due to its length; taller people will feel quite cramped, which shorter ones will be unduly stretched. To some extent this could be addressed with a longer or shorter handlebar stem but Moulton never offered one; except for the longer Milremo stem on the M4, which was mainly for a sportier position. Eventually the Moulton Mini was launched for shorter riders, blowing the theory out of the water almost completely. Why the Moulton Mini had to have smaller wheels as well (a huge problem these days, with virtually no tyre availability) is a mystery to me – 16″ wheels wouldn’t have been a problem at all.

Luggage-carrying is an afterthought on most bicycles, requires accessories to be purchased, and can be difficult to fit

Since the Moulton was made, front and rear luggage rack fittings have become reasonably standardised on touring and hybrid bikes. Touring bikes normally use the fork to carry the load, rather than the frame as on a Moulton, but low riders don’t generally have a negative effect on handling. It is also possible to quickly unbolt the racks when not needed, and use a saddlebag or courier bag instead. A Moulton Mk1 (and most Mk2) rear rack is not designed for easy removal. It’s true that Alex’s Hetchins would have had trouble carrying luggage, but that wasn’t part of its design; it was a racing bicycle.

The top tube makes mounting and dismounting very difficult

This was true of diamond frames (not roadsters) in the 60s but is less so now; “compact” frames with sloping top tubes, popularised by Brodie/Kona for MTBs and then by Dave Lloyd for road bikes, are almost universal. Alternative frame designs such as mixte frames or cross frames with conventional wheels can be made acceptably stiff. In any case, the standover height of a modern spaceframe Moulton is not all that low.

Small wheels accelerate faster and roll more easily than large ones

The first part would be true if the F-frames hadn’t used such heavy steel rims and stodgy, low pressure tyres. I don’t have a typical 1960s 27″ road wheel to weigh but I doubt it was any heavier, given a narrower tyre – probably a tub – and alloy rim. Rolling resistance is always higher for a small wheel except on a perfectly smooth surface, like a velodrome. To some extent, the difference can be narrowed with a really good 16″ tyre and a latex inner tube, but you can do the same things to a large wheel. The fact that wheel sizes on modern Moultons have crept up to 17″ or the BMX 20″ size (really 18.5″) show you that the F-frame wheels were smaller than they should have been.

A road vehicle without suspension is wrong

All bicycles have suspension: the tyres. A good, supple tyre is an extremely effective form of suspension because it adds no extra weight (you need it for traction anyway) and the unsprung weight is minimised as far as possible. It would be more accurate to say that a bicycle with small wheels and no suspension is wrong, although millions of Brompton owners may disagree. A Moulton with the same tyres as a Brompton takes large bumps better, but isn’t vastly quicker. I’ve tried it. The significantly lower weight of the Brompton gives it an edge over rolling terrain, as well. Mechanical suspension is necessary for absorbing the biggest hits but adds a lot of weight and is invariably subject to wear. A steel road frame from the 1960s (or a lot earlier) just needs a repaint to make it as good as new. A Moulton from the 1960s normally requires a lot of fairly specialist work to renew worn bushes, if you can find any new parts. So, on a bike that might be expected to last a lifetime or two, suspension should really be the very last resort.

Latex inner tubes

This isn’t about small-wheeled bikes, as no-one makes off-the-shelf latex tubes for them.

The wired-on (or HP, or clincher, if you must speak USanian) tyre has been the most popular option on bikes for many decades, due to the relative ease of fitting and puncture repair. Previously, tubular (or sew-up) tyres were very popular, but now only really exist for track use and for top-end road racing.

The new kid on the block is tubeless tyres, which are very much like car tyres; there is no inner tube and they form their own airtight seal to the rim. Claimed advantages of tubeless tyres include lower rolling resistance, as there is no friction between an inner tube and the tyre, and the ability to run at lower pressures without pinch punctures. Another claimed “advantage” is an ability to seal punctures, but that is really a function of the slime that is used in them to seal any small leaks, whether from foreign objects or just from poor sealing against the rim. Tyres with tubes can use sealant too.

Tubulars maintain a niche because they can be run at sky-high pressures of over 200psi on a smooth track, where rolling resistance is lowest with a vary hard tyre. They can also be ridden when flat, which is an advantage in a road race, while waiting for the service car to catch up with a new wheel.

