I`ve changed belts on cars and other bikes (which all happened to have spring belt tensioners which made things easy) but had never done it on a Ducati. The last times the belts were changed it was done by the shop. This time I finally wanted to do it myself so I read quite a lot, watched videos etc. and was confused...
This is not a "how to diy" but just my findings and how I did it.
I`ll explain as simple as I can which might annoy some who know how it`s done but might help the ones who thought about doing it but didn`t find the courage.
Replacing the belts is straight forward. Rotate the engine until the marks on the cam pulleys and the driving (bottom) pulley align with the marks on the engine.
I added my own marks on pulleys and belts so I could copy them on the new belts and be 100% sure the new belts would sit exactly as the old ones without skipping one tooth by mistake. If the belts are going to be reinstalled also mark the cylinder and the running direction.
I bought myself a cheap tool to lock the cam pulley of the rear cylinder. What some don`t know (although it should be obvious) is that all marks that were aligned before mean the front piston is in TDC. The rear piston of course is not. In fact a valve rocker spring is putting pressure on the camshaft rotating it counterclockwise a bit as soon as the belt is removed. (hence the tool which is just stuck with double sided tape) Of course one can just be careful and make sure the pulley mark is aligned again when the belt is installed.
Trying to gather infos about tensioning the belts proved to be really interesting as Ducati used different methods over the years and many of the diy infos you find on the internet are just wrong resulting in a huge amount of engines with incorrect belt tension settings. I`m not saying this because I know but because a lot of the info available is either contradicting or leading to a different result so it can`t all be right.
On our models Ducati used a tensioning tool which was nothing than a spring balance, later there was a more sophisticated tensioning tool and nowadays it`s done by setting a specific resonance frequency.
The first tensioning tool just had a mark on it. According to the SS/Monster Haynes manual the tension would correspond to 4.5kg (10lbs). A german Monster manual would say it corresponds to 1.5-1.8kg...?! (what is it?)
Just out of curiosity I figured I`d just try if setting the tension with a spring balance was possible at all and made a small bracket to fit on the belt tensioner pulley so it`s surface wouldn`t be scratched. The pulley of the horizontal cylinder would have to be pulled upwards which is already a bit tricky lacking space and the rear cylinder pulley would have to be pulled to the rear which is impossible with the engine in the frame so I scrapped this method.
I went on looking for belt tensioning tools and there are many like the Dayco Krikit but for what it seems they are all for belt tensions higher than what`s required on a Ducati (i.e. 15-75kg) and some can only be set to "Newton" which is no help. There is no spec to how many N the Ducati belt should be set. (Newton is force, kg is mass which isn`t the same) Ended up scrapping this idea as well. The tensioning tool would be positioned between the fixed pulley and the cam pulley. (On a 4V engine between the cam pulleys)
For what I read the tool that Ducati uses to measure the resonance frequency does this via an optical sensor. These systems are very expensive so no option.
As alternatives there are of course other methods. One is the so often mentioned allen key method and at last it`s using a microphone and software to set the correct frequency.
To be honest I always doubted the allen key method as I didn`t know in what tension it would result. I started having more doubts when I started reading about belt frequency specs of Ducati engines which were (!) quite different depending on the engine.
A 996 would be set to 150Hz, S4RS: 110Hz, ST3: 142Hz, 900SSie 121Hz (vertical cylinder)/146Hz (horizontal cylinder) etc.
The looser rear cylinder belt tension would be because the rear cylinder would become hotter and expand more tensioning the belt more than the front cylinder but wouldn`t that be the same on all Ducati engines and which one of these settings does the allen key method correspond to?!
Another question would be how strong would one have to move the allen key.
The instruction from "ducatisuite" shows that a 5mm allen key is positioned between the fixed pulley and the belt, then the belt is tensioned so that the allen key can still be passed between pulley and belt. A 6mm allen key shouldn`t pass (I guess it would if one just pulled stronger)
Other instructions I found incl. a video on Youtube clearly show and explain that the allen key should be used between the tensioner pulley and the belt not the fixed pulley.
Personally I don`t like the idea to pull a tool with edges over a running surface.
No doubt though the allen key method is the quickest one.
