Sunday, March 5, 2017

More Minnie the 400 Monster dyno runs

I took Minnie to the dyno today to see how the 2-1 works.  Not well as it happens - red is the 2-1 curve, blue the previous session with the std headers and Megacycle mufflers.

Pretty much dyno'd how it feels on the road - goes well under 8, no point going over.  I was going to set up one of the previously dyno'd Megacycle muflers to give just the change due to the 2-1 header, but didn't get to it.  I hadn’t revved it hard before the 2-1 went on.  Now that I have a tacho fitted to know how hard I am revving it, I can see that it would have been very unlikely that I was going over 8 anyway.  As such, I really didn't notice the now crap (ish) top end in comparison to the previous set up.  Maybe I need to make a twin muffler mounting bracket for under the tail and cut down the second Danmoto muffler to fit with original header.

I don't have any air/fuel for the original exhaust and Megacycle mufflers, as the bike killed the dyno on those trips.  The rpm and air/fuel were shutting down when we tried to run it.  After the last session I fitted resistor spark plugs, which you need to run with the Ignitech and which all these carby SS bikes (should have) had oem.  I assume that’s what made the dyno control happy.

First runs were baffle in and then baffle out.  I modified one of the Danmoto baffles I had.  It has 12 holes around it, and the total area of the holes was about the same as the muffler internal diameter anyway.  I machined the end cap of the muffler just a little to allow the baffle to slide in, and them machined a circlip groove to hold it in.  Of course, I didn’t have any circlips big enough to start with, so I bent up a round clip from coat hanger wire.  Gotta love coat hanger wire.  As below – without, machining, and with.

Made so little difference you’d have to call it none.  Then I found that the Showa fork spring preload tubes are a perfect fit over the Danmoto baffle, so I slid a tube over, tacked it on then cut off the original end cap.  The internal diameter of the baffle exit is now probably a bit under 20mm.  Again, not a lot of difference, but I ran with it. 

You can hear the slight change in these clips I took in the dyno room.  Not great sound, but definitely quieter with baffle from the bottom.

As expected, it hurt the bottom end a little.  Not as much as I’d hurt the top end though.  Baffling the mufflers usually effects the peak torque more than the peak power, but in this instance that’s a bit hard to see.

It was a bit (fair bit) on the rich side, so I thought I'd try it out without the airbox lid.

It was a bit better power wise, but the mixture change was pretty extreme.  More than I think I've seen before on a 2v motor.  Too rich to more too lean.

Then I tried it with lid, but without the snorkels.  That helped a bit, but still too lean on the needle.

I have some of the original 400ss carb springs, that are a lot lighter than every other OEM Mikuni spring I’ve seen them fitted with.  I might try them, not sure if that'll make it better due to the increased slide opening lifting the needle more, or worse because it'll have less vacuum over the jet and the needles aren't tapered enough anyway.  I do have some jet kit needles that might be the go.  Probably with springs as well - you’d just call that fitting a jet kit.
I rode it back from the dyno with one snorkel fitted and without the baffle.  It felt a bit stronger without it, but the missing snorkel may also have been helping there.

The Ignitech unit I fitted before this dyno session has the same advance as the std boxes.  Well (you know when you make assumptions and then later you think "oh"), maybe I should check that - it should have the same advance.  
I know that from the previous 750 engine testing that lack of advance hurts the top end and can give an exaggerated choppiness to the curve.

This early Ignitech TCIP4 has a delay that increases as rpm climbs.  The newer units (this one I've had since 2005 or so) don't.  I thought I had allowed for it in my curve, but maybe not enough.  I did have the oem boxes with me at the dyno, but my time was pretty much up there and I had to get back to work anyway so I didn't.  One thing I hate about not having my own dyno running is not being able to act on all the things you think of later when it's too late.

So maybe it didn't have enough top end advance, and it was rich anyway.  
Fairly extreme fuelling (well, air entry affecting fuelling) changes didn't make a big difference either, although it did go from too rich to too lean.  The last run I did had the mixture back at 13.5 or so in the crap top end area, so that pretty much rules mixture out.

Although I ran these same carbs years ago when it was a std 600, and I would have thought with a smaller engine it should be leaner due to less suck.  For some of the production years the M400 and M600 are fitted with the same carb spec (same part number).  I did shoot some video through the airbox lid without an air filter fitted of the slide, and it did look like it was lifting all the way up.  I was wondering if it would, given it's a pretty big carb for a 400.  Lots of fuel flying around too.

Through the snorkel hole

As it sits at work now (waiting for a nice day and time to go for a ride, it's pretty hot here now) it has one of the Megacycle mufflers fitted to rule that change out, and I'll put some oem ignition units in my pocket and fit them on the road. I really can't see the exhaust alone hurting it this much, as nothing I've done exhaust wise in the past has, but I'm often wrong.  And I do have concerns that the 2-1 merge isn't deep enough into the V of the original connector.  It may actually allow flow from one cylinder to go back up the pipe toward the other cylinder. Not so bad from the vertical as the pipe in is angled, but from the horizontal it can.  That's possibly an issue and not one I can do a lot about with this header set, especially now they’re ceramic coated.

