Deceleration Noise

The term “deceleration noise” is an oversimplification of the problem. The noise is sometimes apparent while running under light load on flat, level ground. This is more of a “neutral throttle” condition.

I was pretty well convinced the noise is coming from the clutch assembly. The adjacent audio clip was recorded in a large workshop that produced a bit of an echo.

I'll start with the exploded view of the clutch assembly to have a point of reference. The clutch hub is held in place by a screw and washer (item #10). This fastening method is also used on the GasGas Pro models — in fact, the entire clutch assembly looks quite similar.

Dragonfly Decel Noise.mp3

Exploded view from Mecatecno parts list

The Problem

When the clutch basket (item #1) is fully inboard, its teeth align with the motor pinion gear's teeth.

However, the clutch basket is free to move approximately 0.07 inch (1.8 mm) outboard until it is constrained by the clutch hub (#2). The condition of maximum outboard movement is shown by the yellow circled area in the photo below.

This movement is apparent when the clutch plates, pressure plate, spring, etcetera are not installed. When I run the motor in this condition, I can see the basket move outboard during acceleration and move inboard during deceleration.

I think the noise is produced by the clutch basket rattling in and out (possibly because something changes in the gear loading during acceleration versus deceleration?). Curiously, I could not see anywhere that looked as though two metal surfaces have been bumping into one another.

I don't know what, if any, effect/relation the marks on the gear teeth (red circled area) may have on the problem. But I don't recall seeing that on any other clutch basket, and thought it was worth noting.

Axial movement of clutch basket, shown fully outboard

Wear marks typical on every primary gear tooth

History

In August 2025, the former US importer reported that upgraded gear sets would be available shortly to address the noise issue. Sadly, he died soon after our conversation. In January 2026, the new US importer said that the upgraded primary gear and clutch basket would cost around $700 (after the then 50% import tariff) and only provided a 25% reduction in noise. That did not sound very appealing for something I had the impression would be a warranty fix.

Supposedly, the noise improves as the gears wear in. Different gearbox fluids were also said to yield an improvement. My preferred oil (based on clutch feel) is Dexron VI ATF. I tried a GL-5 gear lube without any audible improvement.

I then started looking for a solution on my own.

How Has This Been Overlooked?

With the clutch fully assembled, axial movement of the clutch basket is greatly reduced. But some movement may still be observed by gently prying behind the clutch basket. It would seem the clutch spring's force must be overcome to observe the movement.

With only items #6 and #7 (spring and fixing ring) removed, full axial movement may be observed.

An Experiment

To test my rattling theory, I machined a 6061 aluminum spacer (shown below) that fits between the basket and hub. The intention was to constrain the amount of movement to around 0.1 mm. (Note that it's possible for the spacer to be too thick, which will cause the hub to turn with the basket even when no clutch plates are installed.)

The noise was greatly diminished during deceleration with the bike stationary on a stand.

It occurred to me that the frequency of the noise should increase (as the distance moved decreases, the time between strikes should become shorter) but that theory is difficult to test.

I also learned that the fixing screw and washer (item #10) has a possibly unintended effect. When the screw is tightened, the washer deforms because it is able to pull in towards the shaft. This may be an important clue to solving the problem. I machined a 2.5 mm thick washer to be used in place of the deformable OE part. I added a 0.4 mm shim at the end of the shaft to make it even with the clutch hub.

Without spacer

With aluminum spacer (44.3 mm x 36.18 mm x 5.4 mm)

Modifications in place

0.4 mm shim under 2.5 mm thick non-deformable washer

12 January 2026 Testing

Due to unseasonably warm weather, I was able to test the bike in my driveway. Tests were conducted with ~42% battery voltage. Oil was 80cc of Dexron VI ATF.

Testing showed substantially reduced noise (I estimate only about one-quarter as loud as without the aluminum spacer).

The noise is only noticeable going downhill. Closed-throttle going uphill, it's quiet. Clutch, engaged or not, has no effect on noise.

The strength of Regen braking is also a factor. Notes above are for Green mode.

In Blue mode it's basically silent. This may explain why some Dragonflys are quieter than others. I also wonder if the amount the clutch has been used affects the noise?

16 January 2026 Modification

I was concerned that if the aluminum spacer should wear due to contact friction, the drain plug's magnet was not going to collect the shavings. So I made a different spacer from Nylon 66 that measures 49 mm x 36.18 mm x 5.2 mm.

49 mm is about as large as the OD can be without impinging on cutouts (potential wear-generating areas) in the clutch basket. The spacer thickness was reduced from 5.4 to 5.2 mm. This allowed removal of the 0.4 mm shim washer.

It's actually fairly difficult to determine / measure the exact thickness needed for the spacer. Making one that's undersize and measuring clearance with feeler gauges turned out to work okay.

Now there is less than 0.1 mm of “rattling” clearance. This configuration has been tested under real-world conditions on multiple outings. So far, it's working well.

Nylon “66” spacer 49 mm x 36.18 mm x 5.2 mm

Results

Both the aluminum and Nylon spacers are being tests in different bikes. As of June 2026 the deceleration noise continues to be greatly reduced and no unintended consequences have been noticed. Testing is ongoing.

Page updated

Google Sites

Report abuse