Nevertheless, the lowest rolling resistance on a normal road is provided by a high quality wired-on tyre with a latex tube. A latex tube:

• runs more freely because it is thinner and more flexible than butyl rubber
• is more resistant to some types of puncture because of its flexibility
• is far lighter than butyl
• takes patched repairs extremely well, because they tend to be seamless and rubber cement sticks best to latex

You don’t get owt for nowt, though, and the reasons not everyone uses them are:

• They need topping up with air every day that you ride
• They cost 2-3 times as much as butyl tubes
• They need to be more accurately sized to the tyre width
• When they puncture, they are reputed to deflate very fast

I like latex tubes. They would be a pain on a bike that is ridden daily, and even on a very long ride you have to expect some loss of pressure; I started the 113-mile Dunwich Dynamo with 160psi in both tyres (700 x 23), so there would still be enough pressure to prevent pinch punctures 8 hours later. For a bike that is ridden intermittently, they are one of the few cheap upgrades that gives a readily noticeable boost in speed. The power saving is estimated to be 5W per wheel at a decent riding pace. You can feel the difference and, oddly, hear it; a tyre with latex tubes makes a distinctive swishing noise.

I’ve only had one puncture on latex tubes, so I can’t really confirm the “rapid deflation” thing. The puncture I had was from a shard of glass that also ripped out an inch of tyre sidewall, so any tube would have gone off with a bang.

Vittoria and Michelin make latex tubes in pink or green respectively-take your pick. Check the size is within spec for your tyre; it is common for latex tubes to be suitable for, say, 22-23mm tyres whereas a similar butyl tube will fit a wider range of 23-28mm (and can be used in an even bigger tyre, if needed to get you home).

Latex should not be exposed to direct sunlight, as this rapidly weakens it, something you may have seen if hanging up toy balloons outside; they burst within minutes in strong sun.

Most latex tubes are made for 700c racing tyres but those for cyclo-cross bikes will also fit touring tyres. For small-wheeled bikes, it is possible to “roll your own” from larger tubes, by cutting and splicing them. At least one firm does this commercially to make Moulton tubes. You could try it yourself using rubber cement and an overlapped section.

Hub gears I have used

Introduction

Hub gears, where all the gear ratios are contained in a hub shell and the bike has a single chainring and rear sprocket, have a few advantages over derailleur gears:

• The chain can be made to run completely straight at all times
• The mechanism is protected from contamination and wear due to dirt
• Maintenance is usually easier because there are fewer external parts to clean
• The mechanism is quieter than a derailleur setup
• Gears can be changed when the bike is stationary – no need to struggle away in a high gear after an unexpected stop

The downsides, and why more bikes don’t use them, balance these out:

• A hub geared setup is virtually always heavier; the small size of the parts means they have to be made solidly, from hard and heavy steel
• There is usually a material efficiency loss, although this is completely negligible for most 3-speed hubs, especially as a derailleur’s chain tensioner also adds a lot of drag
• Wheel removal is slightly more complicated because of the need to disconnect a control cable
• Generally, horizontal dropouts are required on the bike to allow for chain tensioning, and most new bikes have vertical dropouts. A chain tensioner can be used but is clunky and reduces efficiency.
• Hub gears are torque-changing machines and the dropouts must resist torque in all gears except direct drive. Normally this is a modest load and can be done with anti-rotation washers that fit the dropout slots. More ambitious hub gears may need special dropouts or a separate torque arm to the chainstay.

Manufacturers

Historically, the big manufacturers were Sturmey-Archer, SRAM (who took over Sachs of Germany) and Shimano. SRAM stopped supplying hub gears and spares in 2017, leaving just two main suppliers. A niche manufacturer is Rohloff of Germany, who produce one hub gear, the 14-speed Speedhub.

How they work

Hub gears work on the epicyclic principle. An epicyclic gear has a central sun gear attached to the axle and a gear ring (or annulus) with internal teeth around it. Filling the space between the two are a number of planet gears, kept at a fixed spacing by a planet cage. The number of teeth on the planet gears is irrelevant to the ratio provided; their size is determined by the space available. The basic ratio of an epicyclic is (teeth on sun+teeth on annulus)/(teeth on annulus). For the most common 3-speed hub, the sun has 20 teeth and the annulus 60, so the ratio is 1.333.

This ratio is used in different ways by locking different parts of the mechanism to the input drive from the rear sprocket and the output to the hub shell (which then turns the spokes and the rim).

Connecting the input to the planet cage and connecting the output to the gear ring gives a speed increase of 1.333 times the sprocket speed.

Connecting the input to the gear ring and connecting the output to the planet cage gives a speed decrease of 1/1.333 = 0.75 times the sprocket speed.

The input can also be connected directly to the hub shell, giving direct drive. This has the highest efficiency; the epicyclic gears still turn in this arrangement, but losses are extremely small as they are not transmitting power.

