SmedSpeed XS650


Technical Stuff


Inertial Torque and Rephased theory

A conventional two cylinder 360 degree engine (XS650) has its crankpins all in the same plane, both pistons reaching TDC (top dead centre) and BDC simultaneously, occasionally called a flat-plane crank. It is this particular configuration which generates a force called inertial torque. This is independent of the main torque output generated by the combustion process and happens entirely because of the crankshaft layout.

To understand it, first imagine a crankshaft on its own, no pistons or con-rods, spinning in friction-free bearings. There's nothing to slow it down or speed it up so it just keeps spinning at a smooth, constant speed. Now attach the con-rods and pistons, for the sake of this experiment, we'll make them frictionless too. As you spin the crank the pistons move up and down, and the whole system keeps on rotating and reciprocating. At this stage there's no combustion or valve gear or anything to confuse the issue, and crucially, there is no energy being put into our system and none being extracted or lost. This matters because it is a fundamental law of Physics that energy cannot be created or destroyed, only converted into another form - Mechanical Engineers refer to this as the first law of thermodynamics.

Within this system, the pistons are travelling at highest speed when they're approximately half way down their cylinders; at this point they have a lot of kinetic energy. Yet 90 degrees of crankshaft rotation later, both pistons are stationary, either at the top, or at the bottom. Their kinetic energy hasn't simply vanished because it can't: instead it's been transferred to the crankshaft, which was responsible for slowing the pistons down. As a result, the crank itself has increased its speed. Another 90 degrees on and the pistons are back up to maximum speed, accelerated by the crank which has returned some energy to them and in turn, it's slowed down again.

In a full rotation the crank will have sped up and slowed down twice, generating rapid negative and positive torque pulses completely independent of the torque produced by the combustion process. This constant pulsing torque is like a background noise to the main torque output. Although this torque production is not a result of burning fuel/air the scale of it is massive, dwarfing the conventional, output torque by up to a factor of four.

In a conventionally laid out XS650 engine at 7500rpm for example, the inertial torque swings from around 250 to –250lb.ft twice in every revolution of the crank.

Eliminating this inertial torque means you no longer have to make such great allowances for it in the driveline. Many clutches for example include metal springs or rubber bushes as a cush-drive in the back plate to soak up the inertial torque pulsing so it doesn’t damage the gearbox. With this gone the cush-drive can be deleted and the connection between crankshaft and rear wheel is more direct, improving feel for the rider.

90 degree V-twins are famous for their drive out of corners, and sure enough, they have almost zero inertial torque. As one piston is accelerating so the other is slowing down, and when one is stopped the other is at maximum speed. This is an important factor in why Ducati’s have been able magically to accelerate out of corners faster than more powerful conventional fours.

On a Yamaha XS650 that has been rephased, these fluctuations are all but eliminated. In this layout one crankpin is located at 83 degrees advanced to the other. So as one piston is slowing down and losing energy to the crank, another is almost exactly at the point of maximum acceleration (actually 75 degrees ATDC) taking the same amount back. At no point do both the pistons stop together, as they do on a traditional XS650 crank. Instead the energy flow is evened out and the rotation of the crank is almost completely smooth and steady.

The Maths…..

I’ve used a XS650 standard bore engine ( 75mm ) as an example, revving to 7500 rpm, and to make the calculations much easier, I’ve assumed the piston movement is symmetrical so it’s at maximum speed half way down the bore. In fact this is only true with infinitely long con-rods but real ones make the mathematics too complex, and the result isn’t very different anyway.

A piston mass of 0.4kg (I’m going metric for this as it makes life much easier...) although as the con-rod and gudgeon pin move with the piston their weight should be included too, so 0.4kg is rather low and the real inertial torque figure will be much higher.

The aim is to calculate the acceleration of the piston as it moves from stationary at TDC (top dead centre) to its maximum speed, and from that we can calculate the force exerted on it by the crankshaft through the con-rod. Knowing this force and the distance of the crankpin from the crankshaft’s centre (half of the engine’s stroke) allows us to calculate the torque.

