Winter Layup - 1995 Ducati 900 Supersport

[OddDuck]

Electrician booked - should be a 15A mains socket in the wall shortly.

In the meantime, the parts and tools from Stein Dinse came in. I need to pull the rocker arms and camshafts from the cylinder heads before replacing bearings and valve guides, but didn't have the rocker arm shaft puller tool yet.

The new valve guides, standard OD, measured at 13.085 mm each. I'll see about a piloted counterbore or similar to drill the old guides out. Anyone know if there are commercial 8mm valve guide drifts sold in New Zealand?

Removing valve gear from cylinder head, first place to start is to remove both valve covers and then the camshaft cover. This just pulls off by hand, no tool is needed.

I was a bit disturbed to see metal shavings sitting on the ends of the rocker shafts. It doesn't seem to be from anything going wrong in the engine, though. I think it's swarf left over from the M5 female thread, it can be hard to clear this out after tapping.

The valve opening rocker arms are simple to remove - lever the spring clip up with a scriber or similar, get a 2.5mm allen key into the gap, finish levering the thing off while keeping it under control with the other hand - otherwise it'll fly off. Then use the shaft extractor tool. Once the shaft is out, the rocker and the three shim washers simply come out.

[OddDuck]

From previous post - see photo of opening rocker arm assembly on bench.

The closing arms are held shut by springs (these are needed for starting and part throttle operation, they're quite light in terms of spring force). They're a bit trickier to remove. The way to do it is to pull the spring tight so that the rocker arm exerts no force on the cylinder head, then it's a simple job of pulling the shaft out to strip the assembly.

Ducati make and sell a special tool for just this purpose, it's affordable but difficult to get. I made do with cable ties. The vertical head wasn't difficult, the horizontal head needed a second cable tie to make sure that the first didn't slip off the end of the spring. The screwdriver in the photo was used to help with tightening the cable ties up, this is a useful trick when checking valve closing shim clearances as well.

If you're doing this, mark everything, or use labelled bags - if you mix these components up it'll change all the valve shimming.

There's a single M6 cap screw on the head, with a small copper washer underneath it. This is an oil galley seal and it's notorious for slow leaks, leading to the tarry buildup of dirt around it. I took one out and was surprised to find that it's made of aluminium.

I'll have to check what the creep properties of aluminium are, but my guess is that they're not that good - the screw shaft will have lengthened very slightly over time and tension, releasing the seal on the copper washer. It wasn't particularly tight when I undid it, anyway.

[OddDuck]

Carrying on with the cylinder heads - with the rockers out of the way, it's time to pull the camshaft bearings.

I heated the heads in a fan bake oven, to 100 C, then used the slide hammer puller as I did for the crankcase halves. No problem whatever, everything came out easy. The camshaft oil seal pulls out with the bearing behind it.

While doing this I noticed a weird dappled pattern of hammered metal in the heads, directly underneath the camshafts. It's in both heads, more noticeable in the vertical head than the horizontal. It looks a lot like ball peening. The metal faces are fresh but I have no idea where this has come from.

There's also very obvious wear on one end of the closing rockers assembly, this was visible on all four points. It looks like the closing rocker spring exerts some axial thrust.

[OddDuck]

Next step: change the valve guides themselves.

People had suggested guide liners or internal knurling earlier - I looked at both. Liners: either pay a fortune for the liners plus tools, or find a tradie in NZ with the gear. Knurling: questions about oil fouling and longetivity. Plus pricey tooling ($500 US or more).

In the end I decided to go with replacing the valve guides since it was just more likely to work. There's a bit to it, though.

Earlier I'd mentioned the idea of drilling the guides out to paper-thin walls before simply breaking them out. I had a look at the tooling needed to do this: 12.5mm counterbore with 8mm pilot. It isn't a standard tool and has to be bought as bits in a modular system of interchangeable cutters, shafts, pilot bushings and pilot screws. The only supplier I can find inside New Zealand is Trade Tools and they want roughly $200 for everything. If I was doing this regularly I'd go for it, but it's just four guides. After Youtubing several valve guide removal videos I've decided to try just drifting the guides out.

The Ducati workshop manual says to heat the head, then take a guide drift and simply drive the guide right out of the head. The accompanying illustration clearly shows the guide being driven from the valve seat side of the head. This has to be done because the guides have a steel retaining ring. In assembly, this ring sits inside a recess machined into the cylinder head's valvegear housing floor. Once on and driven into place, it's impossible to remove. So the guide can't go down, it can only go up.