Tuesday, February 14, 2017

Some more Gates belt stuff

One of the things that surprises me when I write posts are the reactions I sometimes get, specifically the questions that often leave me wondering how people interpret or comprehend what I've written.  It makes me realise how hard it can be to be clear and get what I'm trying to say across.  Compounded by my tendency to ramble and digress and straight up forget what I was rabbiting on about in the first place.  So I'll put in some more belt info, and try to respond to some of the comments and questions.

Another thing is that (especially with cam timing) people ask me for "the good numbers", as if there's some sort of secret squirrel component to what I write.  There's not.  What you see is what you get.

Anyway, belts.

The old Gates belts were available for all the Ducati models pre 98 external to Ducati via Gates distributors.  The three part numbers were "cancelled" in Australia last year, not sure about the rest of the world.  They were all differing construction types, which I found a bit interesting.

T934: the 900 round tooth belt - 900SS, M900, 906, 907, ST2.  1st generation construction, NEO - Neoprene rubber fibreglass tensile cords and nylon tooth facing.  Interesting that the belt introduced last was the most basic construction.  Never gave any issues.

T819: the Pantah square tooth belt - all small blocks (non 900) up to 1997.  2nd generation construction, HTN - High temperature neoprene rubber, fibreglass cords and nylon tooth facing.  Never gave any issues.  When the Pantah was first released, they had Pirelli belts which were rubbish I'm told by people who worked on them.  And gone before my time.  I have also seen Continental belts fitted to these motors.

T917: the 4V belt, also fitted to Renault R19, which ran from 89 - 96 according to the Gates book.  3rd generation construction, HSN - Highly saturated nitrile rubber with aramid fibre loading, aramid or high tenacity fibreglass cords and nylon tooth facing.  The problem belt (from 98 onwards at least).  The interesting thing is the aramid part.  Aramid is a generic name for what is better known by name brands such as Nomex and Kevlar (both owned by DuPont).  So these belts it appears had a Kevlar component, which was allegedly the big point of the red letter belts.

So to clarify, the only Ducati belt made by Gates that gave any issue was the T917 - available via Ducati or Gates.  And the issues only appeared as a consistant problem in 1998, by which time Ducati seemed to have realised there was an issue as the "2 year replacement" directive was communicated to the importer (ours at least) before we had experienced it.

Why there was an issue I don't know.  Lots of theories were expounded, but I like facts backed up with evidence and there didn't seem to be enough of that to generate any conclusions.  Lots expired down the front straight at Philip Island, but lots didn't.  They just broke on bikes that were more than 2 years old, but had done less than 20,000km.  We also saw belts break that had been fitted at 20,000km services, but not replaced again within the 2 year window.  Yes, the pullies are quite small compared to most automotive applications.  And, on the 4V, there are quite extreme changes in direction compared to what you see in most automotive design.  But, anecdotal evidence of belt life before and after, especially with the later belts, would indicate that the actual engine design is not "the" issue. 

The Red Letter Belts, or whatever you wish to call them, were introduced in 1999 or so - I don't actually recall now.  The main claim to them was that they contained Kevlar.  All still made by Gates and marked as made in the UK, but not available outside of Ducati.  Some were a bit more expensive, the Testastretta belts a lot so.

Whether or not all the red letter belts actually contain Kevlar, I have no idea.  If you can find someone at Gates who does know, pass it on.

As to the two year replacement interval thing: Firstly, I have no interest in arguing this in the slightest.  Absolutely none.  Believe it.  Don't believe it.  I don't give a rat's arse.  But every official Ducati workshop manual from 2000 until the MTS1200 has had the two year replacement regardless of km specified.  At the bottom of the scheduled maintenance chart or on the next page, it will have the note corresponding to the (*) or (1) after the Timing belt line saying "Replace every two years, in any case."  With the MTS1200 and Diavel it became 5 years.  As I'm not part of the dealer network, I didn't get the official explanation as to why.

As to how often I change my belts: I have only had one belt drive bike in use long enough to require belt changes, but then again it's also on its third engine in its time with me and there's been a bit of idle time.  I think only the 750 engine has had belts in it long enough to need them to be replaced, and I only did that when I realised (after noticing the dates on dyno runs) they were over 5 years old.  I'd guess the belts in the 851 are getting on for 15 years old now, but it hasn't run in 12 years so that's no big deal.  Everything else has been moved on before time.  But I'm a tight arse too, so any belts in my bikes will be in there for as long as I can withstand the niggling paranoia of impending doom.

I've not used any of the Dayco belts.  As the Australian distributor for California Cycleworks products, I use the Exactfit belts as the aftermarket option.

