A Star is born.

This is a rather long story and if you have no interest in BSA Gold Stars you might be better off going back to browsing Ebay. Below is the full, uncensored story of my life with BSA’s most celebrated single and I decided to post it here is partly because anyone trying to get on with a Goldie for the first time might find it useful and partly because I think it’s a sort of testament to something, maybe a record of building unfamiliar bikes from scratch in pre-internet times when all your knowledge came from old books and magazines and you had actively to seek out people to answer your questions instead of finding them all lined up on here like a row of wise monkeys. Besides it’s too long to go in the My Bikes section! Whatever, if you have time to read it I hope you enjoy it.

.

 Who has never fancied a Gold Star? When I was a schoolkid I remember looking at the restored Goldie in Maidstone Honda dealer Paul Smart’s showroom. By the time I was riding on the road and obsessed with the Rocker era and the 59 Club, it was my dream bike. ’61 Bonnies didn’t do it for me, one Triumph twin is very like another, it had to be a Goldie. But they were too much money, I didn’t have it and even if I had, spending that much would spoil it. I wanted a fast  cafe racer not an investment that I daren’t rev too hard. In 1984, I was just about to turn twenty when I got a job in the Parts Dept of Paul’s shop; the Goldie had gone but poking about in dusty corners I found that there were odd boxes of Gold Star bits left over from the restoration. It was mostly odds and ends but there was some nice stuff; a lovely Taylor Dow alloy tank, a Superleggera fork yoke, alloy engine plates, bits of an RRT2 gearbox, a set of A65 double damped forks and loads of fiddly little bits and bobs. Then a few years later, digging about in the shop’s offsite store, I spotted a BSA frame and was surprised to find a CB32 prefix, meaning that it was a genuine Goldie.

But there was no engine or any major parts. I had a couple of bits my uncle had given me – a rocker box (missing the exhaust rocker) and a GP carb – but assuming I could buy these bits, would it be possible to get the rest? There was a Goldie engine advertised in CB at the time for £1750: scary stuff so I decided to ring George Prew to see if he thought building a bike from bits was practical. He said it was well worth getting the frame but also told me he had already bought that engine and not only had it cost him another £250 to secure it, he’d had to spend £400 on cylinder head repairs. This was well out of my depth but it spurred me to negotiate for the bits at work. I ended up doing a swap for my Honda TL125 trials bike; my boss Roy Francis was building up a collection of trials Hondas so everyone was happy.

 The engine was clearly going to be a problem but soon after, I saw a Goldie replica (B 31 frame/Goldie engine) advertised locally for a reasonable price. Maybe I could just swap frames and turn it into a genuine Goldie.  I went to have a look and it was very pretty but the engine was a monster. It was an ex grass-track motor, early ZB bottom end with a DBD top with the fins all sawn off!  The vendor told me that repro heads and barrels were available to sort it but when I thought about it  I realised it was a bad idea. I only really needed an engine, by the time I’d bought a repro head and barrel and found the right bottom end all I’d have left was the rocker box and I’d got one of them already… So what about these repro parts? I started to wonder what else was available. Looking in the comics, it turned out that ‘Britain’s Gold Star Service’, formerly run by Ken Gardner and sort of the last knockings of the Eddie Dow concern, was now part of Chris Williams’ Autocycle Empire in the West Midlands. I rang Chris and asked him how feasible it would be to construct a complete engine entirely from repro parts – and how much it might cost. He told me nobody had done it before but there was no reason why it shouldn’t work. Heads were not currently in production but were on their way and most of the rest was already available, barrel, crankcases, crank… The crankshaft is the Goldie’s Achilles heel, a riveted up thing derived from the 1920s engines. Chris had liaised with Alpha bearings to produce a greatly improved pressed-up (not riveted) crank using a Jawa speedway-derived, needle-roller big end. The crank I think was £500 and having just been offered a completely clapped out (although allegedly ‘ex-Al Gunter’) crank for £350, this didn’t seem untoward. The barrel was £195, crankcases a further £500. With the parts I already had it looked like I could build a new and improved engine for the £2000 George Prew had paid for that knackered original. It seemed to make a lot of sense. I converted my life savings into a building society cheque and posted it off to Chris.

