In my youth I owned several (mostly terrible) bikes which had a Sturmey Archer three speed gear in them. Albeit heavy, the gear was largely trouble free and resisted the ingress of dirt and water far better than anything else on said bikes.
Later I got a 'proper racing bike' but this wasn't suitable for everyday riding and keeping it in fresh freewheels and chains was getting pricey; more so when I started doing lots of miles for training.
Around this time I realised that maybe the three speed might be a good idea again, for certain types of training, anyway. Better than a fixed gear, especially in the hills. The three speed worked out OK but I quickly noticed that more gears might be a good idea... why not a four speed then? A little dabbling inside an FG hub showed me that the shifting mechanism for the sun pinions and the control rod (already then NLA as a spare part) were potential weak spots. I thought about a five speed hub, but -although they were then the most sophisticated hub gear available- these were both very expensive and had a poor reputation for reliability.
One day I bumped into a chap who had successfully converted a four speed to a five speed; I had previously realised that this must be possible, but had not appreciated that there were good solutions to the main problems.
1) the sun pinion; the FG/FW 'high gear' sun pinion has chamfers on one side of the dogs so that it disengages (or slips) when turned in the correct direction for the 'missing' (on a four speed) high 5th gear. The solution was to modify the original sun pinion or to purchase the equivalent five speed part which has square dogs. I used part number HSA345 but there is another SA part which will do the same job in most hubs.
2) the control rods and toggle keys. On the right side, an AW toggle key and control/rod can be used and this controls gears 2,3,4 in much the same way as an AW hub. On the left side a new toggle key can be made (e.g. from a modified pawl pivot pin) and this is controlled using a second (left side) lever/cable and a pushrod. Pushing the rod in on the left side shifts 2-1 or 4-5. Original SA five speed hubs (Circa 1965) worked like this with a second gear lever and a bellcrank on the left. [Some later SA five speed hubs for a while had different springs inside on the sun pinions and this was meant to allow a conventional pull chain for the left side gear control. Due to poorly chosen springs, these hubs were an unmitigated disaster, and may not be reliable even when fitted with a revised kit of parts. Many long-term users of such hubs have converted to a left-side pushrod in search of proper reliability.] Sneakily, if the LH control chain is routed behind (say) a dropout cutout, onto the end of a specially made 'L' shaped pushrod, no bellcrank or revised sun pinion springing is necessary; a 'pull' on the chain creates a 'push' on the pushrod.
I robbed a Moulton Mk1 of its FW hub internals, and used the longer axle from an FG hub plus the modified parts as above to create a set of innards which would fit in an AW shell and thence into a 120-126mm OLN frameset. I rebuilt the wheel with double spoke washers and DT DB stainless spokes, and proceeded to abuse the heck out of it for several decades. In the meantime I've owned and used MTBs, touring bikes, racing bikes etc etc; all this time the hub gear bike has done the brunt of the dirty donkey work on a diet of minimal maintenance.
It is now over thirty years later; how did it work out?
-No spoke breakages
-Hub has variously been run with; two SA triggers, Campagnolo friction DT levers, and latterly with modified under-stem-mounted thumbshifters, friction on the left, indexed on the right. All schemes worked reasonably well, obviously indexing is more convenient.
-Hub has not broken in use, ever (*).
[edit; (*) that was true in 2012 but as of july 2017, it isn't true any more. The axle, which was old before I used it for ~35 years (and now well over 50000 miles plus whatever it had in its previous life), finally broke. The bike was soon repaired using a stash of parts that I had squirreled away. Despite everything, the other parts of the gear were still in GWO.]
-Only 'in use' incidents are due to indexing going out of adjustment, or massive water ingress into hub (rumbly bearings).
- replacement internal parts used to date are; 1 packet of ball bearings (for the clutch driver ball ring), two pawl hairsprings. In both cases corrosion had necessitated precautionary replacement.
-Due to ill health on my part and another problem with the bike I effectively left it parked outside in a damp hedge for 18 months at one point; this is when the worst of the water got inside it.