Saturday, February 11, 2017

I've gone bigger font. Someone's eyes aren't what they used to be.

Ducati timing belt failures and replacement intervals

I thought I’d do a piece about belts as there’s a thread going on MS at the moment with responses from people who weren’t there when the shit went down and so don’t understand why.  I’ll try to clarify it from where I saw it.

The first reference to replacing timing belts based on time as well as km that I can find in any of the scheduled maintenance charts in the factory workshop manuals is 1999 996.  For the 1999 MY I also have ST4, M900 and 750SSie manuals, and none of those have it.  All the 2000 MY manuals have an * after “Timing Belt” on the chart and “(*) Replace every two years, in any case” at the bottom of the chart.

As to whether there was a service bulletin about the two year thing, I don’t know.  I have quite a lot of Ducati service bulletins, some back to 1990 or so, but nothing on belt change interval, nor a listing for it in the index.

The first broken belt I saw was in 1995 or so, a 900SS that came in for an opinion to be used in a court case against another dealership.  Typically of the very few 900 that I have seen with broken belts, it also had a loose vertical cam pulley.  As in flogged out and flopping around.  It went again without us touching it.

I think I saw one Pantah with a broken belt.  Asking the owner when they were last changed just drew a blank look.  It got a new belt and was fine, one of the few.

I recall Rob doing a head that was bought in, probably in late 1997 or so as I recall it being in the then new engine room, from an 888 that had broken a belt.  Didn’t know anything else about it though.

The first broken belt we were really involved with was a 916 owned by a friend of Rob’s.  Sold mid 1996, with the 10,000km service done mid 1998 around two years old.  It broke a belt a few months later at a Moto One Broadford track day.  We got it back to work on the Monday after, pulled it apart and found the horizontal (almost always the horizontal) belt broken.  You can tell if the valves are bent by trying to rotate the cams.  If the valves are bent, they pull the closing rockers down and jam them against the cams, thereby stopping them.  Simple check.  I recall one 916 coming in that we fitted new belts too and away it went.  Only one that lucky.

We rang the importers and got asked “why weren’t the belts changed at two years old” and we asked “why do they need to be”.  Then the typical “well, it’s all your fault because .....” that we heard quite a few times from importers over the years.  I don’t recall ever hearing anything about two years before that day, nor seeing anything from Ducati (service bulletin, etc) saying it, but the message from the importer was very clear.  Two years, no more, no help because the warranty’s up.

That was the start of it.  Probably averaged a couple a month for the next year at least.  748 and 916, 10,000 – 15,000km, 2 ½ to 3 years old.  Owners would come in wailing and wanting warranty, us explaining the Ducati line and the fact it’s out of warranty, us getting abused because somehow the mechanics at a dealership half way around the word from the factory are responsible for any design shortcomings.  All that stuff.  We were the preferred warranty dealership in Melbourne it seemed.  When the warranty claim rejections were faxed back and the owners informed, the bikes got picked up and went elsewhere to be fixed because we were too expensive, arseholes, etc.

Of course, while the message from the importer was simple, their own dealership compounded the issue.  We harassed every owner who came in with a bike from then on to replace the belts at two years old.  Everyone.  But we had customers who would ring up Fraser Motorcycles in Sydney and ask their workshop and they’d get told “we just look at them, and if they look ok, we put them back in”.  Then they’d come back in to us and tell us again we were arseholes trying to rip them off, etc.  The usual, with a new slant.

So we started putting “Timing belts not replaced at owner’s direction” on invoices.  Owners would see that (we’d usually point it out to them) and realise we were serious and if it did go bad it was their problem.  Then they’d get the shits because we didn’t tell them enough to convince them that they really did need to be replaced.

I can recall still getting phone calls into 2001 or so from owners at Philip Island track days.   One guy, who I knew we hadn’t seen since his first service in 1996 or so, had his 916 drop a cylinder down the front straight.  Turns out his bike had only done around 6,000km and probably hadn’t been serviced since the first service because there was no annual service regime back when it was new.  I gave him the usual advice: remove the belt covers and check the belts, the horizontal one is probably broken.  Unfortunately correct.

I saw one 851SP3 do it, and again it was a case of the owner being told, when he bought the bike second hand in 1996, that it was good to go until 20,000km.  I think I saw one 916SPS break a belt, I’m sure I had a 1998 model apart for that reason (no Ti rods!) and one 916 that did it within the warranty and belt life period, but there were some other potential concerns about that bike anyway.

I don’t recall a 996 doing it, and I think that’s because the new “red writing” belts with Kevlar in them were introduced with or around the time of the 996 in 1999.  I forget now.  And we were harassing everyone to replace belts at two years, so none of the bikes we serviced would have gone much past two years anyway.