At the time,  I looked up the Honda prices for similar XBR 500 items and found they were much the same. These were the days when the overall opinion was that ‘60s bits should be 60s prices’ which was pretty stupid and made life difficult for people to make new parts of any quality and of course the same crowd were first to complain that ‘all the pattern bits were rubbish…’ But it seemed to me that to be able to sell limited production Goldie parts for the same price as a current Honda equivalent was good going.  

I made a bit of a mistake when I compiled my parts order. When building a Gold Star from bits the first thing you need is a parts book. The second thing you need is another parts book. The first for a Goldie, the second for a B31. I didn’t realise that many of the ‘Gold Star’ parts I ordered were all straight B31 bits, Chris didn’t take advantage but some would have charged Gold Star prices.

The heads were ready in time for me to pick mine up at the Brands Hatch Festival of 1000 bikes in 1994. I also wandered over to the Gold Star owners club, thinking I should join but to be honest I was disappointed. I asked a couple of questions but received rather dismissive answers. Something bad seems to happen when people sit in deckchairs inside a gazebo, maybe it’s a territorial thing but it afflicts most clubs and makes the public image very unwelcoming. But worse than that, with £550 of my savings converted into a Goldie cylinder head, I looked at the row of identical, immaculate, glittering Goldies and to my horror realised that I didn’t actually like them or want one. They were too showy, like a Rockola juke box or a Cadillac, I’d been dreaming of the bikes I saw in the ‘Rockers!’ book or 1960s copies of  Motor Cycle Mechanics. I wanted the Ton-up experience not a Solvol Autosol habit! It may sound silly but it was what my mate Richard calls ‘Buyer’s Remorse’ and I felt pretty despondent until I spotted this very period 350 for sale in the car park which restored my faith. 

 

I put the engine together in Kent, shortly before the major interruption of moving to Scotland, and it wasn’t until 1995 that the bike was finished.

Come the time to try firing it up, it made a horrible ringing/scraping noise from the barrel, a bit like the film sound effect when somebody draws a sword from its scabbard. Squirting oil in the plug hole made no difference but it didn’t seem to be ‘tight’ and after running it a few times round the block I began to wonder if the noise was coming from the crank shock absorber (I’d never had a bike with one before) or if it was simply the noise of running in with a very acoustic big fin barrel. After a couple of hundred miles the noise hadn’t stopped so I lifted the head. Not good. A big score in the liner: the end was missing from one half of the 2 piece oil ring and the broken end was the aggressor. It was now a couple of years since I’d built the engine but I did recall something odd about the ring when I fitted it. Since there had been no broken off bit in the box I think I’d assumed it was part of the design and unwittingly fitted a ring with a bit broken off the end. Idiot; what do they say about making assumptions?

So I honed out the score with one of those three stone cylinder de-glazers but back together with new rings, it was a bit smoky and stayed that way. Pulling it apart again I found the ring gap was greater in the middle of the stroke than top and bottom. With honing, you pass the middle twice for each time you reach the ends so you need to pause there to accommodate. I’d made the bore barrel-shaped causing the rings to ‘pump’. I’m told those hones cause a lot of problems in inexpert hands; this time I took the barrel for professional honing after which all was well – and there was enough bore clearance not to fear seizures…

Performance was a bit disappointing though, after running-in the bike felt a bit stifled but admittedly the silencer fitted was an autojumble buy and was pretty muted. A mate gave me a tatty (and baffle free) Goldie silencer and it went much better. Everything was at last making sense. Now, you remember I said the exhaust rocker was missing from my rocker box? Well back then rockers were unavailable (I did pick up a NOS one for £2(!) at an autojumble but it was another inlet) To get around it I found B33 rockers will fit but since they have a ball, not cupped, end for the pushrod I needed to swap the pushrod ends to suit. The modified pushrod was slightly short to get the right clearance with the eccentric spindle adjustment so I packed the end cap with a couple of tiny washers. I wasn’t sure how well these would wear so instead of riveting the cap back in place I left it loose so I could replace them – it couldn’t fall off.

Gradually the bike got better, although it was temperamental and I was on a big learning curve with the GP. Nonetheless it became my daily bike and I loved it. It also went pretty well… One day on the Edinburgh City bypass I was chasing a guy on a modern Triumph triple, caught him on a roundabout and as we both wound up on the exit the Goldie kept with him very well until BANG! Flat out in third something bad happened. I rolled over to the hard shoulder with no compression. Back at the ranch I found the exhaust valve had bitten a mighty chunk out of the piston. 

