I’m average with a spanner. I learned a more from this repair than just the shaft replacement and thought others might find it useful.
Denise called from Geelong and said that she had been travelling on the Freeway when the throttle started sticking and she needed to force it open and closed. Then it went bang and the bike was doing 9kph. She managed to chug it a few kms on idle to a service station.
I packed a couple of throttle cables and a tool kit and rode down to Geelong. The idle cable at the handgrip had stretched and broken a few strands. I pulled off the airbox and the idle cable at the bottom was holding by one strand. Other strands were curled up, some inside the casing, and the cable was seized. Denise pointed to the cable wheel (Throttle Lever Assembly) and said “Should that be like that?” Nope, it was just bobbing out on the spring. The butterfly shaft was broken.
Denise’s bike – Halo, is a hard case. A 2006 XB12R with 152,000 km on the clock. Denise has taken Halo as far north, south, east and west in Australia as you can take a Firebolt with a full tank of fuel and 20 litres in jerry cans and, to show it, the bike carries more bulldust than a desert. Denise runs Mobil full synthetic and calls a major service 24,000 km and a minor service is topping up the oil in between. Halo is treated like a lowered Uly and is spared nothing on track days. Despite the hard love, it is her only means of transport and has been very reliable.
Troy trailored the bike to my place so I could look at it. The shaft isn’t available as a part; normally it means $700 for a new throttle body. I’d had a new shaft made for Dave when the one on his 2007 XB12Ss broke a while earlier. The dealer was going to fit the shaft, but ended up petitioning H-D for a new throttle body. H-D came through even though the bike was out of warranty. I got the shaft off Dave to try as a repair for Halo. The 2006 and 2007 XB12 throttle bodies are different part numbers. The obvious difference is that the key for the TPS is near horizontal on a 2006 and near vertical on a 2007. I wouldn’t know if there were any critical differences until I pulled it down. The 99-06 TPS is on extended backorder, and the TPS connector on Halo was taped up while she waited for a new TPS. Taking a 2007 TPS was an advantage.
The butterfly was jammed shut; really shut - with grips on the TPS end of the shaft I couldn’t rotate it. I didn’t want to use too much force; I needed the butterfly unbent. I undid the front screw that was holding in thread on one side only – the shafts seem to break around the front screw hole. The butterfly was still jammed but the rear piece of the shaft was now slightly loose on the butterfly. The screw wasn’t holding it tight.
In the picture with the absent screw you can also see that the wiring harness in front of the throttle body was rubbing through on the front breather. That harness does the coil, injectors and the upper sensors. I’d better check my Firebolt; that one’s a breakdown waiting to happen.
I tried to undo the rear screw. I may have got a fraction of a turn, but the screws are slot cut and bent into elliptical recesses in the shaft to stop them falling out into the engine. This may have saved Halo from engine work. These are stainless screws – so not magnetic.
The only way I could see was to drill the head off the screw, and to do that I had to remove the throttle body to prevent the swarf from going into the intake. I looked up the workshop manual. It made the removal seem pretty straightforward. With the scoops off the pictures showed good access to the bolts. When I got that far it became evident the pictures in the workshop manual were with the front of the engine rotate down. In hindsight removal and replacement could have been faster with an engine rotation.
I’ll describe the removal here. The workshop manual used terms like “loosen this” and “remove that” – which I didn’t find all that helpful. The first challenge was the screw into the throttle body through the bracket that doubles as the coil mount and a strap between the two heads. It is a black mushroom head and I couldn’t tell if it was Allen or TorX. The picture in the manual is a pretty ordinary black and white; the parts list doesn’t say (and that screw isn’t used anywhere else on the bike). Even a torch didn’t help much because it is tucked up well under the frame and the angle of view wasn’t good. It was a TorX 27. I had to buy a ¼” ratchet drive and get to it from above with the ratchet handle between the frame and the head strap.