In hindsight, the whole thing was handled poorly from an administrative sense.  BMW (who we were also dealers for, and who had time based service schedules) would’ve sent letters to all the owners pointing out the two year deal.  Of course, a lot of them would’ve been ignored until the bang came, but at least they had been warned.

The belt that caused all the issues was the Gates T917, which was originally a Renault belt.  Why it all started going bad in 1998 I don’t know.  Maybe it was the simple increase in sample size, with a lot more 4V bike sales, coupled with the increase in track day participation.  Nothing else changed in terms of engine design or layout.  Once that belt was out of the official spare parts system, the problem was effectively fixed.  You could still buy them via Gates, as we did for the Pantah and 900 belts, but we mainly kept away from the T917.  All the Gates parts numbers for Ducati timing belts were cancelled last year (in Australia at least), so now they’re gone for good.

There wasn’t an issue with 2 valve bikes simply due to a bit too old belts.  Very old belts yes, but not like the 4V.  I have had my own 2V bikes go 5+ years simply due to forgetfulness while still in use.  Nor is there an issue with the later models.  I have seen the occasional broken belt since.  I had a 1098 come in that had nothing but fluff on the horizontal and strands with teeth on the vertical, but it had at least one seized roller and I suspect hadn’t had a belt change in its 52,000km life.  The damage on that one was quite amazing.  The valves were Z’d (bent above and below the guides), and I couldn’t get one of the inlets out.  The inlet cam had flats on the noses so beautifully precise it looked like it’d been held in a fixture and stamped with a machine, with the material flared out both sides evenly.  The piston was also damaged, with a crack through the skirt into the pin boss.  Probably the first time I’ve had to do anything with a piston when fixing valves bent due to a broken belt.

But, due to professional paranoia and liability, we hassled everyone to replace belts every two years on all models, as Ducati directed.

Now they have moved the replacement interval out to 5 years on new models.  The MTS1200 and Diavel use the same belt as the 1198, etc.  Maybe, in the main, the SBK models get used a bit harder than the tourers rpm wise, but some of the MTS1200 get used pretty hard and often.

My question about that was, given the warranty is two years, what happens if a belt fails outside the warranty, but within the service time frame.  The answer to that, in the current Audi ownership model, is simply that someone other than Ducati will be paying for it.

I recommend 4 years at least with the genuine and California Cycleworks belts.  I don’t go over the mileage intervals, but again they’re being stretched as time goes on too.  What was only 20,000km became 24,000km and is now 30,000km, with no checks or re-tensions in between.  We used to re-tension belts every service, now you don’t look at them until the replacement time.  And with the massive increase in the amount of time required to just access the belts on some models – MTS1200, Diavel, even the M696 – 1100 series, it’s not easy to have a quick look.

Nowadays lots of owners who didn’t experience any of the belt drama back in the day question the two year thing, and many like to claim it’s a money making exercise, etc.  But the basis for it was a real problem for some models, and cost many people quite a bit more than the price of a set of belts.

Sunday, November 27, 2016

More Minnie Monster photos with her custom look.

I took some photos of her in front of the factory.  I'm really liking the look.  Still haven't fitted a dash - the Acewell 6554 has arrived, but with no pressing need to get it fitted, she's once again back to waiting for the dreaded time.


Friday, November 18, 2016

Diagnosing bike not starting - switches and being inputs to the ecu

I had a fellow trailer a Monster S4 in this week because it wouldn't start.  With it on the side stand I turned on the key and the neutral light didn't come on, even though it was in neutral.  So I played around with the gear lever a bit, got the light on, hit the button and away it went.

The owner stood there with a rather dumbfounded look on his face, clearly having enjoyed the two hour drive leading up to looking like a dill.  But understanding why it worked for me and not him involved a concept that he had a bit of trouble understanding at first.

Back in the day, a neutral light was just a thing that told you the bike was in neutral, powered by a wire from the neutral switch.  Simple enough, and likewise obvious.  On most modern bikes though, it is primarily an input into the ecu.  The ecu uses it for the starter and sidestand logic circuits and also to turn the dash light on.  As such, if the dash light's not on, that's because the ecu doesn't know it's in neutral and therefore thinks it's in gear.  An obvious enough assumption based on the info provided.

When you press the start button (itself an input to the ecu) the ecu checks all the start/sidestand logic circuit inputs (neutral switch, clutch switch, sidestand switch) before deciding if it should activate the starter.  It does that by grounding the earth side of the starter solenoid.

Generally, the starter solenoid is activated under the following conditions:
  • In neutral
  • In gear with clutch disengaged (lever pulled in activating switch) and sidestand up
Although this Monster S4 (2002 MY) would start and run with the stand down and neutral light not on as long as the clutch lever was pulled in activating the clutch switch.  When you let the clutch lever out, it would stop.  Most things will stop as soon as you put them into gear if the engine is running and the stand down, or when you lower the stand with the engine running and in gear, even though the clutch lever is pulled in.