But why did it happen? As I picked up the pushrod, the cap fell off and as I pressed it back home there was a squelching noise. Ah…being the bottom cap, it sits in oil and was sucking some in that it couldn’t squeeze out quickly enough at high revs. To my cost I had inadvertently made my own hydraulic tappet! Around this time I changed the alloy tank for a standard steel one from an A10. For some reason the Goldie didn’t seem to handle as well as my A10 had done and the tank, pretty though it was, sat much higher on the tunnel than a standard one and made the bike feel ungainly . 

 

At Stafford Show I got in early and picked up a complete rocker box for £125; it seemed a lot just to get the rocker I needed but later that day I heard a delaer had paid £250 for one. I learned that things don’t have to be cheap to be bargains.. 

That got rid of the dodgy pushrod. With a new piston fitted the bike was going well and I went back to trying to sort the carburetion. This is a bit of an issue on a Goldie. Thing is, near home there was a good and fairly quiet long straight but to check the carb properly you have to be going well over the speed limit and that means making a lot of noise on a Gold Star. More than one or two passes is likely to mean somebody ‘phones the law. But meddling with the carb on one of these can also make it refuse to start or irretrievably soil a plug so you don’t want to be doing carb tuning in some deserted outpost. I used to regard our late summer Lothian and Borders Classic Club ‘Dam Bike Run’ around the dams and reservoirs of southern Scotland as the pre-Manx fettling run and this year something interesting happened. Part way round, my battered silencer broke in half. This left the bike on a short mega and turned it into a complete maniac. It was staggeringly fast. The megaphone effect meant it spluttered and coughed between 40 and 80 but once you got to 80 it just took off and didn’t stop. Now I knew why Gold Stars were legendary. More importantly, this indicated that the fault was in the carburetion, too rich presumably if it ran so much better without a silencer. So for the 1999 Manx GP, I took a pocket full of jets and looked forward to a week of carb tuning on that wonderful Island where going fast and making noise is all accepted as part of the fun.

Unfortunately the fun didn’t last long. We arrived Saturday and on Sunday night we were all having a bit of a  dice coming back to Douglas from Peel and – again at peak revs – the Goldie went very sick, smoking like hell. I limped back to the digs. No compression again. In the light and shelter of the supermarket multi-storey I stripped the top end. A burr around the piston crown revealed it had been touching the head and had finally transmitted the blow through the rings to the ring lands which had completely collapsed on one side. A replacement piston arrived on Friday morning, or last day, and after enjoying my afternoon on the bike we all rode over to Peel for a final meal of the holiday. When we came to go would you believe I had a nail in my back tyre! My mate Peter kindly rode back to Douglas get my spare tube so I could fit that and ride home to Scotland. 

Back home I decided to check my car park rebuild and disaster struck. I thought I could detect some play in the big end and from there it got worse. The cams and followers were badly worn and when I separated the cases I found the outer race of the timing side main bearing had blued and cracked and the timing side crankcase had a crack from main bearing to camshaft spindle.I rang Chris Williams to see what he made of it all. “What end float did you have on the crank?” he asked. Er…what? It doesn’t mention end float in the BSA manual, I just assumed that since it turned beautifully freely when assembled, that it would be okay… “You need 3 thou.” Was his reply. He also thought the other wear was probably due to the cheapish oil I was using. Now what? My mate Derek Coghill came to the rescue. He turned me out a set of M33 cases and luckily for me the Goldie timing side is more or less standard BSA single – it’s the drive side that’s substantially different, reinforced with a larger bearing. But would the cases pair up?

This time I rebuilt the engine referring to Phil Irving’s ‘Tuning for Speed’ rather than the rather basic BSA handbook/manual. Ah yes, 3 thou end float says Phil. But he also said when mating different cases, make up a mandrel to represent the crank, ie a bar with ends concentrically turned to match the main bearings. Then fit it in place of the crank and tighten everything up. You’ll feel any tightness much easier with that than with the weight of the flywheels helping it roll along. This checked out fine. The cylinder base was very close too, needing only a few thou taken off with an oilstone. The mag platform needed machining to level up and I did this on the work’s Bridgeport mill. I was impressed that one case made in 1951 aligned so closely with another made in 1990. One bonus, the new case had ‘feed’ and ‘return’ cast into the oilways, it had always been a disappointment that this was missing from the repro case.