I also struggled with the flange bolts on the primary side. I could get the first fraction of a turn on them with a ½” open ender. Then I tried ring and ratchet ring spanners, but the play in the spanner was more than the distance that I needed for a part turn. I was getting into the mood to rotate the engine when I tried a ½” socket on a 3/8” drive with a universal and extension. The front cylinder particularly didn’t look like it would give enough space for a universal to work, but all credit to space saving at Buell, I could just get the socket on and I could just get the universal to turn. I had to move the front one around a few times because the universal wouldn’t work on a full turn.
The flange screws on the cam cover side were fairly straightforward with the special key. I had to use a normal Allen key to get the initial unlock, and then after that I used the intake manifold key. I’m not fond of the ball type universal ends so I used the straight side of the key as often as I could. I also held the key into the socket with one hand while I turned the key with the other. The last thing I wanted to do was round off the socket. I had to unscrew them with the key all of the way out. They didn’t become hand tight until just at the end. It was a fraction of a turn with each fit of the key. This is a picture of the key in case it helps.
I drilled the head off the rear butterfly screw. It is a stainless screw and took a bit of time. One edge broke away and I used a Dremel with a grinding wheel to get the remaining part. I was able to slip the screw through the butterfly and extract the TPS side of the shaft. I pressed down on the LH rear part of the butterfly and it came loose. This is how the new and old shafts lined up.
The only difference I could find was the keyway for the TPS. I was going to look for new Delrin split bushings and a new felt washer (to stop air leaking along the shaft on the cable end). In the end they were in good nick. I transferred the e-clip from the groove on the TPS side of the old shaft onto the new, put the shaft in with the bushes and washer and found it was a nice snug fit. New stainless screws were easy to get from a bolt shop. They looked identical except for the slot at the end. Pity I didn’t photograph the bits and pieces before I put it together.
I used a screw in the idle adjustment hole to hold the shaft at an angle where the butterfly wouldn’t stick against the sides. I used loctite red on the screws and wound them up and down through the threads a few times to get good coverage on the threads of the screw and the shaft. I couldn’t slot and bend the screws without damaging the shaft, so I crushed the exposed parts of the threads slightly with grips so the threads would bind if the screws did come loose. Looking at it now, the old arrangement held the screws to the shaft when the shaft broke through the hole. My repair won’t do that.
The butterfly is a few hundred mm off-centre. The angle the picture is taken from makes the gap look bigger than it is. I should have wound the screw in the idle adjuster down a fraction so the butterfly was closer to the point that it sticks. This would have helped with centring.
The angle of the TPS keyway went from this:
The only thing I wasn’t happy with was the stop for the cable wheel. I didn’t have my Firebolt at the time to make a comparison, but the way the wheel doesn’t make full contact with the limiter on the throttle body was a concern. I’ve since been able to check against my bike and it isn’t right. I’ll have to revisit it and try to sort it out. For the moment I’ve told Denise she shouldn’t ring its neck on track days. If the wheel pushes passed the limiter it could break the new shaft.
The rear seal had been leaking. Bull dust is very fine, but it won’t blow around the flange and up under the seal; it has been sucked up.
I decided to have a go at fitting the thicker James seals. Just about every mechanic I know who’s tried says they they are almost impossible to fit and I’ve seen at least one thread on BadWeb that says about the same. I was expecting to have more trouble. I put rubber grease on the seals where they go into the flanges, loosened the primary side bolts right off, positioned the flange slots over the bolts and started the back screw on the cam side by hand. I pushed the throttle body down and back and started the front screw by hand. I nipped up the primary side bolts to get the flanges square and then did the screws up progressively until they just touched the flanges. Only the first couple of turns on the screws were possible by hand. After that it was small bites with the manifold key – using the straight side where possible and pushing the key in with one hand while turning with the other. Once the screws were touching I tightened progressively from each side. The torque on the bolts and screws is only about 2 ft pounds and it looks like the flanges could break if they weren’t tightened correctly. In the end about 2.5 -3 mm of seal was visible. The stock seals aren’t crushed down, but they can’t be seen with the flanges tightened.