* Step 1 – Find the piston’s maximum speed (V)

In our XS650 engine this occurs approximately halfway down the bore where piston speed is the same as the crankpin speed. At 7500 rpm the crank rotates 125 times per second. The diameter of the circle the crankpin (big end) travels is π x the stroke, 74 mm on the XS650. That’s a peak speed (v) of π x 0.074 x 125 = 9.25m/s (metres per second) That’s 0-34mph in 37mm

* Step 2 – Find how much time (t) the piston takes to travel from TDC to max speed:

At 125 revs per second, one rev takes 1/125 seconds = 0.008 seconds. Max speed is reached in 1/4 turn of the crank, so t = 0.008/4 = 0.002 seconds.

* Step 3 – Calculate the piston’s acceleration (a), using a = v/t

a = 9.25/0.002
= 4625 m/s-2 (That’s huge. Earth’s gravity, 1g, is 9.81 m/s-2 so a XS650 piston experiences up to 470 G... this is why they need to be strong

* Step 4 – Calculate the force of the con-rod pulling and pushing on the piston to accelerate it, using force = mass of the piston (m) x acceleration, F = ma:

F = 0.4 x 4625 = 1850N (N is Newton’s, and 9.81 Newton’s is the same as 1kg, so a piston experiences more than one and three quarter tons force as it accelerates at this rpm).

* Step 5 – The torque (T) generated by this force on the crankshaft, using torque = force x distance from rotation centre (half the stroke, 0.0370m)

T = 4625 x 0.037 = 171Nm.

So we British can understand, that’s a massive 123lb.ft of torque more than the engine makes in the usual way! And this is only from one piston, and there are two doing it...

In other words, spinning at 7500rpm, the inertial torque in a XS650 engine due to the pistons peaks at 246 lb.ft.

And you wonder why it vibrated …..!



The XS650 has relatively recently gone through quite a rebirth, although its popularity in the late 80s and 90s waned, with the interest in vintage Japanese bikes at an all time high, it is only right and proper that the best Japanese twin made should be at the height of this interest.

There are a whole host of improvements available for these bikes and there are many manufacturers worldwide who will provide parts and services to keep these twins alive. I supply a few of them, rephasing is one of the best known current modifications. There is an upgrade for almost every part for the bike that is “evolution”. These bikes, like every commercial enterprise, were built to a price and some of the design reflects this. However, now many have the luxury of an attentive owner who is quite willing to improve the original design by spending money on quality upgrades.

One of the improvements possible, although not one often addressed , is cam shaft replacement. The XS650 camshaft was never a performance item; its design brief would have called for good fuel economy, ease and cheapness of production and reasonable power for an engine of its type. Recently I set up a dial gauge and measured the valve lift, measured against crank rotation. I have done this dozens of times to check what the valve is actually getting lifting, compared to what the cam vendor or manufacturer states.

What I found was quite a revelation to me. I was fully aware that cam design along with everything else has taken huge leaps forward since 1968/9 when these engines were on the drawing board. The cams in an XS650 both early and late styles lift the valve of the seat very slowly.

These are the figures for the inlet valve lift measured at degrees before top dead centre (BTDC)

Inlet valve lift

Degrees BTDC








11  BTDC



Similarly the inlet valve closing after top dead centre ATDC

Inlet valve lift

Degrees ATDC












If the inlet valve tappet clearance is set at 0.002 the overall inlet cam timing event is 93+180+122 = 395 degrees, if the tappet clearance is opened up to just 0.006 then the inlet cam event total is 49+180+77 = 306 degrees.

The late XS650 stock cam holds the valve off its seat for nearly 45 degrees before it actually lifts more than 0.005; this is just wasting power, and compression.

Setting the intake valve clearance at .006" versus .002" means you have lost 89 degrees of the valve opening event. (44 degrees on the intake opening side and 45 degrees on the intake closing side) It won’t lose you any flow or power however, as the valve was only a couple of thousandths of its seat.

The intake valve is just loitering off the valve seat but is still open when it could have opened 44 degrees later, and closed 45 degrees earlier, allowing for greater dynamic compression, and more time for the valve to lose its heat through the valve seat. There is no flow advantage with the slow opening and closing rates that leaves the valves open for 89 degrees a few thousandths off the seat except burned valves, lost mixture and compression.

The early Yamaha XS650 ( XS1, XS2 etc) had tappet clearances of 006" and .012" for the intake and exhaust respectively. It’s one of the reasons the earlier engines perform better, they do have a slightly wilder camshaft (not much) but most of the increase in compression comes from just keeping the valves shut longer .

Setting valve clearances to the LATE stock cam specs just contributes to reversion and lost compression with these antique slow cam lobe opening and closing rates.

The design of the cam echoes the “cooking model” cams of the British bike industry, but as Yamaha engineers had no previous four stroke experience what else could they copy err…. benchmark.

I have checked stock and performance cams from the opening to the closing in thousandths per degree every 10° of crank rotation, just to check cam profiles. Most of the newer design (non symmetrical) cams open the valves faster than they close them, but both rates are very much faster than the old style cam designs of the past. Computer modelling cam profiles helps in this area.

The late XS650 cams set at .002" intake tappet setting is ridiculous. I can’t think of another OHC engine with valves this size that has such small clearances, the camshaft has a run-out of over 0.002 in many cases and setting the tappets to this small clearance will allow the valve to stay open perpetually. It is an effort by Yamaha to reduce engine noise, but this is not the way to do it.
I set all the tappet clearance at the early settings of 0.006 and 0.012. It allows in more oil, and allows for a generous growth in the metal parts, remember….. A loose tappet is a happy tappet.



A lot of people ask me what oil to run in their XS650, it is one of those emotive subjects where everyone has their own ideas. I did some research a few years back and decided to run mine on Mobil 10W 60. It is a high end synthetic, designed specifically for use in older engines to prolong their life; it sounded just the ticket for an aging XS650.

It is common to find XS650 owners who won't use synthetic engine oil, claiming it is too slippery and the clutch will pack up. Similarly they say that the modern oils don't cater for engines with combined gearboxes, and to a large extent modern oils do not address this phenomenon. This is due to a lack of an additive called ZDDP; one of the main additives that ensures that the gearbox is “protected” is ZDDP. 

A lot of people will already know that ZDDP is Zinc Dialkyl Dithio Phosphatese, and most importantly it provides excellent anti wear protection. ZDDP used to be in lots of quality engine oils. When it greatly increases in temperature it fragments and coats the metal surfaces with Zinc, particularly useful on gears, camshafts, flat top cam followers etc. The Phosphate part then acts as a dispersant making any carbon produced dissolve in the oil and not stick to the inside of the engine.

However, it was found that it caused problems with Diesel Particulate filters, so it has been dropped from a number of oils. It is also said to be less than friendly to catalytic converters, but it seems this is only the new generation of cats. Bear this last piece of information in mind, when selecting diesel oil as suggested by many US based XS650 based forums and businesses.

Mobil-1 Extended Life 10w60 does indeed include ZDDP. Not only this, but it has 1300 Parts per Million (PPM) which is comparable to the obsolete but superb Mobil 1 Motorsport 15w50.

The "10w60 extended life Mobil-1" product is fully synthetic, made from virgin base products.

I have been running it in my XS1 for years, and all the rest of my fleet. I experience no clutch slip on any of the bikes. Just thought I would share my findings with you out there. Hope it helps.


Yamaha XS650 Hi Performance gearbox


The Yamaha XS650 gearbox is a robust design, a vast improvement in terms of the selection and quality of manufacture compared to its contemporaries. It was improved somewhat when Yamaha changed the gear tooth shape in 1978 to a more rolling design. By increasing the area of the gear tooth in engagement at any one time, it decreased individual tooth loadings. To differentiate the old and new tooth shapes, Yamaha machined a shallow groove around the periphery of the gear.
However the gearbox is not flawless, and if you ride really hard or use the bike for competition purposes, then the gearbox can experience mechanical flaws. Pictured below is a gear taken from a gearbox on a road bike, one of the driving “dogs” has broken off the gear. This is almost always caused by decelerative forces, not the driving power or torque, in effect reverse torque production, the end result when the gearbox is subjected to the immense forces produced when substantial engine braking is used.
gear2broken gear
This is the 3rd gear main-shaft on a late XS650 gearbox; as can be seen the driving dogs are not overly wide, when the bike is ridden hard and subjected to hard downshifts, the inertia of the bike is transmitted through the gearbox, and ultimately in this case through three drive dogs. The forces are high enough to shear the driving dog from the main body of the gear.

However as we all know Yamaha are past masters of designing and using the same parts for many different vehicles in their line up, it’s called parts commonisation, and is one of the secrets to minimising your production costs and maximising your profits.

Often you see for instance a clutch lever appear on bikes from different years, when you check it out you find they have using it for 25 years, clever stuff. This methodology extends (thankfully) to the engine designs as well. One of the ways this can help XS650 owners is in the gearbox, the much stronger gearbox parts from Yamaha XS750/850 (the triple made from 1976 to 1981) can be used in the XS650 gearbox from 1978 to 1984.

The main parts that mostly suffer in the XS650 gearboxes are the sliding gears. These gears are the ones that engage with the freewheeling gears to transfer motion to the rear wheel. The Yamaha XS750 gears can be used to provide a much stronger alternative to the original items. In the photo above, the width of the engagement dogs (arrowed) can clearly be seen to be wider and therefore stronger. This is the 3rd gear main-shaft which is the one that suffers the most from abuse and high torque, the XS750 gear ( 1J7-17251-03-00 ) pictured has a massive 35% wider driving dog.

The countershaft sliding gear can similarly be replaced with an XS750 item and is pictured below; in this case the XS750 3rd gear countershaft (1J7-17131-01-00) has a 25% wider driving dog
The gears listed above have accompanying matched size gears which allow use of the wider drive dogs. These must also be used, this is very important; failure to ensure that the matching wider dogs and wider slot gears are used as pair will result in at the very least terrible gear engagement.
The parts numbers for the matching gears being

1J7-17231-03-00    3rd gear pinion
1J7 17151-04-00     5th gear pinion

The gears that are being replaced have the same number of teeth so the internal gear box ratios will remain the same. Pictured below is the whole new assembly complete with the XS750 gears prior to having the c clips and shims installed.
Assembled gear cluster
I sourced a good used 1978 XS750 gearbox from Fleabay for £35, stripped out the relevant gears and built the whole assembly back together, checking gear engagement and end floats on each gear. I de-burred the individual gear pinions, and then shot-peened the whole lot for increased fatigue resistance. This is the best you can do to “Bullet- proof” the transmission. Check also the Kickstart gear engagement, as it usually only engages ½ way. This can be rectified by turning the back of the shaft away, and milling 1 mm from the crank cases. It does require a spacer be made also. If you need to do this mod or the gearbox work please email me for more detail, as it is quite involved, and surprisingly not shown in the Haynes manual! 


Anatomy of an 880 cc engine build

I build quite a lot of Yamaha XS650 engines, all the way from completely stock to full race and everything in between. The XS650 is an excellent design and very resilient to complete disregard of maintenance and outright abuse, and is well enough made to allow significant power and torque increases before major parts start breaking.

I was approached in 2015 by Mike Guilford and his son Martin to assemble an engine for a sidecar Moto-cross team. Since then we as a team have had a fair deal of success, culminating last year (2016) in Martin winning the Twin-shock chairs championship on an 840cc engined XS650.

They say that “racing improves the breed” and it could not be truer. In the last two years we have learnt a lot about big bore XS650 engines, and how to keep them in one piece.

With this in mind I decided to do a step by step build on an XS880 engine. These engines can be reliable when built properly, and strangely enough it's not that much hard work to get right, attention to detail and cleanliness are the objectives of paramount importance.

To build a rephased 880cc engine the following specialist parts are required. The rest of the engine internals can be checked for wear and tear, and if it's all OK can be re-used. The gearbox does get a tough time with big bore engines, but if you don't ride a motocross rider, they should be OK. (See gearbox modifications article  to see what can be done)

Heiden bare big fin cylinder
87mm bore ductile iron cylinder liners. (Ductile iron grade 700-2 is twice as strong as cast iron)
87.5 mm copper big bore head gasket
Bored crank cases
Rephased 533 crank
277 rephased cam shaft

Step one is to acquire a good set of crankcases, whether it be the set on your existing engine or a set from your spares stock. (All crankcases from 72-84 are essentially identical, only XS1/B cases are different). For this build I purposely chose the roughest set in my spares stock to illustrate what can be done with a little hard work. Inspect them carefully for damage, the kick-start lug breaks, damaged threads, and very rarely cracks near the studs.
Cases before any work
Cases before any work

Clean them till they are spotless, and then clean them again! Remove all the old gasket sealer that Yamaha used, it will have turned brown with age. Remove the starter boss, this can be done by removing the screw and then putting the crank case on the gas hob, in around ten minutes it just falls out.
I rarely run starter motors, they weren't very good when they were new, and even the best ones won't get far with an 880cc engine. I turn the defunct starter motor into an additional sump for increased cooling.
Remove the cylinder studs, do this with heat and two nuts tightened against each other, do NOT undo them with a pair of plumbers Stillsons or vice grips, these studs will keep your head bolted down, and any damage may cause them to fail.
crank cases cleaned with studs removed
Cases cleaned and studs removed

The crankcases will need the crankcase mouths increased to 91.00 mm, to accept the larger ductile iron cylinder liners. This operation is done on the CNC mill at my friends workshop. We set the Heiden big fin cylinder up on the crank-cases, measure very accurately the cylinder bore centres, then remove the cylinder and bore the cases. Doing it this way ensures that the liners will be exactly central in the new crank-case holes, something that they are often not from the factory.

Cases being bored
Cases being bored

As a precaution against the cases cracking, I send them away to be shot-peened. This process utilises steel shot projected at a measured velocity against the work piece. The shot compacts the surface, and decreases the possibility of cracks forming, which always start at the surface of a material. 

Shot peened cases
Shot peened cases

The weak point with the cylinders when an 840 or 880cc bore is considered, is the proximity of the dowels holes to the cylinder bore, these big bore engines can very occasionally blow a head gasket. To prevent this I machine up inserts which are welded into the dowel holes and then re-drilled at 10.5mm to give as much gasket surface as possible. This has proved to be beneficial in our race engine; we have not blown a head gasket in a whole race season. 

Dowel hole welding
Dowel hole welded in cylinder

The next operation is to swing the bare Heiden cylinder in the lathe on a specially made Smedspeed tool, and then bore the cylinder out to accept the larger liners. The alloy cylinder is initially bored 0.010’’ undersize, the cylinder is then precision honed to give a high surface finish, and a 0.005’’ interference fit. These steps ensure the best heat transfer possible between the ductile iron liner and the aluminium cylinder. After boring to a piston clearance of 0.003’’-- 0.0035’’ to suit the Wiseco pistons, the cylinder top is surface ground to ensure complete flatness.
Wiseco piston and cylinder
Wiseco piston and cylinder

The finished work on the cylinder can be seen below. The original dowel holes are blocked, welded, and then re drilled. The dowel holes are moved to the outer two 8mm bolt locations. The dowels are readily available Yamaha outer case dowels, Part number 91810-08016 (9.8-12-16 mm)
On the basis that if a little is good, then a lot must be much better, I decided to do the same thing to the head, not just the dowels holes but all eight bolt holes. It's not a big deal to do actually. I set the head up on the mill, centred a 13.5 mm drill on each head-bolt hole, and then drilled to a depth of 7 mm, followed by a 14mm reamer to a depth of 8mm, then pressed in a slightly oversize 14mm plug. The 12mm diameter dowel location in the head can be seen on the bolts that are located under the spark plugs; the depth of cut is 5mm from the head surface.

Bolt holes plugged
Head bolt holes plugged

After the plugs are welded, the head is skimmed, then the cylinder is bolted to the head, and a 10.5mm drill centred in the the cylinder bolt holes to drill the plugs.

The crankcase to cylinder dowels need ‘reducing’ for the cylinders I supply, the liner within the cylinder is 3 mm thick, whilst outside, the spigot, is only 1.75 mm thick, the picture shows what must be done to the dowels to stop then cracking the liner.
Cylinder dowels
Modified 880 cylinder dowel

Squish clearance

The 880cc engines use a Wiseco 87mm piston, the engine will run very high compression, and to prevent detonation a couple of relatively simple things must be done to prevent his. The first is to set the squish clearance properly. The gap between the outer edge of the piston crown and the corresponding area of the head is referred too as the squish or quench gap. This must be measured and modified either by using a thinner head gasket or by surface grinding the head to give a clearance of 0.030’’. The cylinder will actually grow slightly when it is at operating temperature, the long studs are in effect springs keeping the whole lot together. The Wiseco piston shown has been ceramic coated, this reduces the heat transfer to the piston allowing a tighter running clearance, and therefore less wear. It also reduces oil temperatures a little.