At last the frequency measuring method. A microphone is held near the belt while its plucked like a guitar string. The looser the tension the lower the read frequency, the higher the tension ... of course there`s lots of info in the internet about this. The issue is that some pages explain to take the measurements between the driving pulley and the tensioner and others between the fixed pulley and the cam pulley. The latter piece of belt is a lot longer than the first so it`s obvious that the readings would be quite different/lower. My assumption is that some writers still have the position of the last Ducati belt tensioning tool in their minds which indeed would be positione there.
What to do with all this info? The more info is gathered the more confusing it gets and harder to figure out what is right or not.
It`s been a long while I`ve been planning to compare some methods to see if they had the same result which is what I did or better attempted to do yesterday.
This is from 2009 I think which means that a lot of the specs that can be found online are obsolete. The list of course doesn not include any Paso or Sport but mentions the ST2 or M620 which makes me conclude it`s the same for older engines as well.Sonic Belt Tension - Ducati update
The following service bulletin has been issued by Ducati North America. Note that the specs are now the same for all models 2V and 4V. There are a few omissions from the list - 748 and ST4s, but I would think that's just an oversight.
All previous and new model Ducati’s requiring belt tensioning now will be standardized to the below specifications.
IMPORTANT: When checking and adjusting timing belt tension at the recommended scheduled maintenance intervals, if tension reading is less than 70 Hz (minimum allowed value on all Ducati motorcycles), reset belt tension as follows:
Refer to the appropriate shop manuals for cam timing procedure and belt tension setup, then reset to the below tensions for service.
Model
Belt tension spec for new belts upon replacement
Belt tension adjustment settings for used belts when inspected value is determined to be 70 Hz or less
All models:
Streetfighter, 1198, 1098, 1098R 999, 998, 996, 916, 848, 749, M1100, MS4, MS4R, MS4RS, ST4, SC, ST3, ST2, SS1000 HYM 1100, MTS 1100,MTS 1000, MS2R 1000, M696, MS2R 800, M620, M400, M695, SS800, MTS 620
New Belts:
110 Hz +/- 5, Horizontal and Vertical *
Used Belts:
90-100 Hz +/- 5, Horizontal and Vertical *
* Please note, these values may be inaccurate on previous Ducati Technical Data posters and service manuals. Data listed above should be considered the new standard and supersedes all previous bulletins and documentation. Please print and post this bulletin to be displayed with the most current Ducati service poster.
Sincerely,
Ducati North America
Service Department
It isn`t mentioned so it can be assumed these settings are with a cold engine.
The official Monster 696 service manual shows were the measurements should be taken. (Measuring between fixed pulley and cam pulley is definitely wrong while measuring between the driving pulley and the fixed pulleys gets about the same results +/- 1 or 2Hz. In fact the vertical cylinder is measured between the driving and fix pulley, while the horizontal is between driving and tensioner pulley)
There are numerous programs that can be used from spectrum analyzing software to programs used to tune guitars. Nice are programs that have a peak hold function meaning that the loudest frequency reading will be shown on the screen also after you`ve plucked the belt.
Via the italian site http://www.desmodromico.com I found a french software tool which can be downloaded here http://jpdiag.akress.com/ . It`s the last file called "Courroies".
This software was specially programmed for tensioning Ducati belts.
It`s very simple. No settings need to be modified. The program seems to contain the old frequency specs so select a model that has a 110Hz tension setting (i.e. 848/1098) or just change the "Marquer" setting by hand to 110. The red line in the graphic indicates that setting.
When the current frequency is measured a yellow line will indicate the peak. On the bottom the program will also show the measured frequency. Then change the position of the belt tensioner pulley until the yellow line is on the red line. 104,5-115,5Hz is the range you should be in. (110Hz +/-5%)
Of course a computer is required and a microphone as well. I got a used Sony 17" laptop off ebay for 30 Euros. Finding a microphone was a bit more difficult as it should be inexpensive and good enough. Many mics sold don`t mention their frequency response and if they do it`s often from 110Hz up so one cannot really expect a precise reading when using one of these. A Sennheiser computer microphone is what I found for €5. (frequency repsonse 40Hz-12.500Hz)
One small detail that might be interesting is that there are numerous different kinds of microphones. There are for instance unidirectional and cardioid (kidney) microphones. The unidirectional record mainly the sound that is directly in front of the microphone and the other also what`s around. What kind of mic it is influences how it is best positioned. Just experiment a bit.
