Winter Layup - 1995 Ducati 900 Supersport

[OddDuck]

Bike's built up, running and legal. Have WOF, will ride.

There's still work to do. There's the usual shakeout of adjustments after a rebuild. I want to have a look at carburettor tuning. The fuel level sensor is still waiting on replacement (I've been using the tripmeter for years instead). That kind of thing. It's niggles instead of biggies.

Bike's rolling, though...

Thanks to everyone for your support and help through this thread, it's been a blast and I have learned heaps. Hopefully I'll be meeting a few of you out there this summer

The winter layup is over. Time to go have some fun.

[SVboy]

Fascinating thread to follow thank you Sir. You are an inspiration.

[insomnia01]

Awesome read thanks for sharing

[OddDuck]

Thanks guys - it turns out that there's further comment worth making, so I guess that although the layup is over, the thread goes on...

I'll start with the insulator spacer washers I'd made up for the inlet pipes running between carburettors and cylinder heads.

The problem that I was trying to solve was how painfully loud the bike became if it ran low throttle / low RPM for any length of time. Traffic, town, even periods of time in 50 zones would go from embarrassing to painful, and in the heat of summer it'd get flat out ear-ringing nasty, even with earplugs. It was bad for me, it was bad for everyone around me. Must get bike quieter.

Repacking mufflers hadn't done it. After a while I realised that the noise correlated with the temperature of the inlet pipes, when these were cold the sound was OK, when they were hot, the problem occurred.

I spent a while going through the numbers on these and calculated that roughly 70% of the engine heat conducted to these came in via the M8 stud bolts - not through the much greater surface area of the 0.4mm paper washer. Top hat washers for the stud bolts were what was needed, not a thicker gasket or insulator spacer plate.

I am very pleased to say that idea + theory + numbers + making something and trying it = success. Problem sorted. The bike's kept the bass but it's lost the hearing-damage ringing it was making before.

[OddDuck]

Stock airbox, opened lid (with K&N filter) vs K&N pod filters.

The theory, based on everything I've read, is that stock airboxes might look pretty utilitarian but tend to win over pods. Pods are fashion statements, not true go-fast accessories. You get 5% more at the top end, but lose low and midrange torque and power - and that's where most real world riding happens. There's also more induction noise, plus carburettor re-jetting becomes necessary before lean conditions cause engine damage.

At least, that's the theory.

I've had the bike running on a few carbie tuning runs (more about that below) and it's early days yet, but the result seems to be that the pods actually don't breathe any different to the stock airbox. In a lot of ways they're significantly better.

Getting in for carburettor adjustments / sorting cables / etc is a lot easier. Induction noise has actually reduced. There might be a very slight reduction in bottom end / midrange torque, but I'd need dyno runs to make a call on this. I honestly can't tell by riding it.

One improvement has been how smoothly the motor runs. The stock airbox runs super smooth at some RPM, a bit rough at others. The pods have smoothed this effect out very nicely. I'd guess that non-equal induction intervals in a shared resonant airbox would cause this; one cylinder robs the other, depending on RPM, airflow and resonant waves. The fix is to separate the intakes.

Top end power is unknown as yet, but for now the conclusion is that pods (on this bike) work just fine.

[OddDuck]

First few days of carburettor tuning with the AEM UEGO wideband O2 sensor and gauge.

Initially this was bloody dangerous.

I don't have a dynamometer so the method of tuning is: find road, ride it, monitor gauge, come back and fiddle with carburettors. Repeat until happy.

Or in hospital / dead, of course... it is very hard to not get distracted by the readings, while riding, while navigating curves and traffic and idiots pulling out, etc etc. It's the same effect as religiously checking your speedometer so you don't get done for speeding, only it's at least twice as bad. This'd be a whole lot easier if I lived close to a quiet 100 K country road, but the closest I've got within reasonable distance is an urban motorway.

Making life more fun, the carbies were badly tuned at first (of course) and when they're like this, the gauge reading leaps around like a demented ninja. Very distracting.

As to the tuning itself... One of the reasons that I bought the Keihin FCR 41's was the ability to adjust. It's covered quite nicely on this page:

http://www.ducatitech.com/2v/inside_fcr.html

Basic settings and initial tuning are described here:

http://www.bikeboy.org/900SSwithkeih...41mmcarbs.html

I'd spent the last two days trying to tune the carbs from idle upwards. I did it this way because before I can go fast, I have to be able to go slow. I didn't want sustained lean-outs. It'd suck to cause damage to the motor. Softly softly, catchee monkey, that sort of thing. I'd go for a short ride at 50 k's, come back and turn either the slow fuel screw or the slow air screw, go for another ride, repeat, and so on. On the second day I was onto the needles, trying to tune from 1/8th throttle upward. Unfortunately it didn't work.