-Total mileage is conservatively estimated to be 30-40000 miles, in all weathers, quite a lot of bridleway work as well. Mostly the chain just gets fresh oil, no real cleaning.
-It has gone through about ten chains, half a dozen sprockets, about a dozen sets of tyres. The frame that the hub spent most time in has had to have three or four paint jobs in the meantime.
-The chains last about half as long again as I would happily run a derailleur chain/block. I reckon I would have replaced the chain and freewheel about fifteen or twenty times in that use, even if I'd have done more maintenance.
-I have found that with a few tools (two spanners, a hammer (or rock) and a punch) the whole gear can be stripped down inside 15 minutes if necessary.
-I'm still using it; I've done about 500-600 miles in the last month and a half alone. I generally go out with a saddlebag containing a few odds and ends, and might ride for four or five hours, taking in bridleways and so forth as the mood takes me.
What is it like to use then?
Well most of the time it is like a closer ratio three speed gear with 1 and 5 for emergencies. 4-3, 3-2 shifts are simple; as soon as the lever is moved, the gear is definitely in. Shift 2-3 can be performed under a fair load; too much load and it won't go, but damage is very unlikely. Shift 3-4 requires no-pressure, some pedal movement (in either direction) to be sure that the gear is in. Shifts 2-1 and 4-5 require the left lever to be pulled whilst pedalling forwards ; you feel the gear go in, the lever moves to the end of its travel, and then you can apply as much pedal pressure as you like. Shifts 1-2 and 5-4 are a bit like 3-4; you have to move the pedals and wait for the hub to do its stuff. All this sounds more complicated when written down than it is in practice. Similar techniques in fact apply to all known SA five speed hubs (albeit with slight revision to allow for single lever operation). Even though later versions use different internals, the required shifting technique is similar. It should be clear from my experience that if the correct technique is used, this version of the SA-5 is actually very durable and reliable. [As soon as I realised that the later versions need a similar shifting technique, I stopped making foul crunching sounds when riding my Pashley town bike, too.].
The gear is mostly very quiet; the overrun low gear pawls make a tick-tick-tick sound in the high gears but are much quieter in gear #3 for some reason; barely audible at present. In the low gears you can just about hear some gear noise, but not much. It is on average much quieter than any derailleur geared bike I have used.
What about efficiency?
In the HPVA archives there is a test of various transmission efficiencies. Deraillaur gears, hub gears (including a Rohloff) were tested. Just for fun, they also tested some old three-speed gears. The old three-speed gears showed about the best average transmission efficiency as well as about the highest single gear efficiency. From this I take it that the reputedly 'inefficient' hub gears have perhaps been unfairly slighted. On the five speed, in gear three the hub is subject to little or no losses of any kind as it is direct drive (say 98% then). In flat and rolling terrain most riding is in gear three with my set-up and most of the rest is in gears 2 and 4. If we assume that kind of duty cycle and (say) five percent additional losses in 2,4 and ten percent losses in 1,5 then the average efficiency is still 96 or 97 point-something percent. I reckon this knocks spots off anything else for my use and probably for many other people's use too.
If the gearing is set for loaded riding than again gear three might be most used and a similar average efficency obtained. Only riding loaded in extremely hilly terrain would nudge the average well below 95%.
If my estimates of efficiency are hopelessly wrong and it is actually (say) twice as inefficient in the other gears, then again the average efficiency won't drop below 90% for most usage patterns and even in extreme conditions is unlikely to drop below about 85%.
what about gear range and gaps?
With five gears there will likely be big gaps and/or a compromised total range. Arguably the five speed has elements of both. I run 46x18 on 700c tyres. If you wanted to compare the gears you get with a derailleur set-up then (give or take the odd 0.2 of a tooth) it would be like having a 12,14,16,18, 21, 23, 27 T block, which sounds quite sensible until you realise that some cheeky get has pinched the 16 and 21T sprockets. This is surely a mad set-up...? Well yes it is a culture shock and it requires that you are able to pedal at a slightly wider range of speeds or accept that you are not going to be going optimally fast all the time. The low gear isn't really low enough with a 'gear three normal' set-up like mine, so honking or walking are on the cards for some riders on very steep gradients. For moderately loaded riding a 40 x 22 gear is perhaps a better choice; this gives ~62" in 4th for tapping along and a ~32" bottom gear.
