Oh yeah, almost forgot. Two years ago at Laguna Seca I was having trouble passing the 104db sound limits with the side exhaust. So last year I put together a Laguna Seca specific exhaust; entirely different config with downpipes going down the center tunnel and merging just behind the transmission, then 3″ all the way out back, small resonator in the tunnel and then the largest Magnaflow muffler I could find under the trunk space. I didn’t end up going to Laguna Seca last year so I never got a chance to use it.
Always improving:
Local autocross is a strict 91 db limit, and with Miata Reunion right around the corner it was a good opportunity to make sure this exhaust was doing the job. No issues during the autocross. I daresay I could probably even run this car on a normal 92db sound limit day up there!
This config is obviously much heavier than the usual side exits, so those will be back for GTA Finals at Buttonwillow.
So I did some autocrossy stuff just to shake things down and see what wants to fall off. The answer turned out to be absolutely nothing, ran flawlessly. It was pretty goofy trying to tip toe this thing around cones in a bumpy lot with it set up for the completely opposite end of the spectrum. Only a few short spots I could really use throttle but for those brief moments, oh man, it’s moving.
Testing at Buttonwillow showed the fuel starvation issue has been sorted with the new fuel surge tank and external pump.
With that out of the way, I’m able to peel back a layer and find the next issue that was buried a bit deeper. It seems I have a communication issue between the ECU and throttle body, which ALSO sends the ECU into limp mode, just like the fuel starvation. This helps everything make a lot more sense because at GTA Finals we were scratching our heads because sometimes it would go into limp mode under one set of circumstances that we were sure was fuel level related, and then other times it would go into limp mode under circumstances that couldn’t possibly be fuel level related. With one out of the way, it was a lot easier to find the second. In the morning it would go straight into limp mode upon startup or after a short bit of driving, then after a good ol’ code clear and tug on connectors and wires by hand, it started working. Ryan Helmuth, another LFX Miata head, was at the track the same day and when I had the issue narrowed down to ECU, TB, or wiring, he very awesomely offered to loan me his throttle body for a session to see if that made a difference. I ran a session with his TB, then a session with mine, both successfully, so I feel we’ve crossed that off the list and I think it’s down to the wiring and factory connector. I need to try to find that factory connector in fresh form so I can run new wires from the ECU with proper strain relief, heat protection, etc. and pin a fresh connector at the TB, then go test again.
Off to AZ to take the car to UMS Tuning. They weren’t the closest option for me, but that wasn’t my priority. I wanted to put the car in the hands of someone who I knew understood what I needed and would do the job right. Tony knows we aren’t looking for a number to brag about or just a full-throttle tune to get me down a drag strip. It needs to drive smooth, be predictable at partial throttle and in transitions, and be reliable.
I had already been consulting with Tony on configuration details, tapping into his experience with tuning modern GMs so that the ECU would be happy with everything and give us the best shot at all my hard work translating into positive results. So I towed the car 300+ miles through 110° F Arizona summer to get Hyper in the right hands. With swamp coolers, temps in the shop were a humid ~90° F which is about worst case scenario for what we’d see on track.
Hyper was on the dyno all day. Like, 9:30 AM to 9 PM. Lots of time spent on details, playing with cam timing, swapping things back and forth to see what worked best, different pulley sizes, etc.
Now I know EVERYONE wants numbers, but that is going to remain confidential. I’m very happy with the results so far. I’m considering where we’re at now ‘stage 1’, with two pulley sizes available dubbed “extra conservative” and “conservative”. We’re up against a fuel flow limitation on the current hardware with E85, so cannot run any more boost until we sort that further. Once we do, we can throw a tad more boost at it for a ‘stage 2’.
