Tag Archives: Solving the same problem over and over

Can’t stop, won’t stop, thinking about building the Sprint-T

Sometimes it’s a bummer when you get caught in a rut, trying to solve a problem that ultimately can’t be solved. That’s a little like what I’m dealing with on the Sprint-T. Absent a change in the build budget from zero the problem is unsolvable. But that doesn’t keep me from trying to solve it. I mean I have been working on this problem since 1968, why stop now? I’m as close now as I was when I was 10.

With the loss of access to the Subaru transmission the Subaru build got more expensive but still not as expensive as a V8 build, so right now the engine of choice remains the EJ20X at $1K. If I can find a decent Subaru transmission at a good price then I’m ahead of the game, if not I have to figure out how to get a bellhousing and starter to fit so that a good T5 transmission can be installed. Weight wide that would be the way to go, but building the bellhousing might be too much work without access to a 3D printer. I thought I might be able to hydroform a bellhousing, but while I could build the mold and baseplate I don’t have access to the kind of pressure needed to force steel or aluminum thick enough to work as a bellhousing into the mold. If I’m reading the literature right I would need pressures on the order of 3k-4k PSI (207-275 bar) for that deep a stretch, while what I have access to tops out at 1700 PSI (117 bar) That would work for about half the thickness I need in either metal, not good enough.

But this being DFW with lots of aerospace manufacturing going on I might be able to either get equipment to stretch 3/16” steel that far, and a quick Google search shows 2 local companies that can 3D print in several aluminum alloys or in Chromemoly steel, but at “not cheap” costs that are still much less than DIY hydroforming if I try to do it in one pass with materials that are thick enough to pass inspection at the race track (0.25″ at the thinnest point for aluminum, or 0.1875″ in steel). Now the problem here is will a 3D printed bellhousing even pass tech at any thickness. Tech inspectors can get pretty leery of unconventional manufacturing techniques especially on parts that are both structural and safety-related like a bellhousing. The bellhousing has to partially support the weights of the engine and the transmission, making it a structural component, but also has to contain the shrapnel of a potential clutch or flywheel explosion. That would mean making more than one bellhousing and then deliberately exploding a clutch or flywheel inside it and proving nothing made it outside the bellhousing.

Now something that used to be done back in the early ’70s was to make the bellhousing out of aluminum but then put a thin steel liner to contain the exploded bits using the aluminum to support it. That is the liner alone would not contain the explosion because it wouldn’t be strong enough while the aluminum was strong enough but not hard enough to contain the explody bits, so the liner prevents the bits from contacting the aluminum while the aluminum holds the liner together, and the pumps I have access to could hydroform the aluminum and the steel liner using the aluminum as the mold for the steel liner without removing it from the mold used to hydroform the aluminum. In other words it would be a two part process of putting the aluminum sheet in the mold, forming it to the mold, splitting the mold and putting the steel sheet in place inside the aluminum, then forming it to the combination of the mold and the aluminum. The tricky part for the Subaru engine is the starter bolts to a flat spot on the back of the bellhousing which has to be at a certain depth from the block and flywheel for the starter to engage, while the transmission also bolts to that same area and also has a distance from the flywheel it needs to be to function.. The added problem is that T5 transmissions were used in many applications with different mounting depths, so I would need to find a Subaru transmission case and cut off the bellhousing to get the starter depth right, then rig up a plate to mount the transmission at the right depth, fill in the gaps and smooth out the differences and pull the mold off that. The mold would not have to be super strong, just strong enough to hold together for two forming cycles, and the second cycle would put most of the forces on the aluminum outer layer.

And it’s getting late and I got extreeeemely wordy, so time to put the post and the writer to bed.

How do I bolt the body to the frame OMT

Yep, that issue again. Every iteration of the frame requires a new solution to mounting the body to the frame. This version is no exception. The issue this time is the frame rails and the body mounting flange literally have only a passing relationship, as in the frame rails pass under the front corner of the body for a few inches and they never see each other again. Now this might make for a great movie, but it makes for a crappy car.

This problem arose because to reduce flex in the frame I had to remove all the bends to get the frame to tuck up under the body. The mounting flange on the bottom of the body is 27″ wide at the firewall, spreading out to 34″ at the pinch of the cowl and then a nice straight run to 34½” where the back of the body meets the bottom 38½” behind the firewall. The topside of the body is considerably wider with the windshield base across the top of the cowl pinch being between 38 and 39″ depending on how you measure it, and the widest part of the body just a hair over 45″ right about the same distance behind the firewall as where the back meets the bottom. This results in conflict between straight frame members and running the frame under the body instead of through it. If this was a Locost instead of a bucket there wouldn’t be a problem because that car has vertical body panels that pop-rivet to the frame. On a Sprint-T the frame goes around the body so the body can be painted and installed after the frame is fully finished. The conflict arises anew when the ugly triad of light, strong, and cheap also conflict with “looks pretty”. So we have to make the mount light and strong, and cheap, and pretty as it runs from under the car to where the frame rail ended up, and what I’m getting is either light and strong but ugly and expensive, strong and cheap but heavy and still ugly, and everything pretty is either heavy, weak, expensive or all of the above.

The sticking point is the frame rail on the bottom runs from 30″ wide outside to outside at the front hoop to 48″ outside to outside at the rear hoop 58″ away, with the front hoop just outside and in front of the firewall and the rear hoop behind the rearmost part of the body. I tried to move the bottom rail in closer at the back but the diagonal that runs from the front hoop to the bottom of the rear hoop runs into the body. So far the best compromise is 0.060″ wall 1.5″ square across where the back meets the bottom of the body and forward under the body until it meets the main rails, and welding the floor to the top of that. It’s ugly with a Capital UGH and the aero is Not Good, but it’s strong and light and reasonably cheap.