The problem with this sort of tuning is that it's very easy to converge to the wrong destination. There are simply too many variables. Even changing just one setting by one notch at a time won't work. The carburettor has to be tested the whole way between idle to WOT in order to gain decent information. There isn't really a safe way to do this, it has to be roughly set up and then monitored the whole way through the throttle range with the engine under load.

The booklet from Keihin which came with my FCR's show that there are four broad tuning areas within the FCR 41 carburettor's range, all set by throttle:

Closed to 1/8th: slow fuel screw, slow air jet

1/8th to 1/4: slow air jet, needle position

1/4 to 3/4: needle position and main jet

3/4 to WOT: main jet.

Engine RPM doesn't really come into it. That's the whole point of a flatslide carburettor: it tells the engine what to do, not the other way around (as it is with CV's). Unfortunately this business of four areas (and various carburettor circuits) is simplified. The truth is that there's massive overlap of the circuits.

This page goes into a lot of detail about tuning these:

http://www.factorypro.com/tech/tech_...Burns,Pat.html

The major point he makes (and after the last few days I believe him) is that the WOT result has to be obtained first. You've got to get that main jet right. There's no point in playing with needle settings (or the choice of needle itself) before that's done. Regardless, I'm riding tomorrow and over the weekend, something needs to be done temporarily while jets are on order.

I've gone for the Sudco out-of-the-box settings as a starting point and then tried my tuning methods from there, moving only the needle position and leaving the slow screws as suggested. My current opinion is that the slow jet is too rich and should be reduced a bit, and the main jet is far too lean and needs opening up.

Currently I'm on these settings:

155 mains, 200 main air jet, 60 pilot, EMT needle on the fifth notch from top, slow air screw 1 1/2 turns out and slow fuel screw 3/4 turn out.

The AFR gauge indicates ratios of 10-ish at closed throttle, 11 - 12 at up to 1/8th throttle, 12 - 15 between 1/8th and 1/4, and 15+ above 1/4 throttle. The engine gets progressively more lean as the throttle is opened further, hence my belief that the main jet is badly undersized. Raising the needle can only go so far.

That said... checking gauge readings and fiddling is working. It's working amazingly well. The bike has never run this smoothly before. It gets better every time the carbs get closer to proper tune. Gauge readings are settling down, too. There's a lot less jumping than there had been when I started. Provided I ride normally, the bike should be perfectly OK to ride over the weekend.

The pics below show the gauge on startup: it insists on displaying the magic number of 14.7 while the sensor heats up. I've done my reading and this is very much an ideal ratio, obtained with complete combustion of perfect fuel. In the real world, fuel's got additives, combustion doesn't get very long to take place, and ratios of 13 to 14 are probably the best place to be in terms of economy, power, and safe engine operating conditions.

I'll see about ordering some more jets and tuning from there.

[OddDuck]

On Saturday I had a bit of a knock... out for the first ride of summer with friends, and before the first tank of petrol, the bike broke down. Charging system failure. I hadn't gone over it during the rebuild work, I'd just refitted it.

She's laid up again. There's about 80 miles on the clock since rebuild.

I spent much of the weekend doing a tidy up in the garage. Six months to two years worth of trash got hauled to the tip. I got the bike stands hanging off the wall so they're to hand for easy use. That kind of thing. I couldn't bring myself to look at the bike properly, let alone get back to work on it, until last night.

Unfortunately this is part of this kind of work. There are ups and downs. I mean, it's really not that big a deal in the great scheme of things, it's just to have it happen so soon after getting back on the road... it really did feel like I couldn't win, for a while there. This is the kind of thing that gets people very upset, being upset means IQ replaced with emotion, which is why when this stuff does happen it's vitally important to take time off until calm again. Don't make any decisions, that kind of thing. Difficult to do when you feel like you've let your mates down, though.

My planned bike trip had been postponed anyway. One of the crew had also had issues with his ride, only sorted on the day we were supposed to go. I'll have to see what consensus is, but everyone's already pretty fed up with dramas mechanical and electrical. It might be smarter to postpone until early next year and ride with proven bikes.