Anyway, digressing.  So, to expand on this, the reason this bike wouldn't start is because the neutral light wasn't on.  If I hadn't been able to get it to show, the next thing I would have done is pulled the clutch lever in (to the bars) and tried again.

Which leads to a possible total killer of starting desire - faulty neutral and clutch switches.  Then it definitely won't go, no matter what you try.

To get back to the fella with the perplexed look.  He said several times "but it was in neutral" to which I responded "was the light on?"  The answer to that was "no, but we know it was in neutral."  And that, folks, is the crux of the issue.  It doesn't matter what you think is happening, it matters what the ecu thinks is happening.

This is why my first response to people asking me questions about poor running injected bikes is "get it on a diagnostic tool and see what the ecu thinks is happening."  The ecu is running the show, so its opinion is the one you need to consult.  When you have a tool, it's a quick check.  Logged faults and plausibility of inputs is what you're looking at.  If it's all fine, move on to the next thing.  The diagnostic tool is probably not going to tell you what is wrong, but it will let you see what is right.

Monday, November 14, 2016

A makeover for Minnie Monster

I decided to do a 3 week rush job on Minnie for this year's Festival Of Italian Motorcycles, which was last Sunday.  I ran a thread on the forum about it, and am way too lazy to redo it here.  Maybe I'll do something here later - I still have to buy and fit a real dashboard.  The iPhone with speedo app isn't long term workable.

A makeover for Minnie Monster


Tuesday, October 11, 2016

Ducati ST3 front sprocket incorrect position take 2

See this post for more background: Ducati ST3 front sprocket

I had another ST3 in this week with the same issue, so took some more photos.

Spacers: ST3 (999, etc) on left, 1000SS on right.  19.7mm vs 14.2mm (or something like that, forgotten already).

Sprocket fitted with incorrect spacer.

Sprocket fitted with correct spacer.

Saturday, August 27, 2016

Part 2 of Dyno run comparison of things not the same - dunno, couldn't think of what to call it, it's a bit odd

I was curious to see how the torque curves of the std comp motors compared, so I made up a spreadsheet generated graph using exported data form the Dynojet dyno software to compare the old dyno to new dyno runs I did of the high comped 750.  Using the factor between those two runs (average of 0.882, 0.860 - 0.912) I calculated a std comp run for the big valved, 900 cam 750, albeit with std carbs for the new dyno.  I also took the dip out of the previous 750 run, just to make it look nicer.

The result is the graph below.  Black is 900 with FCR39, slip on muffers and open airbox lid.  Red is std comp 750 with jet kit, slip on muffers, open airbox lid, 900 cams and big valves.  Blue is high comp 750 with FCR39, slip on muffers, open airbox lid, 900 cams and big valves.  The yellow line is the comparison between the two std comp torque curves, as % difference of 750 from 900.  The 900 engine is 20% bigger than the 750, meaning an ideal situation should show this as a straight line at 20.  But with the reality of engines being engines, you don't get that.  Interesting that the centre part of the curve is just under 20% (17 - 19), which is as expected and it's always nice to see some correlation to the theory.  Once the 900 torque peaks just under 6,000 rpm the factor drops as the 750 both peaks later and doesn't suffer the same rate of drop off after the peak.

Thinking about it some more just now, the 750 was running the cam timing at 107 degree inlet centrelines, but the 900 was "as delivered".  Meaning not checked.  At a later date I did check the 900, and found both cams at 117 degree inlet centreline.  They were advanced to 107 degrees with a couple of 10 degree offset keys.  At 107 the 900's torque curve would drop away even quicker, possibly to less than the 750 at the end.

And there you go.  I found it somewhat interesting, hope you did too.

Tuesday, August 23, 2016

Dyno run comparison of things not the same - dunno, couldn't think of what to call it, it's a bit odd

I was going through some dyno runs yesterday looking for some 750 carb jetting info, and looking for a comparison when I came up with this one.

Looks like two runs with different jetting?  Not, as it happens.  The blue is a 900SS fitted with FCR39, open airbox lid, some sort of slip on muffler and no other mods.  The red is my hotted up 750 motor when fitted to Minnie.  Same cams as the 900, similar valve sizes, different FCR39 (you'd hope very much the same) and the exact same 152 main jets.  I have a bit of a running joke going with this pair of FCR 152 main jets.  I fit them to everything I do with FCR and see how they compare.  So far they've never kept a permanent gig.  Anyway, two things come to mind.

1/ Massive difference in fuelling where the only real difference is engine capacity, and therefore airspeed through the carbs.  The 750 has lower airflow for the same time frame (same cams), so less vacuum drop across the venturi (Bernoulli's equation) and therefore sucking out less fuel.

2/ Almost identical power curves run against road speed, with the red run stopping earlier appearing to be just a "as run" thing.  For the gearing to match it up this way is pretty neat, and shows that, on a short straight or in this gear at least (possibly 4th on both, 900 has two more to go) it'd be a fair fight to the next corner.

