Radial engine build
- Thread starter PeterT
- Start date
I left the turning about 0.003" oversize. The first 0.001" or comes off with full width homebrew sanding sticks made from 2" wide MDF with wet-o-dry bonded with 3M spray adhesive. The next 0.001-0.0015"" comes off with a lapping tool. This is kind of an experimental thing I tried because CI is so friggen hard it takes a while to work the surface to both a consistent diameter and finish. I made a CAD drawing & send it to a waterjet outfit. They cut the profile & then I drilled a retention screw to apply set pressure. The lapping compound is cheapo oil based Ali diamond paste I had good results with on the valves. I left the last 0.0005" for 1000 or 1200 paper. Interesting how CI can get a finish, eh?
Attachments
The skirt has a chamfer to give allowance for the rods and a small relief groove under the lip. Knock down any burrs, one last finish check, then part off.
Then apply some protection tape, hold in collet chuck & dress the lip to size
Then apply some protection tape, hold in collet chuck & dress the lip to size
Attachments
I heated the cylinder to 450F in my little toaster oven & the liner dropped in no problem. It was at ambient temperature, I didn't have to chill it. But reversing the process didn't quite go according to plan. After another 5 minute baking session the 2 parts did not want to separate. Supposedly the thermal expansion between the 2 materials is completely reversible. Even though the surface finish is pretty decent, those tiny machining marks on either surface can add some gription. I also remembered that in between my prototype I only had 0.0005-0.001" diameter interference. I subsequently ran across an article that suggested this increased interference for proper heat dissipation.
Anyways, the whole plan was to do the finish ID boring operation on all liners outside of the cylinders mostly because they can be held properly. These were my sacrificial testers so I decided they were going to come apart now. I hate using a torch but have had to do it occasionally on the RC engines when they get gummed up with caked fuel residue. So I put the assembly into a corner of fire bricks & danced the torch around until they parted ways. Interestingly after cooling the bores were identical so nothing distorted. That was good to know.
So one more liner to go and then its finishing time. I have a few things to do to prepare for that, so fingers crossed. The learning continues
Anyways, the whole plan was to do the finish ID boring operation on all liners outside of the cylinders mostly because they can be held properly. These were my sacrificial testers so I decided they were going to come apart now. I hate using a torch but have had to do it occasionally on the RC engines when they get gummed up with caked fuel residue. So I put the assembly into a corner of fire bricks & danced the torch around until they parted ways. Interestingly after cooling the bores were identical so nothing distorted. That was good to know.
So one more liner to go and then its finishing time. I have a few things to do to prepare for that, so fingers crossed. The learning continues
Attachments
Last edited:
I put in a fair amount of work lately but one would never know it just looking at the parts. I've had to learn some important things about interference fit & tolerances that resulted in taking the long road around. I mentioned the CI liner is heat shrunk in the aluminum cylinder. My plans/instructions were very vague about this & tester in previous post showed this. In lightweight air cooled engines like this, liners cant really be a loose-ish sliding fit because you wont get proper contact for thermal dissipation once its running & other mechanical issues. It also cant be too much interference or else removing the parts is difficult if not impossible. Plus its not good when both parts are under strain like that.
Another consideration is that any amount of interference literally puts the squeeze on the liner. So assuming you had a perfect liner bore beforehand, it now will be smaller once shrunk in place. I assumed this amount would be quite small because CI is stronger than aluminum & slightly more wall thickness. In reality it was quite significant, about 50-75% times the diametric interference in my case. Another complication is the diameter squeeze reduction shrink may not be uniform if there is more cylinder mass at the top than the bottom as is the case. Anyways I naively lapped my liners within 0.0001" each (outside the cylinder), which in hindsight was a complete waste of time for reasons above. They needed to be re-lapped once married to the cylinder. More on this later but for now some more build pics.
After the liners were OD finished & rough turned inside all in same initial lathe setup, they all got returned to the 5C collet chuck for consistent finish boring. This was my first serious go with a tool post grinder. The plan was to bring them to within about 0.002" & leave the last for lapping operation to simultaneously true any egging + final bore + final finish. I wasn't super thrilled by the grinder finish but it looks worse than it is. Lapping removed the harmonic pattern very quickly so they are probably less than tenth deep. I have some ideas of how to improve the TPG with stiffer arbor. The grinding was quite accurate & process was controlled so that's important. In reality I need a 1/4 scale Sunnen hone LOL
Another consideration is that any amount of interference literally puts the squeeze on the liner. So assuming you had a perfect liner bore beforehand, it now will be smaller once shrunk in place. I assumed this amount would be quite small because CI is stronger than aluminum & slightly more wall thickness. In reality it was quite significant, about 50-75% times the diametric interference in my case. Another complication is the diameter squeeze reduction shrink may not be uniform if there is more cylinder mass at the top than the bottom as is the case. Anyways I naively lapped my liners within 0.0001" each (outside the cylinder), which in hindsight was a complete waste of time for reasons above. They needed to be re-lapped once married to the cylinder. More on this later but for now some more build pics.
