Ford 300 Inline Six

[dnalot]

eek

"eek a mouse" one of my favorite musicians.

Mark T

 

[propclock]

Beautiful work and documentation. I was wondering if the crank distorted any after bead blasting?
I have had some major revelations after the seemingly trivial stress removal from bead blasting.
Again thank you for all your posts.

 

[mayhugh1]

Beautiful work and documentation. I was wondering if the crank distorted any after bead blasting?
I have had some major revelations after the seemingly trivial stress removal from bead blasting.
Again thank you for all your posts.

No, there wasn't any additional distortion that I could measure. For media I'm using glass 'beads' (really ground up glass) purchased from Harbor Freight. A grit wasn't spec'd for the material, but I would guess it started at about 60 and has been getting finer during the past 20 years. - Terry

 

[Eccentric]

Great new project, looking forward to following along.

 

[mayhugh1]

Block construction began with squaring up a 7" long block of 7075 aluminum with a bit of excess stock on all surfaces. This particular alloy is harder than 6061 and should provide additional wear resistance to the splash lubricated camshaft that will spin directly in the block. The 5/16" through-hole for the camshaft was the block's first machining operation.

The usual approach to a hole like this would be to work from both ends of the workpiece. However, I had an extra-long .294" diameter drill and 5/16" reamer and was curious to see how accurately I could place the hole while working from only one end. With the workpiece clamped in a vise and indicated to my mill's vertical axis, the drill had to be gripped in a 5/16" R8 collet to get the needed head room. The hole took half an hour to complete with the table's z-axis used in conjunction with the quill for peck drilling. Measurements show the hole had wondered off vertical with a .001" error on one axis and a .005" error on the other. I could probably have lived with the .005" error but, since the workpiece was oversize, it was re-squared around the hole.

The next operations were performed through what will become the bottom surface of the block. These included the lifter bores, main bearing supports, and the clearances around the crankshaft and rods. These operations began with drilled-through pilot holes for the sleeves and removal of the excess stock on block's bottom surface. This seemingly trivial facing operation was critical because it established the distance between the main bearings and the camshaft which is important for a proper mesh of the timing gears. When completed, and with the crankshaft resting in its outer ball bearings, this distance was within a thousandth of its target value. The complex clearances were machined on the Tormach.

The workpiece was then flipped over so the bores for the piston sleeves could be machined through the top of the block. Slip-fit sleeves will eventually be sealed in these bores with high-temp Loctite before the excess material on the block's top is finally removed. A 3/4" Woodruff cutter with a turned-down shank was used to machine the undercut spaces for the water jackets. The water jacket shapes are fairly complex and biased toward the port side of the engine. They were designed around the starboard-side head bolt locations to maximize the volume of coolant around the sleeves. Their enlarged shapes allowed me to slightly increase the diameter of the transfer holes that will carry coolant into the head.

My CAM software wasn't happy with the undercut operations. It only understands full length constant diameter cutters for operations that must keep track of the workpiece, and it refused to generate the needed code below the deck surface. I eventually got around this by lying to the software about the shapes of the cutter and the part and by modifying the tool's approach and retract code by hand. Some previous attempts to trick this CAM software haven't ended well, and so the code was developed on a piece of scrap before being run on the block.

Although most of the block's critical machining has been completed, it's probably only 10% finished. Most of its remaining machining will be cosmetic, and at least half of the remaining workpiece will be turned into chips. - Terry

 

[mayhugh1]

More Photos ...

 

[pickleford75]

Wow! Beautiful work as always Terry!

 

[SmithDoor]

The journals were turned using side-to-side cutting motions and .005" depths of cut. The depths of these shallow cuts were set while simultaneously moving the carriage and the cross slide in order to achieve the best journal circularities and to reduce the chances of a dig-in due to workpiece deflection. In order to reduce this deflection, the spaces between the rod journal webs were packed with custom ground buttons held in place with lacing cord. All turning operations were done at 80 rpm and feeding was done manually. The finished rod journals wound up with .001" TIRs, and the four workpiece taper measurements remained at their previously measured values indicating a minimum workpiece distortion so far.