[OddDuck]

And the fault itself... the charging system isn't.

Diagnosis for this is easy: take a DVM (digital voltmeter) and measure battery voltage, at the battery terminals. Run the engine. Voltage should be around 12V at idle and increase with RPM's, holding at a maximum voltage after about a third to halfway through the rev range. This test can get quite loud so keep the neighbors in mind. What I saw was 12 V, then more 12 V. No response to revving.

Before anyone says anything about 'crap Ducati electricals' neither the stator nor the RR unit is OEM Ducati. They're both Electrosport aftermarket. Ducati originals are still available, but they're big money (450 Euros for a generator anyone?) and I simply didn't (and don't) have the grand and a half necessary to get stuff in the branded bags.

So, before going further:

What is the fault?
Where did it happen?
How did it happen?
Is everything else OK?

I've got to investigate completely before making any decisions. Electrical fault finding demands a systematic, step-by-step approach. It's pedantic, it's slow, it can be infuriating. It's got to be done, though. Miss one detail and I'll never sort it out.

I went and borrowed the Haynes manual again. The first line in charging system problem diagnosis was unexpected: check leakage current through the system with the ignition off, by putting a current meter onto the battery negative terminal and seeing how many milliamps are trickling away when the bike's standing.

Allowable is 0.1 mA. What I've got is 13 to 30 mA (it cycles up and down) with accessories attached and around 8 mA steady with the bike-only wiring loom. This means I've got a short somewhere in the loom, as well as accessories causing a problem. Both of these issues have to be traced and fixed, otherwise I'm in for more problems further down the line.

The stator is easy to test for continuity: detach the two bullet connectors to the RR unit and check for ohms through the winding. Mine is coming up open circuit at all resistance ranges. The winding has failed. I'll have to drain the oil and get the left hand engine casing off to get at it. That'll be tonight's job.

Testing the RR (rectifier-regulator) involves checking the input rectifier diodes for forward conductance, using the diode setting on the DVM. There are four measurements to make (a three-input lead RR would have six):

- yellow input to red output
- red output to yellow input
- yellow input to black (ground) output
- black output to yellow input

These should all come back with 0.400 to 0.600 V showing on the DVM. That's forward voltage through two functional diodes. In this case they don't, a couple are testing higher, and the epoxy potting on the unit is showing that something within has cooked itself.

I was halfway through these measurements when I realised I'd made a bad mistake installing the RR unit about a year ago: I'd never understood that the ground line is through the casing. It's not through any of the plugged leads. I'd just assumed that it was, after finding that trying to work out the wiring schematic made my head spin.

This time around, I've got my head around the wiring diagram and can work out what's happening. The Ducati workshop manual's electrical schematic makes it quite clear that the 30A return for the ground line is via the eyelet that tags onto the RR casing, where it mounts with a pair of M6 bolts... it's pretty basic. I can't believe that I missed this earlier. It's been a while, I've long since thrown the paperwork that came with the RR away, but I'm fairly sure that Electrosport don't spell out the need to scrape the paint back for the ground connection.

The positive connection carries up to 30 amps. Current goes in a loop, so even though there's no potential showing, there'll be 30 amps on the ground return line, too. That means scraping paint back on the RR, cleaning paint out of the threads of the mounting tab on the frame, using a clean new bolt - in short, taking care to give a clean connection which can handle high current. I've got a couple of spots of slightly shiny metal. That's it. There's no way it's handling high currents properly, although it'd test OK for ohms with a meter.

There's a shift in DVM's in the photos - on the basis that I'll be doing a bit of work, I decided to update my tooling and went shopping. It turns out that you get a lot of multimeter for your buck these days. This one was chosen because it'll test inductance, quite an unusual feature which is very helpful if measuring coils.

[Voltaire]

How do you find the cold starting with the FCRs'?, mine is a pig to start with its 39's and been thru the carbs and settings and they are ok.
I'd say your electrical problems is the regulator if the alt windings are good.

[OddDuck]

How do you find the cold starting with the FCRs'?, mine is a pig to start with its 39's and been thru the carbs and settings and they are ok.
I'd say your electrical problems is the regulator if the alt windings are good.

Starting: not sure of your setup, but if you have the tank vent breather line just tucked away, try routing it into whatever air intake system you're using. The tank fumes will act like easy-start. It'll help. I haven't had the chance to set this up with my pod filters yet but it's on the list.