@ 8.5 psi of boost the fuel juuust begins to fall behind. Commanded rail pressure is 20 mpa (2900 psi) and that tapers to 15 mpa in the last 1,000 rpms. Doesn’t appear to be a shortage on the low pressure side because pressure there is 120 psi tapering to ~95 psi. When we checked that low side pressure we were really surprised to see it that high. The Corvette fuel filter is supposed to be pulling that down to ~60 psi, so it seems my filter is bad. That’s getting replaced this week. But at any rate we weren’t wanting for fuel on the low side. We have some curiosity about whether the LF4 HPFP might be under-driven by the LFX cam, but with the difficulty of swapping them, it wasn’t something I could do while we were there. But, the fact that the LFX HPFP operates at 15 mpa, which is what we were tapering to, up top, makes me doubt that we’d see any better results on the LFX pump. I could be wrong about that, not sure yet.
I’ve ordered a LF4 exhaust cam (on national backorder) to check it against an LFX cam whenever it finally arrives. I’ve also ordered one LF4 fuel injector so see if it fits the LFX. We’ll figure things out, just need to work through the variables. At this time I wouldn’t suggest the LF4 HPFP to anyone until we know more.
I will say that I had been expecting a rather peaky setup with the Rotrex, going by what I am familiar with on BPs, but the setup has exceeded my expectations there – it retains a very flat torque curve similar to naturally aspirated, just with much more torque than before. Horsepower grows linearly with rpms with the largest gains up top, just like expected with the Rotrex. This is exactly what I was aiming for: all the good stuff about the engine package we have, just turned up a notch or two.
To avoid the pitchforks and riots I’ll give this away: we’re currently putting better power and more torque to the wheels than a Ferrari F430 on similar dyno.
Since we have Miata Reunion @ Laguna Seca coming up where I’ll need to run the big muffler, I have the alternate down-the-center rear exhaust on the car. One of the few times I get to hear it sounding more like a traditional V6:
Can’t wait to get in the driver’s seat and start shaking it down. First up is a local autocross this coming weekend where I can make sure nothing major is falling off the car and things are operating as expected, then I’ll be scheduling some track testing time in the next few weeks.
In Miata terms, we’re making good power. But we’re the underdog by a huge margin against the cars we compete with. No doubt, we make the most with what we have, but with the largest Time Attack event moving to Circuit of the Americas where horsepower reigns supreme, I knew we needed to turn the wick up. We’ll always be down on power compared to the other guys, but that doesn’t mean we can’t give David a bigger rock against Goliath. However, save for some transmission hiccups, reliability has been one of our strengths and I did not want to sacrifice that. I had an idea.
This didn’t happen overnight. I’ve been working on the solution all year in the background.
A year ago, I placed a Rotrex on a box beside my spare motor, and got started:
The first time I took this C38-92 out of the box I realized the challenge I had ahead. I’m used to the C30 units used on Miatas, and had that size in my mind when I was initially looking at the space in the engine bay, placement, etc. The C38 is a behemoth. It would turn out that packaging would be one of the largest challenges at just about every step in this project.
In July I posted a teaser, focusing on the new rolling engine lift. But a few people noticed the blower hanging on the side of the motor…
That was after months of work on the bracket, and the final is actually version 3 after scrapping two prior. That gets expensive when you’re cutting 7000 series aluminum, but it has to be perfect. Clearances are mere millimeters everywhere, and rigidity of the bracket is critical to avoid having belt issues.
The frame rail had to be cut substantially and reshaped, with care taken to add new internal structure to ensure loads from the suspension are still transferred through the frame rail appropriately.
The subframe also had to be cut!
Did the math on pulley sizes and picked a few final candidates to test. We will not be spinning the blower to its redline, no running on the ragged edge here. Then ordered lots of belts and worked out a suitable belt for every pulley:
The packaging challenges continued with the coolers. I needed an intercooler and a cooler for the Rotrex fluid in addition to the engine oil cooler. I also wanted the shortest charge piping possible, made a bit more complicated with the throttle body at center top of the V6 rather than on the side of an I4. Off the shelf intercooler options weren’t going to work, I was going to have to make it. I chose a Vibrant core with the right dimensions for what I had in mind:
To hang everything in place I had to toss the old radiator mounts and make an entirely new frame to hold everything tightly together:
I’ve never made intercooler endtanks before. This project was going to give me a lot of practice on the TIG welder. Drew up and cut pieces to make up the endtank:
The top tank took a while with a tube hand notched mid-bend to merge into the curved tank:
Finished intercooler after quite a bit of welding.