The selection of different gears is effected by sliding dog clutches, sprung pawls or roller clutches (the last two of which can be “overrun” when not required). Conventional automatic gearboxes for cars use a similar arrangement, although the locking in those is most often done with brake bands and multi-plate clutches, which allow a more gentle takeup of drive.

Hubs with more than three speeds vary in design but most will use multiple epicyclic stages where drive is cascaded from one to the other or where only one is used at a time, by locking a sun to the axle and allowing the unused suns to rotate freely.

The theoretical efficiency loss of a perfect epicyclic stage is 3%, so a cascade of three stages in a poor hub design would lose 9.3% of the input power – quite detectable by the rider. The shaping of gear teeth is rarely optimal and there are additional frictional losses in the multiple shaft bearings, so losses are greater than this. Nevertheless, a suitably run-in and adjusted Sturmey-Archer AW with oil lubrication approaches the theoretical limit and there is no detectable drag in any gear. Remember that the jockey wheels of a rear derailleur also lose a few per cent efficiency, and a hub gear doesn’t have jockey wheels.

The hubs

Sturmey-Archer AW

The most common hub gear, with a 1.333 ratio. Steel shell. Older UK-made hubs use oil lubrication, which can sometimes make a mess of your rear wheel but is more efficient and saves any need to open the hub up for regular service, since any wear particles generally leak out with the excess oil. Later Taiwan-made hubs use grease lubrication and are generally less efficient and more problematic. A redesign (N.I.G. or “No Intermediate Gear”) adds an extra set of pawls in the driver to prevent a “no drive” position between gears 2 and 3. Unfortunately the pawl actuator tends to rattle or become bent. Grease lubrication means regular service is required, as dirt and wear particles build up inside the hub.

I like the older oil-lubricated AW and it is especially nice to use on a bike with full-size 700c wheels. I don’t like the modern one.

Sturmey-Archer SRF3 (aka Brompton BSR)

Basically a N.I.G. AW in a nice alloy shell, which ironically weighs no less than the old steel shell. I don’t like it much. It gets very sticky with age, and a five year-old hub is like pedalling through treacle if it hasn’t been serviced; a grim job involving a total stripdown, solvent clean and relube with special sticky brown grease. The hub is also less well-sealed against wet weather than an old AW, as it uses a plastic ball retainer under the sprocket instead of the old metal labyrinth seal.

Sturmey-Archer SRC3

An SRF3 with a coaster brake. Better than the SRF3 as the N.I.G. design is done differently. Still needs a regular stripdown and three types of grease when rebuilding (the coaster brake part needs an extremely temperature-resistant grease – I used Ceratec brake grease). I commuted with one of these for many years, with two-yearly rebuilds, and it was entirely satisfactory.

Sturmey-Archer AM

A “medium ratio” cousin of the AW, this went out of production in the 1950s but is a truly marvellous hub. A closer ratio is achieved by using compound (stepped) planets where one end engages the gear ring and the other engages the sun (giving the effect of a much bigger gear ring). It is still a single-stage hub, with excellent efficiency. The ratios are about half as large as for the AW (1.155), so in flattish country you have three very usable gears. The only downsides are that most are in 40-hole shells (sometimes steel, sometimes a very thin alloy) and they use a fiddly two-part indicator rod that tends to unscrew itself in use. No problem: the mechanism screws straight into an AW or SRF3 shell, and you can convert it to use an AW one-piece toggle chain with a few easily-available new parts. In fact, most of an AM can be swapped out for current production AW parts except the epicyclic gears themselves.

Sturmey-Archer FW

The four-speed wide-ratio hub used on most F-frame Moultons. Not made since 1969. There are two suns which can be alternately locked to the axle, and stepped planets to engage whichever sun is in use. Gearing is closer than an AW, and reasonably good for general riding. First gear (or “super-low”, as SA would have it) can be tricky to engage, as you are overcoming a lot of springs to lock the second sun gear to the axle. The shifters tend to wear out in this position due to the high cable tension rounding off the 1st gear detent. They can be sharpened up on a grinder, if you have the time and patience. I’m not keen on the FW: it’s a clever design but needs a lot of mechanical sympathy to use, and spares are very scarce now.