The next graph shows the power versus RPM, which shows them to be different.  As you would expect, the 900 will make more power at any given RPM under the power peak as it is larger, but also points out how good the 750 is comparatively.  

Comparatively is better shown in the torque curve, as below.  The 900 engine is 20% larger than the 750, so it should make 20% more torque all other things being equal.  In this case, the big variable is compression, as the difference in peak torque is only 8%.  The 900 has a bit over 9:1, the 750 allegedly around 12:1 (I never did check).  The runs I ran with the 750 as a std compression motor were done on a previous dyno which gave higher readings, so I don't have a std comp run that directly compares.  

Sunday, July 31, 2016

Rebuilding MV Agusta F4 mirror mount posts AKA when those stupid little knobs on the bottom break off and the bloody indicators won't work anymore.

I had an MV F4 come in with one of the front indicators not working unless the screw was almost loose because the stupid little cast on lugs had broken off the bottom of the mirror mount.  If you've worked on an F4 you've probably had it happen as you tighten the front fairing mount bolts.  Do them up tight on some - no problemo.  Gentle as can be on the next one - break.  They truly give me the hibbidyjibbidies.

On this occasion I thought I might do some thing about it.  I measured the knobs on the other side and they were 5mm diameter, so I drilled the mirror mount at the broken points, tapped some M5 threads into the holes, loctited some M5 screws in (long shank ones with the threaded section precut to the desired short length) and then cut the end off the screws so the desired shank length was left.  As per the photos below.  I was rather impressed with the results I must say, even if one is a bit angled.  It's all done freehand due to the impossibility of holding anything in a fixture.  Unless, of course, you'd made a custom fixture beforehand.  I didn't have that much time to hand.  And yes, the indicator now worked with the screw tight.

Clutch pressure plate bearings and the like

Third 748 story, this one a 1997 SP.  The owner was complaining of the clutch losing pressure over time.  Fluid level was ok, colour typically not, no obvious leaks, aftermarket slave.

I removed the slave to check for the rubber bellows (there, but fitted the wrong way around), but found a short section of pushrod stuck in the end of it.  There are two o-rings on the pushrod that goes across the engine, just inside the cases on the slave side.  On the later 3 phase alternator bikes, a pushrod broken at one of the o-ring grooves is a very good indicator of the flywheel nut being loose.  But this was an older single phase alternator bike, and it didn't sound like it had a loose nut to me.

But I thought it might be more likely to have a seized bearing in the pressure plate, which it did.  Totally seized.  This is one of those issues that suddenly started appearing after many years of none at all.  I see quite a few 2000> era bikes with the bearing seized (or at least rough), but not older ones.  Late model bikes have a pin through the pushrod at the LH end that locates in the slave cylinder boss and stops the pushrod spinning, which is a typically involved solution to a seemingly simple issue.

The bearing in the pressure plate is a normal ball bearing.  I always figured it was used somewhat poorly, given it's not specifically a thrust bearing but someone who knows bearings told me once that it was probably just fine.  Maybe, like the 2002 ish onwards front wheel bearings, it's a bearing quality issue.  Although on this bike the pressure plate was aftermarket, so a non Ducati bearing.  Possibly worse.

As the pushrod had been spinning, it was quite stuck in the slave piston and had probably damaged the slave bore.  The easiest solution was to replace the pushrod with the later, 2001> longer part and fit a new aftermarket slave, in this case an Oberon as they are both well priced and, from my experience, well made.

It also had the usual wear at the centre of the hub, where the big steel washer slowly eats its way back into the hub.  I see a lot of this, and if the wear is not great you can add another washer to take up the clearance.  Clearance here allows the hub outer section - the splined aluminium piece - to move in and out on the cush drive rubbers that sit over the steel inner section.  This translates as excessive freeplay at the lever, as the movement of the hub moves the pack and pressure plate with the hub without lifting the pressure plate off the pack.  As an aside, I have seen (and indeed fitted) up to 4 washers here to take up the wear when the budget doesn't extend to a new hub.  It's quite amazing how thin the hub can get and not fail.

I found a photo of the one that needed 4 washers.  You can also see the wear on the pushrod - the stepped section runs in the needle roller bearing just inside the end of the input shaft.  It's not meant to have a step.

Saturday, July 30, 2016

Odd alternator cover leaks

I'm on a roll now, so possibly the second in a three part 748 tale of woe.

I had another 748 in recently, just purchased sight unseen from interstate from a shop I believe.  Allegedly a low km (3,500 miles) import, clearly bought from a bloke with a very long nose and charred pants.  It's really sad to see people get ripped off like this.