What effect do the gear gaps have then?
Well it depends on the rider. Everyone has a range of cadence and effort that they are happy with. If this isn't enough to straddle the gaps in the gear range then there is a problem; most likely you will have to use a lower gear than you might otherwise for a while. Thus 20% gearing gaps are dealt with OK if cadence and effort can vary +/- 10%. But if the gear gap is 25% then maybe the speed will be up to 5% less than optimum, for about 20% of the time. Supposing you are a fussy pedaller and are only happy +/- 7.5% on cadence/effort, this leaves up to a 10% speed reduction which will come into play about 40% of the time. In the latter case the net effect is that your average speed may be reduced by 2% and in the former case it will be reduced by just 0.5%. This is a much smaller effect than I expected. In fact it is probably not as bad as that even, because if you ride at a lower effort in a lower gear for a while, you are having a bit of a rest, and may be able to work a bit harder than normal later on.
I think that the ability to pedal smoothly over a wider range of speeds is worth cultivating; it is in good part why riding a fixed gear is such good training, and I think it may make you a better cyclist when all is said and done.
What of the future?
I am experimenting with a contact seal near the clutch driver ball ring to help stop the water from getting in. I am also experimenting with the use of a thixotropic semi-fluid grease instead of gear oil. The seal should help the grease stay inside the hub as well as keep the water out, and this may allow a higher fill of grease (I'm aiming for about 1/3 full) and longer intervals between lubricant replenishment.
A recent stripdown (one of only three or four in as many decades) showed that there isn't much wear on any of the gear parts but there is some pitting from (self-inflicted) corrosion. I lightly reground the corroded clutch driver ball race and later fitted new balls in the ball ring. The new balls made a bigger reduction to the rumbling than the regrind. I may yet replace the threaded ball-ring race as this is somewhat pitted from corrosion and I don't fancy regrinding it.
Overall
I have been very happy with my SA5 speed gear; although I am sure that many others will say that they couldn't ever do this, I found that -for the use I have put it to- I can live with its shortcomings and I certainly have come to appreciate its strengths.
The potential drawbacks of such a hub gear (gear range, gear spacing, efficiency etc) in reality mostly have much smaller consequences than you might think likely. The benefits are that you get to spend much more time riding your bike and much less time faffing about cleaning or replacing transmission parts. The chains I use presently cost £2.40 each and I expect to get 3000-5000 miles out of them. The gear itself should last decades with just a little care.
I went on a (for me) fairly typical ride a couple of days ago; 15 or more miles on bridleways variously bumpy, rutted and muddy, followed by about 25 miles on tarmac to get home again, the last 20 of which were knocked off at evens or better courtesy of a tailwind. I can't think of another bike/gearing setup (except perhaps another hub gear) that would do that job as well with so little collateral damage/maintenance required.
I hope this write-up is of interest to others, I can post pictures if anyone is interested.
{edit- pictures of the tatty bike added}
- 'bout 130 miles in the last four days, maybe 40 of which on bridleways and byways...
- frame is a Claud Butler (circa 1968), 531 main tubes and forks, pencil stays. Braze-ons (inc. canti bosses) are mostly Brucey-add ons. Bike weighs about 28lbs or so without the saddlebag.
- partial black paint is recent and is to test for a reaction with the (many) underlying layers prior to the (long overdue) next respray. I am thinking of Black with white panels. Not that there is much danger of it I suppose but I am wary of it looking too nice in case someone (mad) takes a fancy to it.
cheers
Before that, the Moulton was fitted with an FW 4 speed. I've also got an 4 speed FG on another bike.