Sturmey-Archer ASC

A 3-speed fixed gear. Never sold in huge numbers, and if you ride one, you’ll see why. In theory a fixie that makes getting up and down hills easier is a great idea, but changing gear requires your legs to instantaneously change speed too! To make the hub work, there are no-drive positions between all gears, and significant backlash (play) when the hub is in a gear. This feels like riding with a very slack chain. First gear requires huge cable tension (an ASC is a hack of an FW) and is prone to slipping out of gear, just as you are straining to get up a hill. It sells for big money and the unique ASC shifter sells for even more than the hub, but it’s best avoided. If you must try one and can’t find a shifter, you can modify an FW shifter to work by grinding the 1st gear position deeper on the pawl plate. It’s a fiddly job and yes, I have done it. The best thing about the ASC is that it’s better than the…

Sturmey-Archer S3X

SEX – get it? A modern version of the ASC. Looks nice, has a choice of nice shifters (which don’t work with an old ASC),is otherwise not great. There is even more backlash than in the ASC, it is poorly-sealed, there are huge no-drive positions, it’s inefficient and noisy in 1st and 2nd gears, and it’s just as prone to slip out of 1st gear as is the ASC. The design is also poor, with the sun gears being locked by very small dogs on a sliding axle key. Mine looked worn after 200 miles. The best thing about it is that it is unusually easy to strip down.

Shimano Nexus 8

An 8-speed gear aimed at utility bikes. Usually comes with a roller brake, which is a kind of drum brake that runs in grease. I quite like it. The range is good enough for most terrain and the ratios are reasonably close. It is very, very heavy and there is a detectable loss of efficiency in some gears. Taking the rear wheel out is an odyssey best not attempted at the roadside; get Marathon Plus tyres to avoid punctures. The main disadvantage of these is that they run in grease and no-one can overhaul one fully due to its complexity. Even a Shimano dealer will just dunk the internals in a refresher oil, which you could do yourself. They are fairly cheap so you could do the same dunking operation yourself (the internal unit is easy enough to remove) and buy a new one if anything goes wrong.

There are various models available. Some have more ball or needle bearings inside for a small* efficiency increase, and are worth seeking out.

Sachs/SRAM Spectro T3

A jewel-like little hub fitted to Bromptons in the mid-2000s while Sturmey-Archer was moving to Taiwan. Ratios are slightly wider than an AW/SRF3/BSR, but not by much. Always greased, but the original SRAM grease for them is no longer made. SA grease would be OK. The hub is a bit draggy due to tiny gears but it is fairly easy to rebuild. Sealing is nowhere near as good as on SA hubs. The main problem with these now is parts availability – most of the remaining spare parts stock is in Germany, and the retailers with stock won’t ship to the UK since Brexit.

Sachs Duomatic R102

A two-speed hub where a slight kickback on the pedals swaps from one ratio to the other, and pedalling back further engages a coaster brake. Reliable and beautifully made, these were used on the Moulton Stowaway. Otherwise, it has the same pros and cons as the Spectro T3 hub. These, peculiarly, use two types of grease AND oil lubrication. I imagine it all mixes together in the end!

Sachs Duomatic R2110

A redesign of the R102. I have rebuilt one of these but not actually ridden it – sorry! It has plastic components and is not as robust. You won’t see many around as they weren’t made in huge numbers.

*Sturmey-Archer tried ball bearings for the planet gears about 100 years ago and concluded that they made no real difference. However, that was a simple 3-speed running in oil. A modern hub gear with grease lubrication and multiple stages is likely to benefit more.

Brakes for F-frames

The brakes that come with most F-frames are very heavy chromed steel calipers with equally heavy (and none too comfortable) levers. Despite being flat bar levers, they take a “road” cable, with a barrel nipple on the end. The brake blocks were originally John Bull Rubber no.25 in metal shoes. After 60 years or so, the stopping power of this setup is very limited, and almost non-existent in the wet, due to the chromed steel wheel rims used.

Some models come with GB (Gerry Burgess) Sport Mk.III alloy calipers and matching levers, which weigh far less but are hardly more effective at stopping the bike.

If you must stick with the chromed steel rims (and they would need to be in exceptional condition – most well-used bikes have worn them down to rusty unplated steel on the braking surfaces), the best upgrade you can make is to fit Fibrax Raincheater blocks. These have been around since at least the 1970s and have a leather face to absorb water and give higher friction against wet chromed steel. They must never be used with alloy rims, as they will chew through them in no time.

For alloy rims, Kool-Stop salmon pads are the best. They have high friction, work very well in the wet and don’t embed grit or chips of alloy, which kill rims before their time.

Replacing the calipers is fraught with difficulty for two reasons:

• the brakes should pull from the right-hand side, looking at the caliper, because this is how the frame cable guides are arranged. Most modern brakes pull from the left, because they are designed for countries where the left lever operates the front brake, not the UK where the right lever operates the front brake.
• there is limited space between the rear brake mounting hole and the “cup” to which the suspension block is riveted.

Brakes with a right-hand pull tend to be vintage single pivots: GB, Alesa, Weinmann. These are worth upgrading to if you want to save weight.