I did my reading and found a suggestion from another forum: drill the head of the guide out down to the ring, snap the remainder off, pull the retaining ring, and then drive the remainder of the guide down into the port / combustion space.

The reason for this idea is the burnt-on carbon on the guide nose. It can be very hard. If driven upwards, this will gouge the guide's housing bore in the head. Bead blasting for guide nose cleanup is possible, so is sanding, but getting every last trace of baked-on carbon fouling off at the junction between guide and cylinder port isn't easy. However if you drive the guide down, you completely avoid the carbon problem.

I bought and modified a 1/2" (12.7mm) drill bit for brass / bronze, then found that my current drill will only accept up to 10mm bits. Damn. Never mind, I'd also bought a 12mm pin punch with the idea that I could either use it directly or lathe it down a bit so it had an 8mm pilot.

Lathe's at work, using it during the day can be frustrating. Perks take second place to work, can't leave it set up over an afternoon or two. So I tried a couple of the old gearbox bearings for fit on the pin punch. Couldn't get a match on the steel bits but I could on the rubberised handle.

This went into the bench vise and I had at it with angle grinder, flap disc, cut-off wheel (for the sharp 90 degree shoulder) and Dremel (finishing and tidying up). The trick with this technique is to get the work rotating at a consistent pace while being ground, this keeps it cylindrical and on-center. The right amount of pressure with the grinder got it spinning, then I used braking with my fingers to stop it spinning too fast. There was some eccentricity and the grinder would start to bounce if it spun too quickly. Intermittent water cooling with a soaked sponge was needed. I had to manually turn the punch when using the Dremel. 400 grit paper was used for final smoothing of the surface. It's not exactly precision, best fit is about 0.2mm, but it should work.

[eldog]

If you're doing this, mark everything, or use labelled bags - if you mix these components up

There's a single M6 cap screw on the head, with a small copper washer underneath it. This is an oil galley seal and it's notorious for slow leaks, leading to the tarry buildup of dirt around it. I took one out and was surprised to find that it's made of aluminium.

I'll have to check what the creep properties of aluminium are, but my guess is that they're not that good - the screw shaft will have lengthened very slightly over time and tension, releasing the seal on the copper washer. It wasn't particularly tight when I undid it, anyway.

I thought I had labelled and bagged stuff on another project myself, I didnt realise that even though all the parts looked the same, they weren't, so much fun and time rebuilding it several times and finding that I had used the wrong parts in some places. Got it right in the end - all good fun

Has your copper washer gone, hard (work hardened) we used to heat copper washers to anneal them. I am sure some one else here might be able to provide better info.

[OddDuck]

I thought I had labelled and bagged stuff on another project myself, I didnt realise that even though all the parts looked the same, they weren't, so much fun and time rebuilding it several times and finding that I had used the wrong parts in some places. Got it right in the end - all good fun

Has your copper washer gone, hard (work hardened) we used to heat copper washers to anneal them. I am sure some one else here might be able to provide better info.

Yeah, it's easy to mix it up isn't it?

I'm sure it's the aluminium bolt releasing. I've undone old copper washers before, under steel fasteners, and they've always been alright. They release abruptly but initially they're still tight. This one was loose from the word go.

I've annealed copper washers. Don't know if you've done this one, so putting it out there for the other garage enthusiasts... a useful trick if you're short on LPG torches and the like is to use a standard domestic stove top element and a bright, clean stainless steel bowl. Washer on element, bowl inverted over the top, fresh bright foil wrapped around the catch tray under the element. The reflective metal above and below makes a radiation reflection cavity and increases the heating capability of the element. On full power, it's quite easy to take the washer to red heat - 800 C - and soak for a minute or so, enough heat and time to anneal completely. Some wet'n'dry needed afterward for the oxides, just takes a minute or so to scrub it up.

[OddDuck]

Drilling the valve guide heads out in order to remove the steel retaining rings.

I used a power drill and yesterday's modified 12.7mm drill - the rake angles were deliberately ground down to 0 degrees. Positive rake tends to feed itself into brass and then jam up or start breaking things.

Not the way a pro would do it: shavings everywhere and the drill bounced all over the place, also had a nasty habit of walking off center. The ideal is a piloted drill with a 180 degree point angle (i.e. flat), didn't have one of course so hack away it is... a spot of work with a Dremel and some breaking away of metal with a screwdriver got the rings to come off.

I checked the setup for driving old guides out and then new guides in. The 4x2 at rear of the cylinder head is there to brace it against jumping backwards across the bench.