Welding the tanks to the core was a next-level learning experience; you’re welding a butt-joint with the core quadruple the thickness of the endtank, and the core by its nature is trying to suck all the heat out of the weld as you go. Rather proud of how it turned out.
New coolers in new locations meant making all new lines:
On to the intake, and more packaging challenges. I did the math on filter size for expected flow. The filter I had was big for a Miata. The new filter is massive:
The intake tube for the C38-92 is 3.5” and that does not want to fit… anywhere.
More fun welding:
In terms of sensor and component locations, there was some debate about what locations would work best. I made a couple different configurations for charge piping to try the options:
Using a TiAL QRJ blowoff valve, configurable for either recirculating or vent to atmo. Which MAF sensor location we found worked best would determine which config we wanted, so I did both VTA and recirc setups:
The recirculating config uses a hard line that was fun to bend:
The Rotrex points the air inlet straight back, directly at the downpipe. So inevitably, the silicone bend is riiiiight by the exhaust. To give the coupler the best darn shot at nice long life possible, both coupler and downpipe are wrapped, and then separated by a titanium shield with a reflective layer on the hot side:
Crankcase ventilation becomes critical when adding boost. A direct injection engine really benefits from catch cans even naturally aspirated, so it had been on my to-do list for a while. Since this is a race car only I simplified things a bit. Eliminated the PCV, drilled out and tapped both valve covers, and ran lines to a pair of catch cans that vent to atmosphere. With so little space in the engine bay, I ended up placing the cans in the hole that the HVAC system pulls air from on a street car:
The fuel side is tricky. Direct injection can’t be easily upgraded with aftermarket parts the way a traditional fuel injection setup can be. Expecting that we might find we need more fuel than we could flow on the stock system, I preemptively bought a high pressure pump from the Cadillac ATS-V (which runs the twin turbo LF4 engine), to see if the pump would swap in on the LFX:
It is almost identical. The only difference is that the plunger is ~0.100” shorter.
After some measuring we felt pretty sure that the lowest point on the cam lobe that drives the high pressure pump was tall enough to maintain contact with the LF4 pump so we swapped it in on the motor in the car. That is not a fun job. Think coolant reroute but with the engine even closer to the firewall and more stuff in the way.
Colder spark plugs were in order, Brisk Silver RR14S:
t’s all a big paper weight without tuning. That is outside my wheelhouse, and I wanted someone who really knew their field and would give the car the time and attention it needed. For this I teamed up with UMS Tuning. I consulted with Tony through the final stages of the build to ensure everything was configured the way he felt would get us the best run at all things working smoothly. With everything done he made some tweaks to the calibrations of a few sensors on the old tune for startup. It fired first try and idled pretty well.
Pulled it out of the shop, first time sun hit the car since February!
Then loaded up for the 350 miles tow to Arizona to get tuned.
Working on intercooler, routing, etc. for the past several days. Making my own endtanks, which is a first for me, but required with how tight everything is.
Filter comparison, old N/A intake vs. new filter. Much more bigger.
Currently working on the new resident in the nose. Lots of work, had to fab a new frame to carry the rad, intercooler, oil cooler, and rotrex cooler all nice and tight together. I dug through just about every brand’s catalog for intercooler core until I found this Vibrant core that perfectly fits in this funky vertical config I dreamed up. Piping length is crazy short; ~15″ from Rotrex to IC and ~17″ from IC to TB.