Alhonga make dual-pivot brakes with right-hand pull but these will not fit into the space available. You will need to include a deep stand-off washer of about 12mm tall between the caliper and the rear forks. These are sold for BMX use and can be filed as necessary. I found it was necessary to chamfer the washer to get the caliper at the right angle, more or less parallel to the rear forks. Inserting a spacer has the added problem that it brings the RH rear brake block mounting nut very close to the chain, and a normal cylindrical washer tends to leave the caliper angled upwards, making things worse.

I used Tektro drop bar levers to pull the Alhonga calipers on my retromod speed, and they are not ideally matched; the levers pull rather too much cable, so the brakes are firm but not very powerful, much like they would be if using V-brake levers. Unfortunately you will not find much compatibility information available for obscure low-end components like these. The brakes are safe enough but not a patch on a properly matched set of Shimano dual-pivots and levers (which you can’t use, as the brakes are left-hand pull).

If you want good reliable braking above all else, an optimal solution (without frame modification) would be a rear coaster brake and a good front dual-pivot. A caliper with left-hand pull works well enough at the front as there is a good distance between the single cable guide and the brake, far enough for the cable housing not to be contorted into a tight radius. All coaster brakes have a reaction arm which must be clipped to the LH chainstay, but this is no problem. You can also get internally-geared hubs with coaster brakes, including the 2-speed models (vintage Sachs Duomatic or modern Sturmey-Archer S2C) that need no cables to the rear of the bike.

TSR in bits

I’ve been meaning to have the TSR resprayed for years, since the factory ivory powdercoat is (a) not shiny (b) shows the dirt, especially where black grease leaks from the pivot and (c) it’s flaking off wherever there’s a sharp edge, like at the dropouts. To be fair, it’s not the worst powdercoat I’ve seen (that was on a 2008 Brompton, where most of the paint flaked off after a year of dry rides), but it’s nowhere near as good as the one I had done round the corner on an F-frame.

Dismantling one of these is pretty straightforward. The pivot bolt was very tight (as it should be) due to my liberal use of threadlock after I took it apart to check it in 2019. The pivot itself was in very good condition with no play; it hasn’t done a vast mileage, and I have kept it well pumped with grease. It will get a new pivot kit anyway, since the bushes have to be drifted out and get a bit dented in the process.

The bottom bracket only just came out and it pulled apart in the process. I therefore don’t recommend Token cartridge BBs any more. As the threads are obviously a bit tight and/or rough, I ran the BB tap through both sides and also faced the shell, which really needed it on the RH side. Moulton obviously don’t prep them perfectly.

The fork and stirrup are in horrible rusty condition where the friction discs have been rubbing. This is unavoidable unless you keep the area soaked in GT85, which you’re not supposed to do. The SST has stainless fork ends to avoid this.

I’d forgotten how short a TSR spring is. Mine has the “race” spring with progressive winding. It makes very little difference.

I took off the head badge with dental floss and a sharp knife. I’m going to fit the older style AM head badge because the bright blue of the current type will clash horribly with British Racing Green.

And here it is in bits. When you have it like this, two things are apparent: the individual pieces are very lightweight, and the complexity of the construction shows you why these cost about £2000 new. And this is the cheap not-really-a-spaceframe model!

Interesting trivia: apparently TSR frame parts are made as a “bike set” on the Stratford-upon-Avon production line, and the front end from one will not properly attach to the rear end of another. So you can’t swap frame parts with a friend to get a two-tone TSR. The fork and (probably) the rear triangle can be swapped but the two main frame halves are matched. The F-frame has similar foibles, with rear suspension blocks not generally being interchangeable due to the rivet holes being hand-drilled.

Moulton F frame tyres and rims

With hindsight, the choice of the 16″/349mm wheel size for the original Moulton seems a little small. It was an existing size used on children’s bikes and may have been the best option available at the time, but subsequent spaceframe Moultons use 17″ or 20″ (BMX-sized, but really about 18 1/4″ with narrow road tyres) wheels for lower rolling resistance on indifferent surfaces.

The popularity of Brompton folding bikes means there is now a very good selection of road-oriented tyres in the 349 size. Schwalbe is the most prolific manufacturer and there are tyres available from about £10 to £50 depending on the weight, pressure and rolling resistance being offered.

The most important difference, however, is down to the rims. All F-frame Moultons originally came with Dunlop chromed steel rims. These rims do not have a hooked profile and are straight-sided (something currently making a comeback on road racing bikes, for low-pressure tubeless tyres). The hooked rim (or crotchet rim) was introduced to allow higher pressures, as the tyre bead locks under the hook when inflated and is far more resistant to blowing off. As a side benefit, the hook also neutralises the effect of tyre pressure on spoke tension, because the compressive action of an inflated tyre is counteracted by the outward pull on the hook.