The workpiece was returned to the mill where the main journals and inch long ends were roughed in. Back on the lathe and between centers, workpiece deflection was now a major problem even with the rod journal packings. Light thumb pressure on the center of the workpiece easily created a .006" deflection - too much for accurate main journal turning. Some of this deflection was coming from the roughed-in ends which hindsight could have been done later.

I had some 1.250" i.d. seamless tubing on hand that I cut into a number of split pieces in order to stiffen the workpiece during turning. These snapped into place perfectly around the workpiece and were retained with hose clamps. These stiffeners reduced the center deflection to just over .001".

The main journals and their adjacent web walls were then finish turned. The measured TIR's were on the order of .001" with the stiffeners in place. However when they were removed, the workpiece relaxed, and the change in its shape caused two of the TIR's to increase to .004", and one changed to .003". Since the affected journals were still circular to within a thousandth, the increased runout was due to their centers shifting off the crankshaft's main axis. With the semi-finished crankshaft resting on v-blocks, it was apparent that the workpiece distortion had likely occurred sometime after finishing the rod journals and before the main journals were turned.

Next, the weird counterweight shapes were machined into the webs using the 4-axis step indexer still setup on my Tormach. After dealing with the frustrating machining errors on an uninteresting workpiece for so long, it was satisfying to finally see a crankshaft emerge. These operations removed a lot of additional material from the workpiece which added a bit more distortion and another thousandth or so to the main journal runouts. The TIR's of the inner journals were now at .005", .005", .0015", .003" and .0035". These runouts were high enough that if left uncorrected would create difficult fitting problems inside the crankcase, and the final result would most likely be an unsatisfying sloppy fit.

With the off-axis machining operations completed, the end spigots could be safely parted off with the help of a steady rest. I was finally able to measure the runouts with the ends of the crankshaft running in the ball bearings that will eventually be installed in the block. I was hoping the TIR's would improve, but they didn't change measurably.

During my Offy build I came up with a 'scraping' technique that enabled me to reduce the main journal TIR's of that crankshaft by slightly relocating their displaced axes. The journal diameters are reduced in the process, but since work on the crankcase hasn't yet started, this won't be a problem. With the crank resting between centers in the lathe and without any packings or stiffeners, the high areas of the journals were manually rotated back and forth against a razor sharp tool in the tool post. (I continued using the diamond lapped Kennametal carbide insert.) Working carefully while removing a few tenths at a time, the final TIR's were eventually reduced to .002". Progress was monitored with bluing and frequent trips between the lathe and surface plate where the TIR's were measured with the crankshaft running in its ball bearings. The relocated journals received blending polishings with 200g followed by 400g paper to bring them to a common .365"/.366" diameter. All journals then received a final polishing with 600g followed by 1000g paper. - Terry

View attachment 124715View attachment 124716View attachment 124717View attachment 124718View attachment 124719View attachment 124720View attachment 124721View attachment 124722View attachment 124723View attachment 124724

I saw crank machine from bar stock 144" long and about 8" stoke about 12 cylinder.
I would hate to machine that one it may take year.

Dave

 

[gbritnell]

Absolutely gorgeous Terry! What is your plan for the main bearings?

 

[Vixen]

Terry, More gorgeous high precision work for us to admire. Well done.

Quote: "Measurements show the hole had wondered off vertical with a .001" error on one axis and a .005" error on the other. I could probably have lived with the .005" error but, since the workpiece was oversize, it was re-squared around the hole."

So, how did you measure the hole to determine it had wondered off axis by 0.001" and 0.005"? Then how did you re-jig the workpiece to re-square it around the hole?

Thanks for posting your work

Mike

 

[mayhugh1]

Terry, More gorgeous high precision work for us to admire. Well done.

Quote: "Measurements show the hole had wondered off vertical with a .001" error on one axis and a .005" error on the other. I could probably have lived with the .005" error but, since the workpiece was oversize, it was re-squared around the hole."

So, how did you measure the hole to determine it had wondered off axis by 0.001" and 0.005"? Then how did you re-jig the workpiece to re-square it around the hole?