Some noticed the easter egg hiding in the last picture I posted; a Rotrex supercharger hanging off the side of the LFX.
That’s a C38-92. The biggest one they make. Massive packaging challenge for everything involved. Have been working on this for a while.
C38s max at 90,000 rpms. I’ve done the math and chosen three pulley sizes for testing that are all in the “safe” zone. The smallest of the three still spins the Rotrex at a safe margin below that 90k at the engine’s 7200 rpm redline. There’s no reason to run it for all it’s worth and risk damage to the blower. The C38-92 flows some serious air and I expect I’ll be getting all I want without needing to go for broke.
Rotrex specifies that the oil should not exceed 176° F at the inlet. I’ll have a temperature sensor at the inlet and the Racepak programmed with a warning. (re-purposing the same temp sensor that previously measured power steering fluid temp, now that the blower has superseded the PS pump).
The LFX is 11.5:1. I consider the E85 mandatory for that reason. Given E85, hot damn that compression should make this killer.
There are a lot of reasons I chose this particular path. Right now I’m equal parts excited and apprehensive since there are some unknowns in the mix. It’s a lot of work to do without knowing how it will all work in the end, or if it will even work well at all. But, I have faith in the recipe I’ve dreamed up. Ultimately the same reasons I chose the LFX in the beginning compared to some other ways to make similar power (big turbo or V8) are what are behind this choice. The strengths of the LFX and the Rotrex should complement each other very well.
Haven’t set any power goals… seems a pointless exercise. Venturing into the unknown here. So few examples of boost of any form on this motor, and zero of them on E85, or with a Rotrex, or that are track-focused. There’s a lot of question marks. Going to build things ‘right’ on paper and then go see what’s what. Could say I want “X” number but that really means very little… ultimately it will make what it makes. Will be beginning with three pulley sizes to test. Very likely that the limitation will be fuel flow with E85 and direct injection, but there’s only one way to find out.
Fresh 22k mile MV5 transmission going in. One of the few times I can say something cost less than the Miata equivalent.
Went to drill and tap the case for the temperature sensor and turns out that unlike the MV7, the MV5 has a temp sensor just below the fill port, very similar to the place I added a sensor on the MV7. One 1/2″ to 1/8″ NPT adapter and the sensor’s mounted.
Need a different driveshaft/adapter and different crossmember to fit the MV5. I have all that here. The reason for the change to MV5 is simply availability and cost. MV5s are everywhere. So long as I’m on a factory junkyard trans I’d like it to not be unobtanium, in the event that I need to find one quick. Also, I can have one in the car plus a spare on hand for the same price as one MV7.
Time to get cracking. Now with GTA Finals @ Buttonwillow in November followed by Superlap Battle at COTA in February, I’m shifting to a winter-centric test/race schedule with a break in early summer. It’s been refreshing, but it’s time to get to work. Lots of stuff on the to-do list and time is going to start flying.
Before doing anything on the car itself, I had a side project to tackle. So far, removing the drivetrain required a lift; drop the subframe onto a dolly and then lift the car up and off. I wanted to be able to drop the assembly with the car sitting on jackstands, making it something I could do anywhere.
Started with a Harbor Freight motorcycle lift. For stability I made four legs with casters on each end. The legs are removable so the lift can store easily. Then built a simple base on the lift that bolts to attachment points that I added to the V8R front subframe.
Works like a dream. Can drop the assembly out of the chassis, roll it around the shop, lift it up to work on it easily, etc. It’s stable even with the suspension and transmission attached.
I’m using our experience and data gained from COTA to guide the development plans this year, with the primary focus being a return to COTA in 2020 with significant updates. I’ve already jumped into a couple of those projects but some things will take time and there’s always the matter of tuning and revising that follows any major changes, so for now I’m leaving the calendar clear for the next few months. Definite plans to be up and running strong for Miatas @ Laguna Seca and then GTA Finals @ Buttonwillow, but those are both later this year.