Apart from a couple of basic and unmissable compliance issues - LH drive headlight and non approved braided brake hoses - it had undersize front brake discs, front brake and neutral lights not working, steering stop broken off the frame, rusty steering head bearings and more.  Minor detail stuff too, like normal screws holding the mirrors on instead of the brass, breakaway head screws, screws holding the seat on instead of the plastic pins and rubber keeper and no tank pin grommet in the front of the airbox, which was contributing to the steering damper taking the paint of the front of the tank on the RH side.  Again (self righteous wanker statement time) just work done by someone who doesn't know the bikes or what causes issues.  Some of it well intentioned, some just sloppy.

I worked out the neutral light issue was that the dash / front section of wiring loom was for a late style single wire neutral switch, where the switch completes an earth circuit through the engine, while the neutral switch and rear section of loom was the older two wire 12V continuity switch style.  To make the light work I left the switch wire from the dash globe connected, and made a little wire to go from battery negative to the other (originally 12V power) switch wire to earth out the circuit and bring the light on.  Simple, but never going to work as was.

Another was oil leaks from the alternator cover.  Usually they leak from the clutch pushrod hole, caused by the clutch slave (which is bolted to the cover) trying to pull the cover off the crankcases every time you pull the clutch lever and stressing the goop over time until it no longer seals.  Often the cover screw behind the slave is loose too, in this case it might have been too long and not bottomed. Anyway, I pulled the cover, looked for other obvious issues, cleaned, gooped and refitted.  I also replaced the rotation sensor o-ring (old two screw pickup, original o-ring now NLA) and applied some more goop, which I really don't like doing when o-rings are involved unless I really have to.

Then I went for a ride and parked it to see what it dropped and to my profound dismay it started immediately.  At which point the source of the leaks became a little more apparent.  Certainly nothing I've ever seen before either.  The video shows oil bubbling out from under some corroded paint near the top of the cover, around the unused boss originally for the crank sensor on P8 ecu models.


It had more leaks, under the DUCATI logo at the bottom of the cover and a couple up at the front of the water pump below the rotation sensor.  Again, just oil coming through.  I don't know if it's just porous or if it is residual crash damage.  While trying to find a replacement cover I spoke to Shaun at D Moto and he said in his DDT days that they used to see the race bike covers do similar things after crashes.  No obvious marks or damage on the outside or inside, just leaking.

Steel fuel tanks and rust - the sort that makes holes and creates much agro

I had a 748 in recently - 20,000km, overall quite clean, purchased a couple of years ago out of Queensland.  It had some marks on the engine cases showing some sort of "liquid from above" staining, and when I pulled the tank there were a couple of small areas of paint bubbling with a wet look about them (never good).  Removing the pump assembly showed the extent of internal rust.  Photos show some of the tale.  The crud around the fuel pump pickup was quite amazing.  The rust holes shown were exposed using a piece of Scotchbrite only, there was no real effort involved.  It's amazing to think it wasn't just pouring out.  I was discussing this with Brad at BikeCraft and he said he has had customers come in to him with tales of no leaks until it dumped a whole tank in seconds from this sort of rust damage.

If your bike has a steel tank, it's important to look inside it at regular intervals.  Make sure the cap recess drain is clear, so any water that sits in the cap recess will drain out, not run into the tank the next time the cap is opened.  Add something to remove water occasionally.  Every time I service a bike I add about 100ml of injector cleaner.  I use Wurth brand cleaner, I think what I get is the imported stuff that you can't buy retail afaik.    It's great stuff for cleaning injectors - I've had to reset idle mixtures after road tests sometimes just due to it doing it's job - but it's also good for removing water from the tank.  Metho also will do the water removal job.

Even plastic tanks can benefit.  Sometimes you can get a jelly sort of stuff build up in them, or you get the likes of the Cagiva Raptor, where its plastic tank has a bolt in fuel pump assembly with exposed steel fuel pump wiring.  In a triumphant masterstroke of engineering design, those terminals sit at the lowest point of the tank.  Right where the water collects.  Genius.  Luckily you can get the complete plate assembly from Suzuki as a TL1000R part - at around $800 (from memory) it's about half the price of the probably now non existant Cagiva part.

Another thing I see with plastic tanks is their cap recess drains blocking at the little screw in aluminium barbs in the bottoms of the tanks.  Often the rubber hose routing after the barb is such that there is an upward run causing water to collect in the barb.  They just fill with white corrosion, which can go surprisingly hard.  I find the easiest way to clear them is to run a 2.5mm drill through them, which also illustrates how small and easily blocked the hole is.  Best to remove them before you go drilling into them while they're mounted in a plastic tank, but then again I've had a few break off before coming out.  I now keep a couple on the shelf.  Any of the late model bikes with plastic tanks can suffer from this, including Aprilia and Moto Guzzi.