Thanks for posting your work

Mike

Mike,
I inserted a length of .312" drill rod in the hole and measured its height at both ends on both axes using a height gage over a surface plate to get the relative offset errors. I was satisfied with the .001" error, but I shimmed the workpiece in the mill vise to re-face the two surface straddling the .005" error. I then corrected the ends with a long 1" diameter end mill. Since the workpiece still had excess stock all around, I just had to correct my modeling to the new values. - Terry


George,
I plan to use split bronze bearings. There is a picture of them in my model in my very first post. They're similar to what I did in my split crankcase Offy. Instead of milling the oiling slot in the crankcase and drilling up through the bearing to lubricate the crankshaft as you did, my round bearings will be large enough in diameter that I'll mill the slot directly in them. - Terry

 

[kvom]

Where did you order the 7075? I've only seen it locally in round rod.

 

[mayhugh1]

Where did you order the 7075? I've only seen it locally in round rod.

Speedy Metals ...

 

[e.picler]

Hi Mayhugh!
Congratulations on one more wonderful work. I will follow this project with close interest because I also have the same plans from George.
Beautiful work.

TKS,
Edi

 

[mayhugh1]

Although most of the block's machining could be done in a single four-axis setup, I opted to machine it in a vise one face at a time. With this simpler approach, I was able to continue tweaking the block's design while making chips. The already completed machining through the block's top and bottom surfaces left them flat and usable for subsequent work-holding.

The front and rear faces contain features that wrap around the sides of the block and potentially span two different setups. Machine backlash in addition to tiny errors in tool definitions and vise setups can create significant defects in the surfaces trapped between overlapping operations. In order to eliminate manual cleanup, boundaries need to be defined that avoid splitting complex features across multiple setups.

My CAM software provides tools for limiting its tool path generation to user-defined areas on the part, but they don't always work as needed, and I wind up spending most of my time with the software trying to convince it to not cross boundaries. On the other hand, its own internal checks to ensure the actual part isn't gouged during machining are very reliable. So, I created a unique model for each setup in which I used SolidWorks to extrude material (in the same shape as the workpiece) over the part's keep-out areas. To the software this additional material is an inviolate feature on the final part. Since it's easier to visualize than explain, I've included a rendering of the model used to machine the front and rear ends.

The front-end was machined with the workpiece gripped vertically in a vise. The operations' depth was chosen so the fillets connecting the gear case to the sides of the block would fall inside only one setup. In my version of the block, the gear case was extended to accommodate the crankshaft's thicker front bearing and a few changes planned for the timing gear. Instead of securing this gear to the camshaft with a setscrew, it will be slotted and attached with SHCS's to a flange on the camshaft to provide a vernier for valve timing. A bolt-in retainer for the front bearing was also added inside the gear case.

Even though the workpiece sat pretty high in the vise, it still had plenty of mass to dampen any surface spoiling vibrations created during milling. Just to be sure, though, a pair of 4-5-6 blocks were clamped to it.

The rear-end machining setup was similar. The rear of the block was also extended to accommodate the thicker rear bearing and its retainer. The front and rear o-ring shaft seals in the original design were omitted since the outer bearings are being replaced with sealed ball bearings. The extra space inside the bell housing will be filled with a faux torque converter that I plan to add for a bit more angular momentum. I also hope to investigate using an electric starter. - Terry

 

[Scott_M]

Hi Terry
Gorgeous as always !! Your fits and finishes are always a treat to see.
You have become very proficient at lying to sprutcam . I am still using version 7 and know all too well what you mean.

I still think the best part of your builds is your documentation, photography and willingness to share it all with us. And sometimes "Warts and all" but there is always an in depth resolution on how the "Wart" was dealt with. Outstanding

Thanks again

Scott

 

[kuhncw]

Terry,

Very nice work as always. Thanks for explaining your CNC machining sequences.

Chuck

 

[kuhncw]

Terry,
Which surfaces or features are you using to locate the part when you flip it end for end?

Chuck

 

[mayhugh1]

Chuck, I used the same corner of the workpiece for both ends. The ends of the workpiece match to within a thousandth even over the 7" length.
-Terry

 

[kuhncw]

Thank you, Terry.

Chuck