A hookless, straight, rim cannot support high tyre pressures. The maximum pressure for a typical 1 3/8″ tyre is about 60psi (4 bar). This means that, if you are running the original steel rims, your tyre choice is far more limited. You could fit any tyre and just inflate it to 60psi, but a tyre designed for 100psi or more in a hooked rim is unlikely to perform at its best, and the sidewalls may be excessively soft. There is also a risk of blowing off, even at 60psi, if the wrong type of tyre is used. This ruins the original Moulton concept of suspension and high-pressure tyres, but Moulton had to use the technology available in the early 1960s.

The best choices for steel rims are the Raleigh Record (55psi) or the Schwalbe HS110 (60psi). Neither of these are particularly fast. If you want your F-frame to perform better, you will need to change the rims (and probably the spokes).

If you have 28 hole hubs, the older type of Brompton alloy rim is very inexpensive, at about £13 each. It doesn’t have the ideal box section but it is a reliable and fairly light rim. For 36 hole hubs on early F-frames, the only real choice in the 349 size is the Sun CR18. This is far more expensive but is a good rim and has a polished finish which doesn’t look too different to the original chromed steel.

You can use any 349 size tyres if you change to alloy rims, and run them at their maximum sidewall pressure – something that is recommended, as it reduces rolling resistance and lets the suspension do all the work, as Alex Moulton intended. Schwalbe Marathons are a popular choice. Schwalbe Kojaks are lighter but don’t roll as well as you would expect from the price and the lightweight construction. The folding bead ones are never cheap but the wired bead ones are worth buying if you see them on sale for about £15. They are reliable, if not the fastest. The Schwalbe Pro One is probably the best but, at about £50 each, you would need to be building a very special F-frame indeed. The old Schwalbe Stelvios were narrower than Kojaks and roll very well, but have been discontinued for years. if you find a NOS pair that has been well stored, it is worth buying.

There are tyres available from other makers but most are too wide for an F-frame with mudguards. Beware Chinese knock-offs – there are amberwall tyres that look almost exactly like the Schwalbe Pro One but these are actually an obscure brand using a similar logo, and are unlikely to be particularly good.

One odd issue you may encounter when fitting tyres to the original steel rims is that, because of the lack of a hook, the tyre will not centre properly. This may be because the tyre is badly made (even some Schwalbe tyres can be wobbly) but, assuming the tyre is ok, you can encourage it to sit straighter by lubricating the rim with soapy water before fitting, and by slightly over-inflating to begin with.

The Moulton TSR as a fixed-gear bike

Can we fix it? Yes, we can!

The TSR has a unified rear triangle. This means the bottom bracket and rear hub are part of the same subframe and therefore the chain length doesn’t vary with suspension movement or pedalling. This makes it possible to run a TSR as a fixed gear bike (and is also beneficial if you want to run one as a singlespeed or with a hub gear). It also has horizontal dropouts (or, at least, my TSR does), which are essential for running a fixed gear as they allow the chain to be tensioned.

A fixed gear bike requires the chain to be as tight as possible without binding, or two things happen:

• There is undesirable free play between “driving” and “braking”, which makes the bike unpleasant to ride
• There is a higher risk of the chain becoming unshipped, which can be disastrous on a fixie, especially if the chain gets caught around a pedal spindle.

Singlespeeds and hub gears can tolerate a bit of chain slack, as they have freewheels and the consequences of an unshipped chain are not usually severe.

Benefits

Turning a TSR into a fixie is the simplest and cheapest way to get the weight down to that approaching a normal road racing bike; in the low 20lb range, if you are prepared to lose the rear brake and mudguards. The other benefits of riding fixed are too many to go into here, but this is about the only way you can ever experience riding fixed with full suspension, something of a revelation on a bike where you have to pedal through the bumps.

Parts swap time

You can (in the UK, at least) lose the rear brake caliper and its cable. If you have dropped bars, you will need to keep the lever so you have the hood to hold onto. Losing the mudguards is optional for maximum weight saving, but not strictly related to the conversion.