Ducati 2V timing belt adjuster bearings

I had an M1100 in recently (same one I pushed some time ago) to follow up on a poor starting issue.  I see a few of these in for valve clearance services (first clearance service, or first by me if I can say that without sounding like a self righteous wanker) where the owner will comment, often once finished, that the bike can be hard to start.  Or they'll ask me "did you have any trouble starting it?"  Usually I respond with a bit of a blank look and "no, fired right up".  But when doing the clearances I would have found lots of closing clearance and no opening clearance on the exhausts.  After setting the clearances the starting issues seem to go away, so it's the next thing I would normally check after cranking compression, leakdown and fuel pressure.  In that order because it's the most time consuming.

On this particular bike I found one exhaust opener with zero clearance, and the expected lots of closing clearance.  When I say lots of closing clearance I mean in the region of 0.15mm.  Many will recognise that as being within the Ducati specified range of 0.00 - 0.20mm.  In my experience, 0.20mm will cause you all sorts of idle and low speed issues.

Not that it made any difference to this bike.  Still not a happy starter, taking 3 or 4 cranking cycles to fire and run.  I did notice, when doing the clearances, that it was very, very hard to turn the engine backwards with the rear wheel.  Fine going forward, very hard backwards.  The difference here is that turning the rear wheel in the direction of normal rotation just turns the gearbox and crank.  Going backwards engages the starter motor system through the starter clutch, and although the drive ratio means the starter spins much faster than the crank, it still should turn rather easily.  I have seen a few idler gears weld themselves to their stud over the years, and I'm curious about this one.  It will have a Motolectric starter lead kit fitted next, but I get the feeling I'll be pulling the alt cover following that.

Anyway, none of that is why I'm writing this.  Today's rambling is about what I found when I pulled the belts.  This bike has done 26,000km, and had a 24,000km (meant to be a valve clearance) service done within the last year I believe.  I always spin the belt adjuster and idler bearings to see how they feel, and on this one both adjusters felt a little rough.  A little rough means they get replaced.  The horizontal idler was making a very dry noise, so I added it to the list, then looked at the vertical idler.

It was completely seized, and you can see how hot it has been from the belt rubbing over it.  I have seen a few M696 with this bearing failing or failed on the vertical cylinder, all around 36,000km.  But not one this bad.  So it got all idler and adjuster bearings (supplied as complete assemblies only these days from Ducati) and a pair of belts.

Something to keep an eye on.

Monday, July 11, 2016

Another dyno debacle with Minnie the 400 - Ducati Monster 400 dyno runs

So, I had some time today to return to Dynobike with Minnie.  Last time, some weird stuff went down with dyno interference.  This time, I had some different Kokusan ignition units fitted and took along a couple of Ignitechs - my old one with the last curve I ran, and a new one just in case.  I checked them all with the timing light before hand and all the units (2 pairs of original Kokusan and 2 Ignitechs) had the same idle and full advance, with my old Ignitech taking longer to get there.

My old Ignitech and the Exactfit ignition coils have been on this bike with the 750 engine on this dyno before, so I figured that, if required, that set up should ultimately rule out any issues.

Bup bow.  Wrong again.  I forgot the Ignitechs would have had the rev limit set at a normal-ish 9,000 rpm.  The 400 goes to 11.  Plus they seemed to have a misfire of sorts.

Anyway, eventually some runs without rpm or air/fuel (the bike shuts both of those dyno functions down) were made, comparing cam timing settings and std vs the all conquering Megacycle mufflers.  As below.  46 mph = 5,500 rpm.  80 mph = 9,500 rpm, 93 mph = 11,000 rpm.  Ish.

Megacycle mufflers fitted.  Comparison of 108 degree inlet centreline cam timing in red, 120 degree cam timing in blue.

Same as above, but with std mufflers.  The interesting thing here is is the angle of the power curve right at the end of the 120 degree run.  You could hear it on the dyno too.

Clearly, no matter what else you're running, 108 degree inlet centreline is the go.  I had the pullies drilled for 112 and 116 too, but I wasn't going to waste anymore time with what seemed like pointless experimentation.

Next, std mufflers in blue versus Megacycle in red with cam timing at 108 degrees.  Odd linear straight line bit between 25 and 45 mph, hole between 50 and 60 mph.  

With the cam timing at 120 degrees, it's pretty much the same thing with a bit less everywhere.

For a bit of a starting point to final comparison, std mufflers and 120 degree cam timing in blue, Megacycle and 108 degree in red.  Up to 15% better in places, but then again, 15% of not much is quite a bit less.  The red curve power goes flat at 9,000 rpm and peaks around 9,500 rpm, so you can shift earlier too.

Finally, a sample of the frustration.  Blue is the run from the first session two and a half months ago.  Red is the same set up from today's session.  The only change is the ignition control units, one pair of original Kokusan units replaced with another pair.  But even more confusing is the green curve.  That's from today as well, with the Ignitech unit fitted (the new one from memory).  It had a bit of a misfire through the rpm range, but it's about 20% stronger than the runs with the original control units that followed minutes later.  If someone can explain that to me I'd be very happy.

Lastly, what timing belts look like going to 11,000 rpm.