The other trick is to walk up to the bike and twist the throttle a few times before turning the key. Use the accelerator pumps to squirt some petrol. Give it 30 seconds to build up some vapour and try starting then.

My FCR's are set up with 60 slow fuel jets... they're running very rich, around 10 - 11 AFR once warmed up, and dropping below 10 if the bike's stopped at the lights or similar. I'm pretty sure there's a tradeoff between starting and closed to 1/8th throttle running.

Electrical problems so far look like both reg and alt are gone. Alt's testing open circuit, and the reg diodes aren't coming back evenly.

[OddDuck]

Right, got the alternator cover off. I've found the fault and it's 100% mine... I managed to completely forget the two cap screws which hold the stator into the cover during engine reassembly.

The stator has moved inboard from the cover and started skidding on the flywheel. The windings have been cut away to flats in a couple of points. A lead seems to have been cut as well.

Wouldn't have found this without investigating, wouldn't now know about the grounding issue on the RR, but it's still a very mixed blessing. At least it isn't a big deal getting the cover off and on the engine. The gasket is reusable, within limits.

I'd kept the previous stator and put heavier lead wires onto it, I'll get onto fitting that a bit later on. For now I need to sort out a replacement RR unit and get that on its way.

[OddDuck]

RR unit ordered, jets ordered. I spent tonight getting the previous stator (rather higher quality than the Electrosport unit, I think it's OEM Ducati) back into the cover and onto the engine.

The revised leads are furnace heater element connector wire. This stuff is rated for service at high temperature, high current, plus I've increased the cross sectional area from the original leads. High quality wiring is needed here, normal insulation really won't cut it.

Look for chemical stability, temperature ratings to 200 C, that kind of thing... and it's got to survive protracted immersion in engine oil, which can lead to hardened and then cracked plastic with most of the common cable insulation materials. PVC certainly doesn't survive, nylon's marginal at best. I am crossing my fingers with these stator leads because of this issue, unfortunately the only way to find out if this cable will work is to try it. That said, it's glass fibre insulation over silicone rubber. It should make it.

The same problem comes up with securing the leads. Vibration or mechanical force really can't be let anywhere near a soldered joint, especially not at this operating temperature.

One of the photos shows the leads tied into place with a twist of enamel-insulated transformer winding wire. I couldn't find anything else that I was confident would survive. Most cable ties are Nylon 66, maximum service temperature of 80 C, and not rated for immersion in engine oil. The stator windings are enamelled copper, I know they'll survive, so this particular tie should work. I've taken multiple turns to try to broaden the contact area and avoid cutting into insulation.

Bolts were checked for length and fit - I wanted the maximum number of threads used, but also to be sure about tightening the stator into place against the metal below. Loctite 222 was used, as per the workshop manual.

Another concern was sharp corners on the casing cutting into the leads and shorting the windings out. The shadow of the previous leads can be seen if you look closely enough; this shows the leads folded into a smoothed portion of the casting. Should be OK.

My cable gland assembly is back in place. Drilling for a larger cable diameter didn't work. I had to file the holes in my two-core bung out to admit cables from the sides; silicone goo was needed to seal it up against leakage.

I've found that it's helpful to use the cover removal tool when refitting. It's possible to put the cover back on slowly, instead of watching it leap into place when pulled by the magnet.

[OddDuck]

I've spent the last couple of nights indulging my curiosity and digging the potting out of the failed RR unit. Apologies for the wall of text below, this does get detailed.

This wasn't easy... there's a hard layer on the surface about 5mm deep which really wants a drill or rotary burr, then after that it goes a bit rubbery and it comes off readily with a broad-bladed soldering iron. There's smoke, there's dust, it's generally stinky and nasty. I can't see this as being good for a person's health so I don't recommend taking this up as a hobby.

I've also been doing my reading and looking at circuit schematics. This helped greatly with understanding what's going on in my former RR. There's a catch, though. The more I learn, the less I like what I'm finding out. Highlights below:

1) if it's potted, chances of repair are about zero
2) potting covers up cheap build quality very nicely
3) there are no fail-safes of any kind; the first thing going wrong will kill it
4) it's built to only just work, components are tiny and so are internal conductors
5) the casing's anodised finish really screws up the ground connection for the unit.

Seriously, it's crap. Unfortunately so is just about everything else on the aftermarket, as far as I can tell. Giveaways are cheap PVC sleeving and cheap bullet connectors which are clearly under-rated for the current passing through them, going to wiring which is a bit on the thin side.