There are apps for smartphones for tuning guitars but I doubt the microphone in the phone is good enough.
The microphone is positioned about 1cm (0,5") from the belt. Ideally if it`s one with a long neck like mine it doesn`t touch the frame unless at the point it`s mounted to as this might influence the reading.
This job is ideally done in a closed and silent garage as every car driving by or a dog barking two houses down the road will show in the graphics and the yellow line might be jumpy.
Click "Start" on the program and pluck the belt in the middle of both pulleys. If done with the thumb while keeping the hand on the engine the peak reading will be the same but there are more disturbing moises registered. Pluck "normally" without force.
There`s no "peak hold" funtion in the program so you need to look at the screen while doing this. It may take a few attempts until you get a steady reading. Plucking the belt in short intervals will help.
My first reading was 141Hz, hence too tight
There are two bolts on the belt tensioner. One is the rotating axle (the easily accessible bolt) and the other the locking bolt which is partly covered by the belt. Undo both a little bit so the belt tensioner can be moved as required. Handtighten the tensioner and leave the locking bolt loose. (If you`d try to get to it you would have to move the belt out of the way and this might move the tensioner). Then try again.
Must have been beginners luck as it varied between 109 and 112Hz which is perfect.
If that is correct tighten the tensioner pulley to spec (P750: 20-24Nm/14,5-18,1lb/ft). (Note that this might move the tensioner outwards/ away from the belt minimally which will decrease the belt tension so recheck tension). After that the pulley locking bolt can be tightened. The frequency measurement should be repeated and again after the engine was turned over by hand a few times.
The process on the rear cylinder belt is the same but one needs to be aware of the piston not being in TDC. If the tension would be set now it would be incorrect. The engine needs to be turned by 270° in its runnig direction/counterclockwise (or 90° clockwise. This is seen from the alternator side btw) to get the rear piston to TDC. (Newer flywheels have TDC marks, I don`t know if the old ones have. Didn`t check)
Now to the comparison.
When I removed the belts which were installed by the local shop and basically new as the engine had hardly run in the past years an 8mm allen key would fit between the belt and fix pulley. This seemed loose but I doubted the shop had done it incorrectly.
Setting the new belt to 110Hz (measured at the point shown in the Monster696 wsm) actually gets to that result. The frequency measured between the fixed (the upper one) and the cam pulley is about half which I had about expected. This section of the belt is about twice as long. If it was set to 110Hz here the belt would be tensioned way too much which imho proves that taking a reading here is plain wrong.
Using the allen key method which describes that a 5mm allen key should still pass between the belt and the tensioner pulley results in a tension which is too low for a new belt but ok for a used one.
The allen key method which says a 5mm allen key should pass between the fixed pulley and the belt but a 6mm not results in a tension that is too high. Looking at the old obsolete tension specs it is within the limits for a 996 which makes me wonder if this method originally came from the 4V fraction.
It says here http://home.comcast.net/~mmullen38/Beltadjust.html that at warm engine the belt tension increases by 7-8% (on a ST4). If that is similar on other engines the belt tension resulting by this method at cold engine already exceeds the tension of the first method with hot engine.
These are my findings. Results may vary (as I only did the horizontal cylinder so far and didn`t remeasure after turning over the engine yet but I don`t expect the results to vary that much).
Belt tension is as much a correct science as is carb jetting. Different settings will work but some will be closer to the "ideal" than others.
Now make your own conclusions...
G.
PS: my engine has adjustable cam pulleys and I was initally wondering after watching the belt change video of a newer Ducati in which was shown that the bolts of the outer ring of the cam pulleys were loosened before setting the belt tension as it is done on many cars. But in these cases there`s a pin that`s screwed in from the side and which holds the cam in position while the ring can be moved counterclockwise. I don`t have that so left them untouched. Cam dialling will be next anyway. Not that this is necessary after a belt change but it`s another thing I haven`t done on the Duc yet