[OddDuck]

I had a go at making a piloted drill out of a normal 12mm bit, using the same method as for the guide drift - spin and grind.

Didn't work, at least not in this setup. The flutes on the drill cause too much bouncing. The side loading is also a good way to stuff a nice new powerdrill, too. This is really work for a lathe.

OK, so the fancy, elegant stuff wasn't going to happen. Never mind, there's always a way. I decided to just take the risk and drive the guides out directly. Before doing that, everything had to be washed to get the shavings out of the heads.

Initial scrub in tray with full-strength degreaser, then total immersion in degreaser and water solution. I turned both heads over several times, did a lot of plunge and drains, etc etc... looks like almost all the shavings came out.

[OddDuck]

Heads were heated to 150 - 160 C, as per workshop manual, prior to drifting.

A note about standard domestic ovens: the hysteresis on a thermostat controller can be as wide as 35 degrees, at least it was on my oven. I found it necessary to hang around and drive the thing directly, turning it on or off at 150 or 160. I left each head in for at least twenty minutes.

Then I gloved up, wrapped the head in a shop towel for insulation and carted it off to the setup left in place on the bench. Then I just put the drift in and whacked the first two guides out. Didn't need to belt the hell out of things, lots of decent taps with the hammer got the job done.

Close examination of the guides showed that this is a dodgy way to do it. The tapered and not-flat drilling meant a real chance of a high spot, localised mushrooming, and damage to the valve guide bore in the head. I got away with it, but only just.

Right, try another idea I'd read: tap a thread into the free end of the guide, put a bolt in a few turns, put the drift right through the guide body, and drive against the bolt in the front, effectively pulling the guide out from the front instead of pushing it from the rear. A T-shaped tap handle is needed since they're down holes at an angle, I made do with a 1/4" ring spanner. The inlet guide (SG cast iron) turned out to be quite hard and difficult to tap.

This worked on the exhaust guide. The inlet guide simply broke as soon as I used the drift, and I was forced to drive from the back as before. That still worked, though. The material used on the inlet guide seems to be quite hard, it doesn't deform under the punch anywhere as much as the bronze exhaust guide will do.

The guide bores were internally wiped with a new rag wrapped over the head of a screwdriver, push and twist and repeat. The head was then reheated and the new guides were drifted into place. I'm confident that standard sized guides will be fine, there was no need to oversize. The guides self-aligned and the correct insertion depth could be felt through the drift and hammer - the ring stopped the guide moving and the pitch of the hammer blows changed.

I did one head before the other - didn't want to swap inlet and exhaust guides by mistake. Having a built up head helped.

Both sets of guides have developed a little internal rolling / mushrooming at the top, from the drift. The seal area still looks OK though.

[pete376403]

I guess the reason the inlet guide broke (when the exhaust one didn't) is the taper of the inlet guide = less support for the thread. But whatever, you got them out and the new ones in.

[Ocean1]

I guess the reason the inlet guide broke (when the exhaust one didn't) is the taper of the inlet guide = less support for the thread. But whatever, you got them out and the new ones in.

Inlet guide looks like SG iron, bronze guides would normally be aluminium bronze. The iron is more brittle, and AB is surprisingly strong.

A good trick with taping them for a plug too, means the guide is effectively pulled which narrows it, making it looser, as opposed to pressing the top which tends to force material into the bore making it tighter. I've known it make the difference between successfully pressing a bush out and having to laboriously nibble it out in the mill.

[Ocean1]

The aluminium bronze turned out to be quite hard and difficult to tap.

It's wonderful stuff, but it's an absolute bastard to machine.

Next time you want a drift or the likes made just PM me and if I'm around I'll sort it out.

[OddDuck]

It's wonderful stuff, but it's an absolute bastard to machine.

Next time you want a drift or the likes made just PM me and if I'm around I'll sort it out.

Thanks mate!

You're right, the paler (silver) colour guides that I thought were aluminium bronze turned out to be magnetic. Looks like they're SG cast iron. No wonder the one that I tapped a thread into broke as soon as I tried using it.

I measured the guides after drifting and saw how much they collapse under the pressure... 7.99mm on the outer, unsquashed part, 7.95 / 7.96 further in. This would mean direct interference fit with the valve stems on assembly. Time to ream to size so that there's a running clearance.

I had a bright idea: heating the heads prior to reaming, so that I could cut nominally undersize holes (8mm when hot, something less when cooled to room temperature). The reason for this is wear and tear on the valve stem. Running clearance is supposed to be 20 - 45 microns, and in places, if I ream to nominal 8.000 mm, I'm already most of the way through this.