Now this is idle speculation, but I think there's good reason for this. There just isn't much emotion driving purchases in the RR market. We get excited about mufflers and exhaust systems, we love anodised aluminium bling (well, some of us do), we'll pay silly money for go-fast accessories like carbon fibre huggers, but as long as the battery is being charged, why pay any attention to the RR unit. It's not exactly a glamour item. Buying a new one isn't something we look forward to; it's just another bill to pay.

The market, therefore, sets a price. The manufacturer's margin is the difference between that price and however cheaply the unit can be made. The difference is sweet profit, so they're made cheap, cheap, cheap.

Most of the schematics I've seen follow the same pattern as the unit I've just stripped:

There's a permanent magnet rotating on the crankshaft. It's encircled by a winding, coiled on a set of stator plates. That winding is fed, via leads and connectors, to a set of rectifier diodes. These diodes are bridged by silicon-controlled rectifiers, which (in a controlled manner) short these out, dumping excess voltage to ground. The rectifier feeds sensing circuitry (this controls the SCR's) and then outputs main current to the bike and charging current to the bike's battery and ground. There's usually a sensing lead back from the battery, to help control the SCR's.

The construction does pay a lot of attention to heatsinking. The PCB on the base is an aluminium plate, overlaid with circuitry on top and thermally connected with heat sink paste to the inside of the finned aluminium casing. Components that generate heat are nicely spread out, to avoid hot spots.

There's quite a lot that could be there but isn't, though. If a generator coil is disconnected suddenly, there's what's called inductive kick-back. The stored energy in the stator coil's magnetic field discharges with a bang. It can be several hundred or even several thousand volts. Loose connectors, bad wiring etc can lead to an intermittent connection, and this will momentarily disconnect the stator coil. Result: kick-back, and fried rectifier diodes or SCRs. One loose connector will knock a $250 RR unit straight into the bin. We accept this situation as normal.

This is avoidable. Tricks can be done with clamping zener diodes. The inductive kick-back spike can be made to bypass the circuitry safely. That safeguard isn't there though... Neither is any kind of over temperature protection. I'd been reading accounts of Ducati superbike owners who have had up to five RR units fail, due to traffic / stoplights / hot summer days / positioning inside bike fairings close by exhaust headers / no provision for cooling beyond slipstream. In this situation a thermal cutout is needed. That's not there either, much like some kind of current limiting short circuit protection for the RR output. Fuses aren't exact devices, it can be more than twice rated current to make them blow, but sustained high current will definitely cook the RR unit over time.

I'd checked leakage current on the bike again - that is, current between battery negative terminal when disconnected and bike grounding lead. This time the leakage came in at just 0.1 mA, inside specification for a healthy wiring loom. This happened with the RR unit removed and everything else still in place, so that's where the 8 mA leakage was happening.

I've confirmed that a good, clean ground connection to the RR casing is needed. I'm also satisfied that the unit pictured failed with the stator. The bike's wiring loom is OK. Replacement of both stator and RR, together with reliable connectors and cabling, will see the bike running again.

[pete376403]

Shindengen mosfet RRs are considered one of the better options. Higher-end Jap bikes (ie R1 etc) have them. They are being sold as a retrofit item to replace plain old diode rrs http://www.roadstercycle.com/Shinden...or%20about.htm

Also, where is the rr located on the bike. Is it in an airstream, or hidden away somewhere? ON the 1100 Suzuki, the original location was under the battery box, just above the swingarm pivot. I relocated it to the frame downtubes, just below the steering head, to get it exposed to better airflow. hasn't hurt and might have helped.

[OddDuck]

Thanks mate - since I've confirmed the grounding connection issue, and it's now a known quantity, I've already gone ahead and ordered a direct replacement from Electrosport. It's not a flash RR by any means but once again I'm on a clock and have to go with what I know. In parallel I am looking at design and build of my own RR (with safeguards) but that's one of those projects that'll probably blow out in terms of difficulty.

The Shindengen mosfet does sound good... that looks like the way to go for a spare on hand for the next time the Electrosport craps out. Cheaper and way easier than DIY.

Placement is lower front of the frame, directly behind and underneath the lower head bearing. It's right in the path of incoming air and nowhere near the horizontal head or exhaust header; placement is actually pretty good as far as cooling goes, even with the bike stopped.