The two lathe tools shown in one of the photos in my previous post were used to turn the crankshaft journals. Both of them use sharp carbide grooving inserts that were shop modified. The gold colored insert is .120" wide, and in order to reduce chatter its center was notched out. The outer sides of the insert and the insides of the notch were relieved with a diamond burr to allow shallow side-to-side turning. After relieving the insert, its holder had to be ground down to avoid rubbing on the workpiece. At 80 rpm, this tool was capable of turning the main journals to a predictable diameter with a smooth surface finish using a d.o.c. of .010" (dia). Unfortunately, the holder wasn't stiff enough to work as well on the rod journals which required more stick out.
The rod journals were turned using the black .204" wide insert. The stock insert already had relieved outer sides, and so relief had only to be added to the insides of the notch. This insert was also capable of acceptable surface finishes, but its huge contact patch even with the added notch combined with the flex in the system to make actual d.o.c.'s somewhat inconsistent. Turning all four journals to a common diameter was hit and miss, and when I had somehow managed to get all four journals to within a thousandth of .537" I stopped. The TIR's all wound up slightly less than a thousandth.
The rod journals were finished in .008" d.o.c. (dia) steps, and the TIR and diameter of each was checked after every pass. The technique that gave the best results was to start the turning operation in the center of the journal and then power feed to its leftmost side and stop. After reversing the feed direction, the cutter was power fed to the rightmost side. The feed direction was reversed again and the cutter once more power fed to the leftmost side. Because of space limitations, the journals could only be mic'd in their very center, and so it was important to not leave behind a narrow and difficult to detect ridge in the center of the journal. All operations were performed with the lathe in back gear and set to 80 rpm, and so the process was pretty time consuming. The journals were finally polished at 400 rpm using 600g paper.
Earlier, when roughed out on the mill, excess stock was also left on the webs on either side of the journals. This was removed next using a pair of repurposed boring bars. At this point, the offset turning operations were complete with (hopefully) no further need for the offset center-drills. The close-fitting rod journal spacers will remain in position until the crankshaft is completed. I didn't see any change in any of the TIR's that I continually measured during the rod journal turning operations.
The workpiece was re-installed in the lathe with the crankshaft's main axis between centers. With some care in positioning the drive dog and adjusting the tailstock pressure, I was able to get the main journal TIR's matched to the values measured before and after the rod journal turning operations. Even with the rod journal spacers in place, however, it was possible to more than double the runout with too much tailstock force. The excess web material left on either side of the main journals was removed next.
The workpiece was returned to the mill where the rear (ball) bearing and flywheel retaining surfaces were hexagonally roughed. It was then returned to the lathe so these areas could be semi-finished with a conventional turning tool. Again, all the TIR's remained consistent with those previously measured.
The workpiece was returned to the mill where the front (ball) bearing surface was roughed in. The Offy crankshaft has a skinny extended nose that will eventually receive a Loctite'd drive gear and a hardened nose for the starter clutch. This troublesome section will weaken the workpiece and likely create problems for the main journal finishing operations, and so this portion was left for later.
After returning to the lathe, the front (ball) bearing surface was semi-finished. At this point none of the TIR's had changed since that mysterious overnight movement after the initial main journal semi-finishing step. I began final finishing the main journals with the front (ball) bearing journal which went perfectly. However, after finishing the center journal, I found its TIR had ballooned to .004" - likely due to workpiece flex.
After grinding a razor sharp HSS tool and indicating it in the toolpost, I was able to scrape away most of the runout by manually rocking the spindle back and forth over the journal's high spot. With no way of putting material back though, I ended up with a .005" undersized journal. The results were essentially the same with the other two journals even with the HSS tool.
I roughed in the two ends of the crankshaft so it could be trial fitted in the crankcase lower half without the front and rear ball bearings. The fit wasn't as sloppy as I expected, probably because the three TIRs aren't lined up with respect to one another. In any event it will become worse as the engine is run and the bronze bearings become 'wallowed' out.