After the liners were OD finished & rough turned inside all in same initial lathe setup, they all got returned to the 5C collet chuck for consistent finish boring. This was my first serious go with a tool post grinder. The plan was to bring them to within about 0.002" & leave the last for lapping operation to simultaneously true any egging + final bore + final finish. I wasn't super thrilled by the grinder finish but it looks worse than it is. Lapping removed the harmonic pattern very quickly so they are probably less than tenth deep. I have some ideas of how to improve the TPG with stiffer arbor. The grinding was quite accurate & process was controlled so that's important. In reality I need a 1/4 scale Sunnen hone LOL
Attachments
The laps I used are Acro brand. They are reasonably priced for what you get & the difficulty to make your own. The arbor is a steel shaft with a threaded end on the shank. the barrel is brass with double tapered ID & radial slits along the body. Basically you incrementally screw in the tapered end screw, it then expands the lap in the middle. I've tried different lapping compounds, they have their pros & cons in terms of how they cut, how they break down, how/if they embed. The one common factor is they are messy. You have to completely clean the part spotless in order to mic bore, rinse & repeat.
Attachments
So back to the saga. I was between a rock & hard spot. The cylinder bores were nicely finished with reamer. The liner ODs were nicely finished by lapping. But one of the two had to be altered in order to reduce correct ~ 0.002" of interference screw up. The finished liners were very hard to accurately grip on an ID type expanding mandrel & I wasn't about to re-make them. So I opted to lap the aluminum cylinders thing they will benefit by being trued better than the reamer finish & aluminum is softer so should go quickly.
So here we go again... another 45$ lap because its a different size. 800 compound grit leaves a satin finish but its very true. The slight shiny irregularity you see (looks like a scratch in the pic) is just the ball of the bore mic. Into the ultrasonic to remove lapping gunk.
So here we go again... another 45$ lap because its a different size. 800 compound grit leaves a satin finish but its very true. The slight shiny irregularity you see (looks like a scratch in the pic) is just the ball of the bore mic. Into the ultrasonic to remove lapping gunk.
Attachments
Last edited:
And finally, with proper OD/ID interference of ~ 0.0005", the cylinders went into the mini toaster oven at 450F for 5 minutes & then room temp liner dropped in place. Allow cookies to cool to room temp & then lap the assembled CI liner bores. This time I used a different brand (Clover) lapping compound 600 grit thinking it would cut faster & then I'd switch to finer with the last 0.0002" or so. This stuff has more oil content which offers benefits IMO. You can feel the cutting pressure better as opposed to drier compound which can get 'cakey' if that makes sense. It cleans off the surface much better with common paint thinner. The aggregate is a bit sharper & doesn't break down quite so fast. I cant detect any remaining lap embedded in any surfaces so I think that is maybe an issue you hear more abound with diamond.
I think I now have 5 complete cylinder assemblies.
I think I now have 5 complete cylinder assemblies.
Attachments
Last edited:
You are a patient man Peter! Beautiful work as usual.
Piston making time. They are made from 7075 aluminum. I love machining this stuff, to bad its spendy.
Picture shows an RC commercial forged aluminum piston from an OS-56 4-stroke alongside my tester blank. The OS is the same nominal bore as my 5-cylinder so I made a decision early on in this project to use the rings, at least initially, to give me the best chance of running success. Rings are finicky to make being about 0.043" square section. the tolerances are pretty demanding at this scale & rings require special fixtures & oven to heat set the opening. I am going to make rings now that I found a guy with an oven. Anyways I bought the commercial piston to mimic the critical dimensions: the OD, the reduced crown, the ring gap, trough OD, width etc.
Picture shows an RC commercial forged aluminum piston from an OS-56 4-stroke alongside my tester blank. The OS is the same nominal bore as my 5-cylinder so I made a decision early on in this project to use the rings, at least initially, to give me the best chance of running success. Rings are finicky to make being about 0.043" square section. the tolerances are pretty demanding at this scale & rings require special fixtures & oven to heat set the opening. I am going to make rings now that I found a guy with an oven. Anyways I bought the commercial piston to mimic the critical dimensions: the OD, the reduced crown, the ring gap, trough OD, width etc.