I was running 1 x 9, so no changes were needed to the bottom bracket, crank or chainring. The rear mech, bar-end shifter and cabling all come off and go into the TSR spares box. Suddenly there are no cables to the rear of the bike, just like a TSR2 or an old Stowaway. The rear wheel needs to be changed completely, except for the tyre and tube which can be transferred. I used a Surly 130mm fixed hub, a Surly 15T sprocket and an ISO lockring, built onto a Sun CR18 rim to match the front wheel. I could only get a 32h hub, which is probably unnecessarily strong for a 20″ wheel, but it looks fine. The Surly is designed for conversions and gives a chainline of about 45mm from the centreline of the bike. This will line up with many 1 x setups, although you may need a different length bottom bracket for some TSRs. If you have a double or triple chainset, you will want to lose the redundant ring(s) and use single chainring bolts to just fit the outer ring. Finally, you will probably want a new chain to avoid shortening the one from your geared setup, asusming you will be changing back at some point.

My 56 x 15 is just under 70″ in old money, a typical “medium gear” for fixed. This will allow you to keep up down most hills without dragging the front brake, but you can still haul it over a 1 in 6 hill if you need to.

The finished conversion. As minimalist as a TSR can get

Riding experience

I’m assuming that anyone who goes to the trouble of doing this is already an experienced fixie pilot, so I won’t describe the general fixed experience.

Let’s deal with the good parts first. The bike is significantly lighter and there are no rattles due to the tight chain and the lack of, well, stuff (mainly mudguards and 2/3 of the cables). Fixed forces you to pedal through whatever the road can throw at you, and this is where the TSR comes into its own, soaking up the worst of the bumps. The suspension is unfazed by braking through the transmission and works just as normal. It makes for a fun bike for flatter roads, short blasts or just something to freak out fellow cyclists, as if a spaceframe wasn’t enough.

The bad part is that climbing steep hills is harder work than a normal rigid fixie. If I ride fixed on the road, it is usually a track bike. Track bikes have their own problems on the road as the tight angles and very short wheelbase give them toe overlap and the rear wheel can skip over bumps. One thing they have in spades, however, is fork stiffness. Steel track forks, designed for manic sprints, have round fork blades which are stiff in all directions and, when wrestling the thing uphill, there is very little power loss or brake rub. An oval-section road racing fork flexes sideways; a track bike fork does not. So, front end stiffness matters.

The TSR suspension doesn’t help at all when trying to get a 68″ gear over a steep hill. Even with the damping cranked up, the fork turns into a pogo stick and a lot of the effort goes, uselessly, into heating up those Coulomb friction discs. Moultons were designed to be ridden in the saddle, using suitably low gears, not honked uphill with the rider out of the saddle; but that is what you have to do with a fixie, if you ever want to be able to use your knees again. I seriously thought about locking out the fork completely, but that would defeat the whole object of the exercise. I’d just have a slightly heavy fixie with rear suspension.

Verdict

Partial success. I rode the TSR like this for a year, then converted it back to gears (I actually ran it as a 3-speed for a while, then back to 1 x 9). It takes an hour or two to convert between configurations, so I will ride it as a fixie again at some point, but it just doesn’t work well enough for hilly routes. Even a famously flat ride like the Dunwich Dynamo, which I have ridden fixed four times, includes a few hills that require a bit of out-of-the-saddle grunt. To make it a full success would require an on-the-fly fork lockout for the climbs. Most MTB forks have a lockout, and I’m sure Moulton could engineer something suitably lightweight for the road. All it really needs is a beefy locking pin through the bottom of the steerer to restrain the preload adjuster, operated by a thumb or bar end shift lever.

F-frame vs TSR front suspension

It would hardly be possible to design two more different suspension forks than those used on the F-frame Moultons and the TSR (or APB, or many others with the same design).

The F-frame fork is a pure telescopic design, where a long inner steerer slides within an outer steerer. Visually, it is very unobtrusive except for the rubber bellows which cover the splines of the fork. Offset is achieved in the usual way, by curved fork blades.

(Steering characteristics are determined by head angle and offset, which is how far the wheel hub is ahead of the steering axis. Increasing the head angle or increasing the offset both make the steering quicker, or more twitchy. Most bikes need some offset to avoid the steering being “supertanker-like”, where the bike wants to follow a wide curve around corners, and it is hard to dodge obstacles in the road.)

The TSR fork uses a leading link design, where more of the suspension is on show. The springing element is inside the head tube, as with the F-frame, but there are two short leading links connecting the bottom of the true fork to a separate “stirrup” which holds the wheel hub. The stirrup is connected to the spring via a sliding bush or “bobbin”, but the fork itself is rigid. Offset is achieved by the leading links themselves.

The TSR design eliminates the sliding splined bush used in the F-frame fork, a potential (although in practice, not serious) source of wear.