When the bike was back together I reset the TPS and took it out and started it. It wouldn’t idle. Failed? On a ride around the block I could feel it was running on one cylinder only, and the front pipe was cold. The front plug was completely coked. (Denise told me that the front plug runs black and the rear runs light tan, and I guess chugging the bike on idle wouldn’t help). I changed the plug and the bike idled erratically often hunting and fluctuating about 500 RPM. I took it for a ride and the idle was a bit better when I came back but still not great. I set idle as close as possible to 1000 RPM and tried to leak test with brake cleaner. There was no clear indication of a leak but the idle was erratic which makes it difficult to tell.
Sprocket had fitted the James seals to a couple of bikes so I rang him to go through what I’d done. It all sounded good to him including having 2.5 - 3mm of seal showing. He told me that he sometimes rotated the TPS on the larger top screw-hole to smooth out idle. I’d noticed that the TPS was only showing around 3.7 degrees. I had been going to look at that later and hadn’t thought of it affecting the idle. The maximum I could get with the idle set at the 1000 was 4.3 degrees. I slotted the screw-holes to get it to 5.1 degrees (5.9%). (It kept falling back 0.2 degrees when I flicked the throttle so I set it to 5.3 which settled back at 5.1).
I assembled Halo again and this time she started easily on a chill Melbourne winter evening and idled rock solid at 850 RPM then climbed to 900 RPM about a minute later. After a ride the tacho showed idle at 1000 RPM. The thing was the idle, cold or warm, was now steadier than I can recall seeing for any XB. My Firebolt is pretty steady, and I’ll have to check, but Halo may now be better.
I rode the bike for a few days to test it. I set the AFV to 95 and went out early Sunday with no traffic for a run along the freeway. When I got back the AFV was 100. After riding it to work and back in heavy traffic the AFV dropped to around 90. I like it higher than 90 but I’ll need t check with some others with race kits to see if their bikes are about the same.
The bike now goes from lug to high revs with no flat spots. It cruises smoothly with no cough or hesitation. Denise says that the bike has never idled like that and can’t remember it ever running better. Whew.
When I was sure it was running right I changed the rear plug and checked the front plug. The rear plug was light tan (but quite eroded and corroded – so it hasn’t been changed recently). The new front plug was a nice light grey. I think that the front plug fouling is because of the leak in the rear intake manifold seal. The rear cylinder would have been running leaner and since the O2 sensor is on the rear cylinder the AFV should have scaled up to stop the rear cylinder from running lean. The same scaling is applied to the front cylinder and this could make it run too rich. Be good to check the front plug again in a couple of months.
It was interesting talking to Chris in the workshop at a local dealership. He said that occasionally he has had to slot the screw holes on a new TPS to get the angle right. He also showed me his collection of old throttle bodies. He said that before having a shaft made up it would be worth going through to see if there was an old shaft that could be reused.
What caused the shaft to break? If the butterfly weren’t jammed shut I’d say that the idle cable may have jammed and when the throttle was reefed open it pulled on the wheel for the cables and bent the shaft. The thing is that if the idle cable was jammed, it couldn’t pull the butterfly closed and jam it. The idle cable could have been damaged when the shaft came out and left the wheel out of line with the cables. The old shaft looked bowed when the two halves were put together (the new one looked bowed as well, but was straight when I sat it on the desk). It had strange wear marks on the cable side just near the butterfly.
That’s a lot of wear on a hardened steel shaft that is running in a cast throttle body.
No Smoking Gun but I’d say that if the cables start sicking avoid forcing the throttle. Dave rode his bike for a while not knowing that the shaft was broken, the bike was idling and running badly but it took a while to find. He was lucky that the screw around the broken hole didn’t work loose.