Clutch & Gears

Clutch

The diaphragm clutch looks quite conventional, right down to the use of 10 “fingers” (also known as clutch release arms).

Mecatecno's assembly video below shows the installation of an S3 clutch pack.  The plates are the same as those used by GasGas, so their various thickness steel plates would be available for tuning lever effort and travel needed to engage or disengage (and also to compensate for friction plate wear). 

GasGas steel plates are available in the following thicknesses: 1.3, 1.4, 1.5, and 1.6mm.

One notable departure from a standard clutch design is the lack of a large center nut.  Instead, there is a small screw.

I was a bit concerned about the paper gaskets I saw in Mecatecno's assembly video.  They appeared very thin.  But not to worry,  the clutch cover gasket measures 0.75mm and seems adequately robust.  No adhesives are used on the gasket. 

I disassembled the clutch to make measurements for future reference.  The ten M4 screws that hold the spring retainer required a 2.5mm hex key.  I was disappointed to see that one of the ten screw's “drive socket” had been buggered at the factory.  This made removing the screw difficult. 

These screws seem like something that requires proper torquing.  Although I have a variety of 2.5mm hex keys, none could be turned with a torque wrench.  I have ordered Wiha Tools item number 71316 which is described as a 2.5mm hex bit with 3/8" drive.

Overall Gearing

Based on observation (counting teeth) I have determined the following.

Secondary: 10T front sprocket (GasGas spline pattern) drives 44T rear sprocket for a 4.4:1 ratio.  The Iris chain is a 520 with an easily removable master link.  This latter fact is a big deal if you ever need to push or tow the bike a long distance.  Because there is no neutral, removing the chain is the only way to make the bike roll easily.

Primary: 25T motor gear drives a 75T clutch basket gear for a 3.0:1 ratio.

The transmission ratio is 1.78:1.  (This is 3rd gear for a 5-speed Sherco gearbox.)

Thus, the overall ratio is: 4.4 x 3.0 x 1.78 = 23.46

A top speed of 28 mph has been measured via GPS.  This corresponds to a maximum motor speed of about 8500 rpm using a standard trials tire.

It could also explain the US importer's 46 mph top speed statement as a misprint of units.  28 mph is fairly close to 46 km/h.

For comparison, the EM Race (before 2024) had an overall motor speed reduction ratio of 14.39:1 and a measured motor maximum of 8,160 rpm.  This yields a top speed of about 70 km/h (44 mph) with the same tire dimension used for the Dragonfly calculation.

Front Sprocket Issues

I pulled the front sprocket to verify what appeared to be a GasGas spline pattern.  It is.  But what a struggle to remove/reinstall that sprocket!  The swingarm protector/chain guide is hard plastic and part of it extends from the swingarm pivot to the sprocket.  I needed to bend this piece significantly to remove/install the sprocket.  (Later I determined it's easy to just remove the swingarm protector entirely if you want to change the sprocket.)

Then I found it impossible to re-install the sprocket circlip due to the sprocket being too far outboard.  There was a 25 x 30 x 2.5 white plastic spacer inboard of the sprocket.  See adjacent photo.  This spacer is not shown on the gearbox parts diagram. 

So I replaced the 2.5mm thick plastic spacer with a metal one that's 1mm thick.  This allows clearance between the circlip and the sprocket.  With the original spacer in place, the countershaft would need to be pulled outboard forcefully to refit the retaining circlip.

To understand why I substituted a 1-mm spacer, refer to Mecatecno's parts diagram above for the relevant item numbers.  

Upon removal of the cover plate (item 1) I discovered there were three O-rings (item 4) installed rather than the two shown.  Apparently, this was done to move the bushing (item 5) far enough outboard that its end would clear the seal (item 3).  But when attempting to reinstall the sprocket, considerable force was needed to compress the O-rings.  It is interesting to note that the cross-section of each O-ring is 1.5mm (the difference between the original plastic spacer and my steel spacer).

Fettling the Clutch

Upon first receiving the bike, I was curious about the clutch and inspected everything.  The OE fingers appeared to be able to rattle around a lot, and I wondered if they could be contributing to a noise that was present during deceleration.  The clutch release fingers varied in thickness from 0.05925 to 0.05990 (inch).  This is a maximum variation of about 0.016 mm.  I used a friend's surface grinder to bring them all down to the thickness of the thinnest one.  When I was done, the fingers still moved around a lot and there was no improvement in noise.

Another concern I addressed was that one of the studs had not been fully pressed into the hub.  This resulted in great difficulty getting the pressure plate screws installed satisfyingly, as item #7 had to deform to accommodate the error.   In fact, one of the “cheese hardness” screws had been badly buggered at the factory.  I replaced all the screws with higher-quality hardened M4 x 16 button-head screws that have a larger (3 mm) wrenching hex. 

In order to remove the clutch hub, I had to fabricate the simple puller shown below.

Although I don't think it had anything to do with the deceleration noise, I was unimpressed with the quality of the clutch needle bearing (item # 8).  I had never seen one fabricated in that manner, and replaced it with a bearing from an industrial supplier.  

As a troubleshooting exercise, I ran the motor itself with no gears meshing.  It's fairly quiet, but not silent.  There is an RPM band at which the motor is noisier.  This reminded me of the structural resonance noise exhibited by the EM 5.7's motor.

Xiu's Clutch Release Kit

Xiu-rdi Engineering makes a kit for the GasGas Pro models that is said to “stabilize” the clutch fingers (also called release arms).  See: https://www.xiu-rdi.eu/buy/13/-kit-clutch-bearing-plate-assy-clutch-release-arm-set-gas-gas-and-vertigo.html  No one knew whether it would even fit the Dragonfly, but it does fit perfectly.  

It's a nice piece of kit, but did nothing to improve the deceleration noise.