I did a few other Irving mods, making a couple of dowel bolts to aid rigidity etc, spoke to a laser cutting contact to get a 0.020” cylinder base shim cut to relieve the piston clearance problem and put it all back together.

But once back together and run in, it didn’t seem to go anything like so well. At first I blamed the new silencer I’d fitted after the Dam Bike Run. Due to the Isle of Man blow up I hadn’t really had much chance to open up since the megaphone incident. I made the daring move of fitting the absurdly loud mega and going for a risky test ride on the local straight but no, something had changed; the bike felt ‘flat’ – like the valve timing was out. Checking the openings with a degree disc they seemed to be quite a way out even though Chris Williams said they were made to BSA diagrams and anyway a few degrees here or there would make no odds. I wasn’t happy with this. In Arthur Lupton’s book ‘Goldie’ he tells how one member of the team (Roland Pike, presumably) became obsessed with perfecting cam timing. I couldn’t see him being happy with this sort of discrepancy.

At a local jumble I spotted a second hand but unworn DBD inlet cam in a box of bits. Fitting it, I found it timed up on the dot, which seemed to prove my point and even with the dodgy exhaust cam the power was much better. But it still wasn’t quite as good as before so I was delighted to pick up an exhaust cam at Stafford a while later. That opened spot on too. But strangely, when I next got a chance to open the bike up, I had a feeling the power wasn’t so good. Maybe I should go back to the carburetion.

The loose bore meant it was no surprise that the plug was always black but with fuel or oil? The bike had always seemed too rich but weakening the main jet from standard just made the engine ‘pink’. At about £30 quid a go I wasn’t keen to start buying slides of every cutaway but I bought a new standard one and then began cutting back my original to see what happened. Trouble is the old one was badly worn so it wasn’t a direct comparison. I noticed that I had accidentally set the float height about 5/ 16” lower than it should have been and although, strangely, that appeared to have no effect on the overall mixture, it certainly improved starting which was now reliably one or two kicks, hot or cold. I’d never had the expected trouble with tick-over; using the friction adjuster on the throttle (there’s no throttle stop screw on a GP) it would always tick-over reliably under 1000 revs and sing the pilot screw you could weaken it down to 500rpm but at the expense of a clean pick-up. 

The needle and jet is the main governor of everyday riding and it seemed moving mine up and down made little or no difference, I felt I wanted a couple of extra grooves in the needle and I found that GP, Concentric and Monobloc needles all have the ‘jet’ (restricted) portion at different levels,  if you modify the spray tubes to the same heights you can effectively move the jet in relation to the needle groove. What’s more, a 376 Monobloc needle has a very similar taper but starting higher for the same overall length. Crazy as it sounds, by mixing and matching I found I could create over 20 consecutive needle positions! But it didn’t solve my problem.

So what about float height? It was such a faff to adjust; wouldn’t it be great if I could alter it as I rode along… Then I could take the bike up the motorway instead of all these short hops up and down that were annoying the neighbours. I spent ages making a cable operated doofer that raised or lowered the float from the handlebar but again, I didn’t really learn anything from it.   

I fitted a 1000 series Concentirc for a while just to try a completely different carb, it had occurred to me that maybe my GP had been modified, that’s the trouble with the Goldie’s 65mm stud centres and 1.5″ choke, you can’t just swap carbs with another of your bikes. Again it made no appreciable difference and made me ever more sure this was an engine problem.

In 2005 I rode the Goldie to the International West Kent Run from Edinburgh,on the way I holed the piston chasing a TVR down the fast lane of the A1M at a steady 85. I had a spare with me – the Isle of Man incident had taught me that you can borrow tools, a workshop or a bed for the night but you’re best taking your own spares…

Replacing the piston took three hours and would have been quicker were it not for the amount of people who pulled up to see what was going on. If that sounds a bit flash, bear in mind the Gold Star is a very easy engine to work on – racers have to be. First time I took the head off I couldn’t believe how difficult it was to remove the rocker box with all those stupid studs! I slotted the ends to make it easier in future and noticed the manual kept pretty quiet about it. Later I learned my mistake, the rocker box comes off with the head, there’s no need to remove it first. Anyway as I was wiping my hands an old Sarf London type chap pulled over and asked what was wrong. When I handed him the old piston he said, “You ain’t chainged a faaaaakin’ piston, ave yer?!”