Attachments
The crown is slightly reduced which is the norm to ease ring stretch a bit when installing. The ring width was made with a Nikole grooving tool. Because I had to open the width by only 0.0010" I couldnt trust by carriage DRO so I set up a tenths indicator on the carriage after cutting the first depth, zeroed it, then moved the carriage over this time watching the needle deflection. Verify fit with feeler gauge stack that matched the commercial piston. The ring trough diameter was measured with blade mic.
Attachments
The pistons are 0.0025" undersized to liner so just requires careful finishing & measuring using the same micrometer. Protect the ring area with tape & part off. I also use tape to protect the OD from the collet grip when facing the bottoms to length.
Attachments
I decided to drill & ream the wristpin hole while the blank was solid. On my prior tester I drilled them subsequent to inner body material removed & didn't like feel of the drill & reamer breaking into the chamber & restarting hole on the other side. I couldn't actually measure out-of-square but just preferred this solid method, basically pecking 0.050", clear chips, add cutting fluid, rinse & repeat. I had a vise stop but just to make sure of center used an edge finder on either side, used DRO center & compared the digits. Wrist pins are 5mm reamed.
Attachments
I made a fixture to hold the piston with dummy pin to align. Then into mill to make the rod pocket. Clean, de-bur, recheck dimensions.
I installed a ring & did a push fit in the oiled liner, so far so good.
I installed a ring & did a push fit in the oiled liner, so far so good.
Attachments
Chicken lights
Forum Pony Express Driver
Very nice work!!
Rocker covers made from 6061 aluminum. I had 2 successful testers so needed 8 more for 10 total. Apparently optimistic that I would have no boo-boos along the way.
Trimmed stock, drill & c/bore M2 hold-down screws.
Forgot to mention I tried a technique I saw on Tom Lipton video. (Left hand picture). He stacked blocks & took a pass across all to save time doing them onsey-twosey. But this is an accident waiting to happen if there is even a thou difference in thicknesses; the vise will only tighten on the fattest one & others may come loose. Tom used an aluminum TIG rod on the moving jaw to squash & conform to this difference. I don't have those rods but I used ~ 1/16" diameter solder. Its much more malleable & takes up the difference quite well, very positive holding. But. I'm still not convinced I like this technique if you need absolute squareness. Its possible the stack-up gaps can vary a bit among the pieces & by the time you get to the end of row, it may not be 100% square to the first one. It has its place though if you aren't thou chasing.
Trimmed stock, drill & c/bore M2 hold-down screws.
Forgot to mention I tried a technique I saw on Tom Lipton video. (Left hand picture). He stacked blocks & took a pass across all to save time doing them onsey-twosey. But this is an accident waiting to happen if there is even a thou difference in thicknesses; the vise will only tighten on the fattest one & others may come loose. Tom used an aluminum TIG rod on the moving jaw to squash & conform to this difference. I don't have those rods but I used ~ 1/16" diameter solder. Its much more malleable & takes up the difference quite well, very positive holding. But. I'm still not convinced I like this technique if you need absolute squareness. Its possible the stack-up gaps can vary a bit among the pieces & by the time you get to the end of row, it may not be 100% square to the first one. It has its place though if you aren't thou chasing.
Attachments
Last edited:
5mm radius bull nose profiling. This is best done while the block is solid as its kind of an interrupted cut feel & not a lot to grip in the vise in certain orientations. I've learned to make some witness lines beforehand. The vertical depth is a bit easier to set because the cutter has kind of extended wings (not all cutters have this). Then graduate in from the fixed jaw side until the radius just kisses both witness lines. Then always set the part to the fixed jaw. About 0.050" DOC/pass. Rinse & repeat
Attachments
The recess has a deeper slot (10mm EM). I found things went better with a starter relief hole & plunge the EM mostly into that (as opposed to solid metal). Then I can take a deeper DOC. I don't really care about the side wall finish but it actually stayed pretty clean. The big issue in slot milling is clearing chips often or it can make a mess regrinding shrapnel. I vacuumed for the most part. Low pressure compressed air works too but chips go everywhere (potentially your face & eyes). I think I'm going to make a clear plastic shield with a small hole in it, poke my air nozzle tube through that so I can see what I'm shooting at. Maybe that will block the bits from coming back to me?
Attachments
Last edited:
Now the pocket milling, 0.500" depth with 8mm diameter EM. I first plunged each corner hole with X&Y locked. Then on the 4th corner once at depth, I ran the perimeter, conventional milling direction. The inner width is the more critical dimension as it has to slide fit over the rocker perch axle ears. The external corner points get blended to the bull nose profile with a Dremel come finishing time.
Attachments
Last edited: