Tom Louwrier of the Netherlands has detailed a potential weak point in the 668 motors when they are used under extreme racing conditions, centered around a harmonic in the balance shaft.  Before you read any further, be aware that the problems with the balance shaft are pretty much all isolated to engines that have been used in racing conditions.  Bikes used on the street are extremely unlikely to be subject the the sort of abuse that is encountered on the race track.

Here's what Tom Louwrier has to say on the subject:-

"The original balance shaft (in the 500 and 600) had a diameter of about 13 or [so] mm's and smaller counterweights. These engines had a 61 mm stroke. The pistons on these engines were lighter as well I guess. When Laverda created the 650 (which is 668 cc's) they had to reduce the balance shaft's diameter because the stroke went 69 mm's and the big ends would foul the shaft. The counterweight was increased to match the new layout.

"Unfortunately, in this setup the balance shaft became unstable. After the first shafts broke on street bikes (after less than 5000 km's!), people started blaming the shaft for being too thin to take the 668's power. This makes no sense, because there is no transfer of power through the shaft.

"Imagine this: the shaft is a pretty long, thin rod with a free rotating weight on one end and a driving gear on the other. The driving torque is not smooth at all, it's a four stroke twin! What you get is a torsional spring (the shaft) that gets pulses from one end (the gear), leaving the other end (the weight) to vibrate freely. If there were power transmitted through this shaft, the situation would have been much better as the torque would be a preload that eliminates the swinging from the free end. But this is not the case.

"When you calculate the own frequency of this spring/weight combination you find that it gets critical around 9500 rev/min. For safety's sake you should always keep at least 20% away from this speed, so that leaves you with a shaft that shouldn't do more than 7600 revs/min. This of course is well within the stock specs' operating range.

"You have probably seen more broken shafts than you'd care to mention. I bet they look similar: the break occurs at the gear's side just past the neck, where the thin part starts. It's a smooth surface with small lines on it (fatigue break) which ends in a rough part (the final violence break). The smooth part has a helix shape of about 45 degrees (torsional load). Mine were all like that, and so were the ones that we saw at the importer's.

"To cure this 'progammed failure', you can do several things:

  • limit the engine at 7500 revs/min    Ha ha bloody ha... 
  • change the shaft's material. It should have a very high resistance to fatigue. 

"Still, this doesn't take care of the problem's source, so you have to move the shaft's own frequency upwards by:

  1. reducing the counterweight's moment of inertia
  2. making the shaft stiffer

"a- can be done by taking weight off of the original part. This does change the engine's balance, which was very good, but it will stay acceptable.  You can take about a third of the weight's weight off on a lathe, turning the 'edge' from the inside out. You'll end up with a counterweight which still has the original outer diameter and thickness but it will have an edge about as thick as the web: around 5 mm's.  If you want, you can change the crank shaft's balance factor as well, but we found it was very good to start with, balancing the primary forces from the pistons.  BTW: the stock pistons are already pretty light weight). We decided to accept the extra horizontal vibrations.  This setup is known to do well in endurance racing (Gijs van Dijk's team uses it. He is good!)

"Another way of doing a- is to make a new counterweight from scratch. This will have to be as small in diameter as possible, while being deeper than the stock part. This way it is possible to keep the original balance factor in the counterweight. I ended up with a weight that is 24,5 mm's deep (total) with the actual half disc solid and flat on either side, 20 mm thick and 65 mm in diameter. This is a more elegant method, especially for road going machines since it doesn't give extra vibes.

"b-is more difficult. There is no room in the crank case for a straightforward thicker shaft. That is why, in the 750 block, the balance shaft was moved forward a little: to get it away from the big ends and create clearance.  If you have access to a machine shop, try this: make an entirely new shaft from tough material (look for fatigue resistance). The ends stay as per original of course. Between the bearings is another matter. Make the necks as shallow as possible (R20 at the gear end and R80 at the free end), ending in parts that are 11,5 mm (original diameter). Now, in the middle there is room for a thicker part, 16,0 in diameter and 100 mm long. This just clears the crank's webs. Make the steps (from 11,5 to 16,0 and back) gradual: about R10. This shaft is NOT stronger than the original (it still has two sections that are 11,5 mm) but it will make the shaft much stiffer, driving the own frequency up a lot. And that is the source of the problem.

"With the original, unmodified weight this raises the critical speed to 10,500 rev/min, including 20% safety factor. With the all new weight you go beyond 11,500 rev/min, The modified stock part should give about the same result, but with a little more vibration.

"It won't be easy to turn this new shaft. You'll probably need a CNC lathe to do it, unless you are/know someone with very good turning skills. The tough material will be a little difficult to work as well. Make sure the shaft's surface is very smooth, especially where the radiuses change. Finish it on a lathe with fine sandpaper or something. It is a very long and thin part, so take care it doesn't 'sing' either, this will ruin the surface. Any scratches or grooves can lead to cracks, shearing and a new shaft failure.

"Once again, I'm sorry I can't send you drawings or calculation right now, but a good engineer will be able to see what I mean from text. Take a set of original parts and then follow the notes looking at them."

An additional comment by Micheal Moore of Euro Spares in the USA on this subject, regarding using the bike in racing conditions:-

"If the balancers don't drive any other bits (like waterpumps etc) I'd ditch them and have the crank rebalanced.  You'll save rotating weight (better acceleration) and eliminate the balancer problem.  You will need to pay attention to oil galleries - when doing this to a 250 Kawasaki Ninja twin I had to make some blanking discs with annular grooves in them to carry oil to go where the balancer shaft was in the case.  If you don't do something like that you've just lost your oil pressure."

Looking at the 668 parts manual, there is a different countershaft and weights assembly for 668 motors after engine serial number 3032.  This is shown as being for the Strike and Legend only, and there is no indication of  why a change was made or if the later parts can be used in earlier motors.  Given that Tombo has been working on an earlier engine, any mods he suggests may be applicable only to engines with a lower serial number than 3032.  It is likely that this change, at engine serial number 3032, coincides with the change made at the factory to the 750 bottom end, which leads to a much stronger motor.  This has not yet been confirmed.

Balance Shaft

© Steve Carr, 2001-2007

Site created February, 2000.  Last modified Friday, May 15, 2009

Return to 668 Tech start page

Last updated
30 May 2005