Wiseco piston
Ceramic coated Wiseco piston crown

The reduced piston to head clearance will actually lift the compression a little further, however the beneficial squish effect will add considerable turbulence created by the tight clearance and this will actually reduce detonation and increase the burn of the fuel air mixture. The ignition timing can then be reduced from the stock 38 degrees to around 33-34 degrees (the actual figure will be found by dyno tuning).

The 880cc engine must use as a bare minimum NGK BP8ES spark plugs or the fine wire equivalents.

The 880 kit can only be used with a 533 crank. This crank was used on European only XS650 engines, it has a larger 29mm crankpin, and the connecting rods have 140mm centre. This gives a rod to stroke ratio of 1.89:1 which reduces piston side thrust loads, and subsequent wear rates. Ensure you have a 533 crank; the 447 and 256 cranks will not work with the 880 conversion. If your crank is in excellent condition it can be re-used, but if it is in any way suspect then replacement of the worn parts is recommended for long engine life, these engines are very high torque and will “spit out” any suspect parts.
Knackered crank
Knackered crank

The crank above was in a shocking state when I had it given to me. I purposely picked a crank that to all intents and purposes looked trashed, and to illustrate what can be done. I disassembled it, masked off the bearing areas and then blasted it with chilled iron to get the rust off. Amazing what throwing some blasting media at a rusted crank can do!
It has new main bearings (4) and new connecting rods; it has also been dynamically balanced to 53% balance factor to suit the Wiseco pistons. Rephased engines are very smooth compared to stock engines, and the heavier in mass the replacement piston, the nearer it is to the correct balance factor it gets without balancing work.

Rebuilt crank
Rebuilt 880cc 277 degree rephased crank

Dynamically balancing the XS650 crank is a long winded and expensive option. The crank has to be built without the con-rods, trued and then taken along with all the reciprocating items to the specialist balancing firm. They balance the crank weighing all the parts and correcting any imbalance to the chosen factor. The crank has then to be disassembled, and rebuilt this time with the connecting rods, then trued and then TIG welded. It's a lot of work, but I wanted to see how smooth one was compared to the unbalanced option.

Balancing work on crank
Dynamic balancing work on crank

The gearbox can have a hard time on a big bore or bored and stroked XS engine, the sidecar boys regularly get the engines to spit out the 3rd gear mainshaft. I took this one out of a pile of spares I bought, it might have been raced, who knows.

The Wiseco pistons provide for a compression ratio of 10.5:1. This is really on the upper limit for good fuel without detonation occurring, if detonation occurs it will ruin even the best built engine. As a precaution, and to make it a little easier to start and to get some low valve lift air-flow, I remove all the rough edges from the combustion chamber, Yamaha specified this in one of there many racing tips sheets. It drops the compression a little which will only help. I kept the ports essentially the same shape as stock, as I want to keep a lot of velocity in the ports, I just smoothed out the rough bits. I  have written a whole lengthy article on gearbox revisions, it might be worth doing if you ride really hard, otherwise I wouldn't bother unless you just want to.

The assembly is fairly simple, I use a mild cam, it's a copy of a Shell #1 with a bit of minor tweaking to make it less tappety. These larger displacement engines tend to like less rpm,( max rpm 7000) it will keep them together longer, and make them more fun too ride. I have seen a fair few dyno sheets from these engines, and 65-68 RWHP is easily attainable. Thats not much of double the stock output !

The 880cc conversion for the XS650 has been about for many years, we did a few when I was in my twenties and that was quite a while ago. The procedure then was to use Yamaha XT500 pistons and bore the existing cylinders, and fit some sort of adapted car cylinder liner. It's OK I suppose.... just about, but it's not really right. What I have tried to show here is how with modern parts and processes this conversion can be improved a little. This is not an exhaustive guide to building one these engines by any means, it's just a heads up for things to look out for or consider. I couldn't find any information when I considered doing this, so I thought quick, get it down on paper, it might just help someone.

Tel: 0208 979 6972 or email me at