Ducati.ms - The Ultimate Ducati Forum banner

1 - 3 of 3 Posts

355 Posts
Discussion Starter #1

After being slack for 6 months I finally got around to changing the belts on my 996R this Easter weekend. There is much discussion about the process on this board so I'll quickly add my experiences to the mix.

This was just a straight belt change (2.5 years since last one), the valves were adjusted last time 1500km (950mi) ago so this time I didn't even open the rocker covers. As you might guess I don't get much time to ride my Ducati (it shares that garage with other machines and Monday to Friday I use a Triumph Daytona 955i) - quality over quantity is my motto.

Removal and replacement of the belts was a straight-forward mechanical operation and doesn't require much discussion (in my opinion), however we all seem to spend an inordinate amount of time over the belt tension.

To set the belt tension, I used my O2 Phone (running Windows Mobile 2003 Second Edition) and the application Guitar Tuner Demo (by Four Pockets). This application is limited for it's intended use however it does give a continuous readout of the current speaker frequency - which is ideal for our purposes.

Before disassembly I checked the tension, the belts felt quite loose (and I couldn't get a frequency reading at all). Using the old system, about 6mm of deflection was able to be measure between the cam pulleys.

On reassembly, first I ensured that the belts had been correctly applied (centered) and the pulleys aligned (not jumped a tooth). After applying mild tension, approximated as much as before, I rotated the engine several revolutions stopping with the horizontal cylinder at top dead centre (TDC). NB always rotate the engine forwards if you pass TDC then go around again don't reverse the rotation to align the marks.

When trying to set the tension, I initially struggled to get readings with the phone and just when I was about to give-up, I got it to work reliably. The microphone needed to be held about 5mm from the centre of the belt, on plucking the belt a reading would only stay on the screen for about 1 second before the background noise took over.

My garage is fairly quiet, however using acoustic belt frequency measurement seemed to need near total silence to get reliable readings, so I had to time my readings with passing traffic. The phone was kept insulated from the bike at all time and I held my breath too (got the belt to settle at 108hz).

Once the horizontal belt had been set, I turned the engine to set the vertical belt, similar system - access is slightly more difficult because of the chassis frame, possibly a remote microphone would have been better (still achieved 111hz though).

It is important that the belt being tested has the cylinder at TDC, out of curiosity I checked the tension on the horizontal cylinder (when setting the vertical cylinder) and recorded 121hz. Just note the belt tension changes during the revolution of the cylinder (there is another thread that demonstrates this).

A couple more revolutions of the engine and a recheck of the belts revealed the horizontal belt has settled back to 103hz, another adjustment and re-check.

As a back-up I still tried the old way and noted about 5mm deflection between the cam pulleys. When tensioning the belts, I noted it took about 15 degrees of rotation on the tension adjuster to change the frequency by about 2hz, as the tension neared so the sensitivity of the adjuster got greater.

Once everything checked out, refit all the stuff that had been removed. Start the engine and go for a ride - total time about 3.5 hours.


PS this is my experience, use it as a guide only - remember the belts are the life-blood of your engine.

Flavor of the Month
2,359 Posts
Thanks man. I'll be changing my belts soon on my 748 also. It's due to time and not mileage too unfortuneatly. I'm not looking foward to it, but this helps. The testa head bikes had different hertz for the belts from the desmoquattro correct?
1 - 3 of 3 Posts