Augusta freewheel thoughts and ideas.
I’ve spent much time on the Augusta’s freewheel which has delayed putting the gearbox back together but the subject is interesting enough to merit a detailed discussion so here goes!
I had thought all along that using the freewheel was probably not a good idea when driving an old car in current traffic, particularly on the hills around here. Discussions with fellow Augusta owners confirm that few of them use the freewheel and that some even disconnect the Bowden cable and wire the actuating lever in the permanent drive position, as had been done to our car. The previous owner in Italy had commented on YouTube that the freewheel can cause problems, so I expected the worst and never even tried the freewheel on the road, expecting to leave it disabled.
The freewheel is very similar to that used on Rovers from the 1930’s to the 50’s, with a three lobed driven cam which expands three sets of rollers outwards with a wedging action into a surrounding drum which takes the drive to the output shaft of the gearbox. When driving the rollers expand, assisted by a spring, but on the over-run they are rolled back down the cam allowing the drum to turn faster than the input shaft. A sliding internally splined sleeve, permanently engaged with six teeth on the outside of the driven drum is moved forward by a gear selector fork to engage with a six toothed dog on the input shaft, giving permanent drive, or backward, out of engagement with the input dogs to activate the freewheel. All the parts are superbly machined and surprisingly small and delicate, the drum being only 50mm bore and the rollers approximately 7.5, 8.0 and 8.5mm diameter and 20 mm long. At the front of the unit, running within the drum, and on a short length of 32mm diameter shaft between the front driving dog and the cam, is a set of three rollers 9mm diameter and 6mm wide, which act as the front bearing of the output shaft. These are static relative to both shaft and drum when the freewheel is locked, but rotate when the car is freewheeling.
Lots of wear was evident, both cam and the bore of the drum had axial ridges where the rollers took the drive and the rollers themselves had flats. The brass cage surrounding the rollers was quite worn where the ends of the rollers impacted the cage and it looked as though the rollers were trying to adopt a skewed angle to the cam rather than lying parallel. The 7.5 and 8mm rollers had at least two flats each whereas the largest 8.5 mm rollers had none. One obvious cause of the wear was that the bolt which holds the roller cage against the cam had sheared off, and lay loose inside the drum of the freewheel. This bolt is famously known to be a weak feature as despite the 12mm thread size, it has a 27mm hexagon which encourages over tightening.
I think that Lancia intended that the freewheel mode would be the normal way of driving as the freewheel selector is spring loaded towards the freewheel mode and is pulled into lockup mode by the cable. The three rollers which carry the front of the freewheel drum are I suspect too few and too small for the job they are expected to do when in lockup. This may not have been a problem when the free wheel was mostly engaged as the rollers turn and even up any wear. However if always driven in lockup mode the rollers may develop flats, eventually allowing the drum and the front end of the output shaft a small degree of wobble. Permanently locking the free wheel by disabling the actuating cable must be a bad idea and even if the owner has no interest in using the freewheel it should at least be allowed to work once per trip so as to rotate the bearing and discourage flatting of the rollers.
Driving constantly in the lockup mode also takes its toll on the driving dog teeth at the input end of the freewheel which, as we know tend to become tapered, putting a load on the selector fork, assisted by the coil spring trying to disengage the dog from the sliding sleeve and only restrained by the Bowden cable. My freewheel had suffered this and showed signs of overheating due to pressure of the sliding sleeve on the arms of the selector fork.
Having seen all this I decided that, rather than relying on disabling the freewheel and ignoring its worn out state, or going to significant expense and hassle to regrind the worn surfaces and replace the rollers for a freewheel I did not wish to use, I would eliminate it altogether.
This turned out to be quite simple as an idea but needed accurate machining of a split collar to clamp onto the input shaft between the driving dog and the cam to act as a supporting bush for the front end of the driven drum in place of the three small rollers. Since the bush does not rotate relative to either drum or input shaft it can be a tight fit on the shaft and a push fit in the drum, just loose enough that the drum can be slid over the bush during assembly. The brass cage with the freewheel rollers was removed and replaced with a washer drilled to take the locating peg on the end of the cam and for the tag washer for the cam retaining bolt. This all worked out well and there is now no detectable slop between input and output shafts and it all rotates freely and true.
I was concerned about the lubrication of the driving dog and the splined sleeve which is achieved by oil flowing through the shaft between the main gearbox and the freewheel and emerging from holes in the splines to escape through the front (three roller) bearing into the drum and out to the driving dogs. To allow this to continue I grooved the bore of the new bush and its front face and also machined an annular groove in the bore in line with the holes in the splined shaft to help spread the oil.
Following a suggestion by Karl Sänger, an idea which originated with Kees Jan Boosman, I reversed the sliding engagement sleeve. This increases the area of engagement between the splines of the sleeve and the driving dog. In the normal orientation of the sleeve, alternate splines at the front are cut away to facilitate the lining up of dog teeth and splines as lockup is engaged. As mentioned earlier, the splined sleeve was a tight fit on the outside of the driven drum and needed a very small amount grinding from the peaks of the splines, done with air powered grinding wheel mounted on the toolpost of the lathe. The selector fork has to be reversed to suit the repositioned sleeve and this involved drilling a new hole in the selector shaft for the locking screw of the selector fork to locate the fork in its new position. The hardened shaft cannot be drilled with a normal High Speed Steel drill bit but a carbide slot mill worked well. I also replaced the helical spring on the selector shaft with spacer sleeves to front and rear to keep the selector and sleeve in position, taking care to preserve a 3mm gap between the forward face of the driving dog and the dividing partition in the centre of the gearbox. This is a point at which wear can occur if the freewheel Bowden cable is too tightly adjusted. With the 3mm gap, the splined sleeve engages fully with the driving dog teeth, even allowing for some wear between selector fork and sleeve. Finally I removed the freewheel operating lever and blanked off its bearing in the side of the gearbox.
With all new bearings in the gearbox, reconditioned spigot bearings and the removal of the freewheel, the gearbox feels very smooth and free turning, although no doubt first and second gear will still growl away as before. We’ll see once I get the car back on the road.
I’ll add some pictures shortly – I’ve been a bit held up recently as the lawnmower and sundry gardening kit were given notice to quit their habitat in the potting shed which is now required for its proper function by the chief gardener. I’ve had to build a new shed as such mundane functional kit is not allowed to invade the car department!!
Mike