In terms of steering geometry, the TSR fork has the edge. A leading link design, if correctly dimensioned, compensates for the effect on steering geometry of hard braking (or a heavy load on a front rack). With a telescopic fork, the head angle steepens as the suspension is compressed, making the steering more twitchy under braking. With a leading link, the offset is not fixed. As the suspension compresses, the leading link pivots upwards and backwards and the offset reduces slightly. This counteracts the effect of the steepening head angle, and preserves the handling characteristics. For the same reason, it is important to set the preload properly on a TSR, with the leading links horizontal under a rider’s weight. Otherwise, the steering will become quicker under light braking. In the real world, most people will not notice the handling benefit of a leading link design; virtually all current mountain bikes and motorcycles use telescopic forks, despite their suboptimal braking characteristics.

Damping is handled very differently in the two designs. The F-frame relies mainly on the hysteresis of the main rubber spring for damping, with some additional effect from the metal spring rubbing against the inside of the inner steerer.

The TSR only uses a metal spring with very little hysteresis, so adds friction (or Coulomb) fibre discs inside the leading links. These are adjustable by small allen bolts. Most riders set them as loose as possible to reduce static friction, which is the main problem with friction damping.

(Friction damping was used on early car suspensions, too, but disappeared very quickly as it just doesn’t work very well. It is light and simple but static friction gives it a poor response to small road irregularities and it needs to be adjusted frequently as the friction elements wear out. Fluid damping, which does not suffer from static friction except in the seals, is universal now.)

It is also common practice, although not a factory recommendation, to regularly soak the friction discs with a friction-reducing lubricant like one of the “dry” PTFE sprays. Static friction is greater than sliding friction, so the TSR suspension lacks suppleness. I have tried adding more damping, e.g. when using the TSR as a fixed-gear bike where climbing in the saddle is not an option, but it does not significantly reduce the pogo effect on climbs and merely makes the suspension less effective overall.

In the F-frame design, the rubber boot is critical for excluding dirt from the splined connection. The TSR uses a wiper seal to protect the inside of the steerer, where the sliding bush and spring are located. Both work well enough to protect the internal parts but the leading links of the TSR are very exposed. Earlier versions of the leading link fork, such as on the ATB and APB, used a rubber bellows instead of a wiper seal but, as mountain bikers found, a wiper seal does a good enough job without acting as a potential moisture trap.

F-frame suspension is completely unadjustable, as designed. Heavier riders – bearing in mind that an F-frame was designed when people were smaller and less well fed – can add preload by inserting a short section of dowel (19mm diameter aluminium bar works well; wood can absorb moisture and expand) between the spring abutment and the main spring. About 1/2″-3/4″ is best. Add too much, and the rebound spring becomes compressed too, plus the bellows will be overstretched when the bike is parked. Lighter riders should not add preload, but can cut a small section out of the main spring and replace it with dowel of the same length. This is obviously irreversible unless you can source a spare spring. The damping effect of the rubber main spring cannot be adjusted but works well.

The TSR suspension is highly adjustable. It is best set for minimum damping, with the small bolts evenly just nipped up against the leading links. There is a preload adjuster, which should be wound up until the links are horizontal under the weight of the rider (and any luggage, if you are finicky). The video feature on a smartphone is useful here. There are also three different springs available for a TSR: soft, medium, and “race”. The “race” spring has progressive winding, with a close-wound section and a looser section. The idea is that the close-wound section becomes coil-bound under moderate compression, effectively leaving a shorter, stiffer spring for the remainder of the travel. I confess that I went from a medium to a “race” spring and didn’t notice any real difference.

The F-frame fork is relatively difficult to overhaul, and will defeat most bike shops. You need two very long screwdrivers, a lockring spanner and a lot of care (or experience). There are full instructions in my other blog entries. The TSR fork is rather easier; the main risk is of losing or mixing up the small leading link parts, such as the different-length bolts. It can also be tricky to change the bush, or “bobbin” that slides inside the steerer tube, since it uses a tight balljoint to attach to the preload adjuster. At some point, and certainly for a repaint, the shell bearings in the fork and stirrup will need replacement, and these are pressed-in. A small vice and suitably-sized sockets can be used to do this, as they are not large or particularly tight. I found one of mine was a loose fit anyway.

Overall, I don’t like the TSR fork much: it isn’t supple enough, a lot of moving parts are exposed to dirt, the fit of parts and the action of the friction discs are affected by the paint finish (which inevitably wears off and leads to unsightly rust), and spare parts, while available, are very expensive.

The F-frame fork always has a bit of play due to its prodigious length and the use of a splined connection, but all the moving parts are protected by the bellows and it generally just works with no adjustments. Spare parts, if and when you can get them through Moulton Preservation, are reasonably priced.

Moulton has other leading-link fork designs that use Flexitor rubber elements and/or fluid damping, and these are no doubt better. The price puts them out of the reach of mere mortals, though.