I decided to scrap the part and start over. I rechecked the snugness of the rod journal spacers, and even though they felt correct, I can't help but think they were at least part of the problem. Since I'm now reasonably confident with the stability of the material, on my next attempt I'll likely finish the main journals before even roughing in the rod journals. - Terry
The rod journals were turned using the black .204" wide insert. The stock insert already had relieved outer sides, and so relief had only to be added to the insides of the notch. This insert was also capable of acceptable surface finishes, but its huge contact patch even with the added notch combined with the flex in the system to make actual d.o.c.'s somewhat inconsistent. Turning all four journals to a common diameter was hit and miss, and when I had somehow managed to get all four journals to within a thousandth of .537" I stopped. The TIR's all wound up slightly less than a thousandth.
The rod journals were finished in .008" d.o.c. (dia) steps, and the TIR and diameter of each was checked after every pass. The technique that gave the best results was to start the turning operation in the center of the journal and then power feed to its leftmost side and stop. After reversing the feed direction, the cutter was power fed to the rightmost side. The feed direction was reversed again and the cutter once more power fed to the leftmost side. Because of space limitations, the journals could only be mic'd in their very center, and so it was important to not leave behind a narrow and difficult to detect ridge in the center of the journal. All operations were performed with the lathe in back gear and set to 80 rpm, and so the process was pretty time consuming. The journals were finally polished at 400 rpm using 600g paper.
Earlier, when roughed out on the mill, excess stock was also left on the webs on either side of the journals. This was removed next using a pair of repurposed boring bars. At this point, the offset turning operations were complete with (hopefully) no further need for the offset center-drills. The close-fitting rod journal spacers will remain in position until the crankshaft is completed. I didn't see any change in any of the TIR's that I continually measured during the rod journal turning operations.
The workpiece was re-installed in the lathe with the crankshaft's main axis between centers. With some care in positioning the drive dog and adjusting the tailstock pressure, I was able to get the main journal TIR's matched to the values measured before and after the rod journal turning operations. Even with the rod journal spacers in place, however, it was possible to more than double the runout with too much tailstock force. The excess web material left on either side of the main journals was removed next.
The workpiece was returned to the mill where the rear (ball) bearing and flywheel retaining surfaces were hexagonally roughed. It was then returned to the lathe so these areas could be semi-finished with a conventional turning tool. Again, all the TIR's remained consistent with those previously measured.
The workpiece was returned to the mill where the front (ball) bearing surface was roughed in. The Offy crankshaft has a skinny extended nose that will eventually receive a Loctite'd drive gear and a hardened nose for the starter clutch. This troublesome section will weaken the workpiece and likely create problems for the main journal finishing operations, and so this portion was left for later.
After returning to the lathe, the front (ball) bearing surface was semi-finished. At this point none of the TIR's had changed since that mysterious overnight movement after the initial main journal semi-finishing step. I began final finishing the main journals with the front (ball) bearing journal which went perfectly. However, after finishing the center journal, I found its TIR had ballooned to .004" - likely due to workpiece flex.
After grinding a razor sharp HSS tool and indicating it in the toolpost, I was able to scrape away most of the runout by manually rocking the spindle back and forth over the journal's high spot. With no way of putting material back though, I ended up with a .005" undersized journal. The results were essentially the same with the other two journals even with the HSS tool.
I roughed in the two ends of the crankshaft so it could be trial fitted in the crankcase lower half without the front and rear ball bearings. The fit wasn't as sloppy as I expected, probably because the three TIRs aren't lined up with respect to one another. In any event it will become worse as the engine is run and the bronze bearings become 'wallowed' out.
I decided to scrap the part and start over. I rechecked the snugness of the rod journal spacers, and even though they felt correct, I can't help but think they were at least part of the problem. Since I'm now reasonably confident with the stability of the material, on my next attempt I'll likely finish the main journals before even roughing in the rod journals. - Terry