It actually did sorta work when I tried cooking and measuring, but the shift when heads are heated to 60 C is about 5 microns. Covering the 20-ish micron shift I wanted isn't possible without overheating the heads, then there are worries about the reamer heating while I'm working and jamming up as well.

Have to live with it. I reamed the still-warm heads and took everything to 7.995 - 8.005.

Hand reamers (first time I've used one) have a gradual taper on the nose, it means it's possible to feel the reamer align when you start cutting. It was about $29 delivered. I'd been told back in school that reamers lose their edges instantly if they're ever turned backwards, so I kept turning in the cutting direction even when pulling it out.

I haven't fitted valves yet, but given that the seats need reconditioning anyway, I'll be looking at valve seat cutters next. This site turned up somehow:

http://www.cylinderheadsupply.com/ne...t-cutters.html

These guys will sell bits and pieces individually, there's good info in there about what's needed.

[OddDuck]

Valve seat cutter gear ordered - it's pricey tooling so I've taken a risk and just got the 1.5" standard cutters for 30 and 45 degrees. The exhaust valves are 38mm / 1.5", the inlet valves are just over 1.75". The cutter blades are supposed to be expandable by 1/4", which should take me to the size needed for the inlet port. Cross fingers, if it doesn't work then I'll have to reorder.

Electrician's been, I now have a 15A wall socket and a 2.5mm2 conductor extension cable. Went and got a D size gas bottle after, now all I need is a steel plate to serve as a welding table for practice before attempting the frame repairs.

Tried installing valves in heads this morning, just to see how the guide install and ream worked out. Had a surprise: a couple of the valves wouldn't fit.

It turned out that one valve had a raised ridge of metal on its stem, presumably from the opening shim. A spot of work with a needle file and some 400-grit paper sorted this out, before it ripped up my nice new guide. It still didn't fit. I realised that I hadn't taken the reamer deep enough to get the tapered lead portion clear of the guide so did it again.

It was similar for the other valve, but in this case there was a residual carbon deposit left at the very base of the valve stem. The 400-grit sorted this out, then it was necessary to ream again as well.

The valves look like they seat very nicely, from the combustion chamber side. It's a different story looking in the port: daylight is visible through the gaps. I checked clearance with the feeler gauge and the worst result was roughly 0.1mm. Cutting definitely necessary.

[OddDuck]

A while ago I'd had the idea that maybe the shift drum sharkmouth grabber arm thingy had got worn, thus explaining issues with shifting around 5th / 6th gears.

Finally checked new part vs old. They both come in at 26.4mm, with the old unit being just a tad longer - 26.44 or so.

So this component doesn't seem to be prone to wear. After having seen the design of the gearbox, I'm going to call my gearshift troubles as being from slower rotation of the shaft carrying 5 and 6. Shift with higher RPMs, or allow longer, or both.

So, I started bottom end reassembly. Before the crankcase halves can go together, the conrods have to go back onto the crankshaft. Before doing that, it's probably a good idea to check running clearance, even though they're new bearing half-shells.

This was done by assembling them, using the old bolts, and retorquing to full assembly loading. I then measured with transfer gauge and 25 - 50 mm micrometer.

Guess what. They run looser than the old ones. Previous running clearance was about 60 microns, on a specification of 25 to 56 microns. These are running at around 65 to 70 microns.

After a check, it turns out that there are tolerance classes in these bearings, denoted by colour. I'd bought red standard, but it seems that the forums are full of posts about rebuilding with blue, whatever size that is.

This is not to be confused with buying 0.25mm or 0.50mm undersize bearing shells - this is what's done if the crankshaft bearing journal has been damaged and needs regrinding.

Just to make life more fun, Ducati use the colour coding system to denote balance classes in the conrods and crankshafts as well. The balance colour is painted onto the conrod base cap webbing, between the two bolts. My colour is black. There's also a letter engraved onto each conrod, by hand: B. This is for the dimensional tolerance class of the crankshaft journal bearing diameter, there are A's (slightly bigger) and B's (nominal 42mm).

I managed to get the journal galley plug replaced (yeah, it's been that kind of a day). The rag in the picture is marked with ash wiped out of the threads, I wanted this gone before installation. Loctite 222 was used as per the manual, with screwing in and out to make sure it spread over the entire thread surface. There isn't a torque figure quoted anywhere so I went by feel. There isn't a proper shoulder on the inside to tighten against, the thread simply tapers out.

Pending word from the dealer (and I'll try Stein Dinse on Monday), I've made a start on pressing the new bearings into the crankcase halves.