The piston was a GPM and there was a crack from crown to gudgeon pin that had simple begun to burn. The verdict was that it was a casting ‘hot fault’ that had caused the crack before burning through, so at least this was not really the bike’s – or my – fault.

I rode home again, using a lot of oil; the replacement piston – a forged Omega – and its rings were second-hand, so when I got home I stripped it again and fitted a new set.  

I had a couple more bizarre experiences for the record. Once almost home on the city bypass I came to a 50mph restriction slowed down and retarded the ignition slightly, the bike coughed. Bad sign, the more I retarded the spark the more it misfired until eventually it cut out at anything less than full advance. This is generally a sure sign of magneto failure, damn. I pulled over at a mate’s and pulled out the plug. Sure enough, no spark…er, come to that no electrode! The tip had burned completely off an N4G platinum tipped Champion plug! I guess it needed an N3! I banged the earth electrode against the wall to shut the gap and it fired first kick and ran fine. 

On another occasion I completely lost drive riding in traffic. It hadn’t jumped into neutral and the chain was still on…but as I freewheeled slower with the engine revving impotently suddenly it leapt forward again. I realised the clutch had jumped off its taper and somehow re-seated itself on the taper, very odd. Investigating I found a curious problem. When I built the bike I fitted a 4 spring Triumph clutch. To do this you need an adapter that used to be available from Eddie Dow, I got mine from Chris Williams. But I could see that my clutch sprocket teeth were worn on one side. The clutch clearly stuck out further than the engine sprocket so I guess the misaligned chain had reeled the clutch back inward until friction stuck it back on the taper. I had never thought to check the chain line when I fitted it – it was all standard stuff (assuming again).

Chris’ adapter must be wrong…but the difference was at least 1/8” and I couldn’t see how it could be made any differently, there’s not enough material in it. Then somebody gave me an original and it was identical. To make it work you’d need to space out the engine sprocket by nearly 1/ 4″ which would make the crankshaft shock absorber spring virtually coil bound. Thinking about it I realised what was wrong: the Triumph clutch was fitted to late twins and singles and the adapter is, in fact, a factory BSA part. However, the chain line on an A10 is different to a single. Twins have a small, maybe 3/16” spacer where the chaincase lug bolts to the frame gusset that is not found on a single because the twin engine is wider. The clutch adapter pushes the clutch out to suit the A10 chain line not the single’s. The only singles fitted as standard with the Triumph clutch were the alternator models which don’t have the engine shaft shock absorber so this adapter can never have been any good. Writing this, it occurs to me that I could have spaced out the sprocket and replaced the spring with a spacer, locked up solid because there’s already a rubber shock absorber in the Triumph clutch; but I have never heard this said before, certainly not at the time.  Interestingly, looking in my mid 60s Eddie Dow catalogue, the clutch adapter is listed for the twins but not the Goldie… 

The next weird problem was another phantom compression loss, this time at traffic speeds with no nasty noises.  Removing the tappet cover revealed the cause. On a BSA the round-footed tappets are offset to the cam lobe so that they rotate to spread wear. One of my tappet guides had worked loose and slowly followed the rotating tappet until it had unscrewed enough to hold the pushrod open.  While it was easy to screw it back with a pair of pointed pliers to get home, it would require a top strip to tighten it fully. As a quick fix I made up extended bottom screws for the tappet cover that rest in the spanner flats to prevent the guides turning. This held until the next top strip – but I have left the extended screws in there ever since, just in case.

I did a deal for an RRT2 gearbox to replace my STD one but found it hopeless. With a passenger, it was impossible not to make a ridiculous racket pulling away in what’s laughably called ‘first’.  I had a fair bit of BSA gearbox stuff at the time and found that you can fit a STD first in an RRT2 box as long as you also fit the third gear pair. That’s necessary because the dogs between the two are incompatible but the good news is that a STD third has the same number of teeth as RRT2 third. What you get is a first that is 30% higher than it would be in a standard box because it runs through the top gears (themselves a different ratio to standard) and is in fact the same internal ratio as a SCR scrambles first. I can recommend it, there’s a bit of a jump between 1st and 2nd but  it’s nothing compared to the ability to pull away cleanly.

Around 2012 I got talking to someone steeped in Goldie lore who suggested the magneto might be the trouble and in fact it failed soon after, the original Bakelite slip ring proved to be ‘leaky’. Replacing it got the sparks back but made absolutely no difference to performance and to be frank I was starting to lose interest. Over nearly two decades the damn thing had only ever really gone brilliantly that one time on the megaphone. Since then I’d got more into pre-war bikes, first with the 1932 Sunbeam, then the Inter Norton and they just seemed so much more rewarding.  

Last month I decided to have another go. Maybe with another 20 year’s experience under my belt, I would spot something I’d missed before. Incorrect valve timing really knocks the punch out of an engine and I was sure this was a timing issue. I set up the degree disc and rechecked it. As before I found the inlet was spot on. So too was the exhaust opening point…but it closed 15 degrees too early. Not sure I’d ever looked at that, assuming opening was the most important. But I had read in Tuning for Speed recently that the overlap is the most important part of the timing event and should be prioritised where there are inaccuracies. Overlap is the period when both valves are open at once – the exhaust still closing as the inlet begins to open.

It’s an idea that dates back to 1910 or so when Brooklands riders George Stanley and Victor Horsman accidentally mistimed the valves on George’s works Singer and found a power increase. Thinking about it they realised that the exiting gases helped suck in the inlet charge to create a kind of reverse supercharging effect.  After that Stanley used to pop off the timing cover and remove the cams in the paddock in case anybody turned the engine over and noticed both valves open at once. 

So was the loss of 15 degrees overlap a problem? I needed to compare it with the cams I had before but that was another problem. I’d sent the originals back to Chris for hardness testing when they wore and didn’t get them back. The replacement set I had sold on to a mate looking for cheap cams for a B33 project so I had no idea what figures I’d had before. I ordered another cam from Phil Pearson but even this turned out to be 10 degrees short, unfortunately he no longer supplies Vernier cams, which would at least allow me to adjust any discrepancy to the opening rather than closing point. Phil reported that he had no problem with them but it did occur to me that while racers probably use special cams, few road riders these days really thrash their Goldies so maybe nobody had noticed, seemed a bit far fetched though. I asked all sorts of people involved with Gold Stars and camshafts but couldn’t find anybody who could definitely say whether this was likely to be my problem and it seemed incredible that with a bike so popular, with so many experts, tuners and specialists, nobody could say if a 10 degree shortfall mattered! Digging through some old notes, result! I found an old bit of paper upon which I had written the timing figures of the last set of cams; remember I said I thought it had been better with just the exhaust cam from that pair? Well guess what, that cam opened way out but closed bang on 55 degrees. The plot was thickening…  

But when I spoke to Jan de Jong at ABSAF he told me to check the peak lift occurred at opening midway point, conventional thinking on most engine – but contradicting Irving’s dictum that correct overlap was crucial. Moreover de Jong seemed certain that my 20 thou base shim was the problem, interfering with the cylcinder head ‘squish’.

 This was a new thought, after all the shim has been in it since the Isle of Man blow up when the problem appeared. I spoke to local tuner Ron Lewis who has done a bit with Goldies and he told me that the affected area is the base slope of the piston dome and its relationship to the combustion chamber, it exists to squeeze out any mixture lurking around the edge of the combustion space and fire it into the main arena and shouldn’t exceed around 40 thou. I put two pieces of 0.052” solder across the crown of the piston in an X form and replaced the head. After turning over, the solder was untouched so the gap was certainly too wide. I removed the shim and tried again. This time the solder got nipped at each end and measured 40-45 thou.

I’m going to wind up here because the next step is to find another way to keep the piston clear of the head, put it back together and try it. I’m still not happy with the valve timing but you should only fix one thing at a time.

I wonder how I will feel if this is the answer. I can’t see me riding the bike as hard as I used to now, I’ve had enough accidents over the years and am happy getting my kicks at 60 on a flat tanker. Even Ron Lewis suggested Goldies are better with Scrambles cams, although he also understood why I have to solve this problem. As far as I’m concerned this long story won’t be over until my Gold Star goes like a Gold Star. I’ll keep you posted…

Cheers Rick