30 Nov 2017

Get yourself a coffee – it’s a read.

I feel comfortable changing most engine ECU parameters, ones easy to monitor and correct, like idle, fuel maps, sensor calibration, and various safety features. On the other hand, cam and ignition timing scare me because there’s no easy way (on the street) to know what’s right and what’s too much. Images of holes in pistons from detonation, or valves hitting the pistons come to mind, so those were left as-set by the previous tuner.  Reasons for the tune review include: never finding root cause of the engine failure two years ago, having changed cams and the turbo, adding a much larger intercooler and cooling ductwork, and a general insecurity that maybe I or the previous tuner overlooked something. Another thing that’s always puzzled me is how other people with nearly identical setups often report having about 20% more power with less boost. Given the high cost of rebuilding the engine, the peace of mind to have the tune checked seemed worth it. Church Automotive in Long Beach is about the best Honda tuner, and is also well-versed with the Infinity ECU, so I got to enjoy two-hour tows in both directions through Los Angeles’s finest traffic.

Once the tuning started, Daniel, the tuner, found that cam timing had been locked and asked why, given that it’s a fundamental reason why Honda K-series perform so well. I didn’t know and hadn’t touched it. He also found that ignition timing was off. The results of his changes were striking, finding 10-30% more torque between 2000-4000 rpm and 15-20% from 4000 on up. The improvement was enough that we mutually agreed to reduce boost to 15 psi, where it still produced 410 ft-lbs torque on 91 oct gas and 450 ft-lbs on E85. Horsepower was 475 on gas and 530 on E85. (With it as-is, it’s the same as running 1000-hp in a stock-weight Honda, given that Midlana is roughly half the weight). Even though DynaPacks and roller-type dyno readings differ by about 20% (which is nuts, but whatever), the fact that the car left with about 20% more power than it came in with, at lower boost, is what matters. The tuning session went almost flawlessly – except for one rather large event.

After completing a pull, the engine was shut down while various tuning changes were made, but then it wouldn’t start. At first I assumed that he’d “broken” the tune, but it eventually started, albeit with epic amounts of white smoke pouring out the tailpipe (I was too stunned to take pictures). Right then I thought I was looking at another engine rebuilt, but what didn’t fit was that it had completed the previous pull with zero problems – how did it break while sitting still? While it was running, we saw evidence of oil leaking around the turbo, so then I thought the turbo seal had failed, which was strange given that it probably only has 5000 miles on it from brand new. See second pic below of what it barfed up on the floor…

The air filter was removed and there was oil on the compressor side, which meant the oil hadn’t originated from within the engine. Curiously, Daniel wasn’t concerned, saying that he’d seen this before on other cars and asked if I wanted to proceed. Not understanding what happened really made me want to say “no,” and if it was anyone but Daniel I would have, but given that he knows way more about Honda engines than just about anyone, I uncharacteristically agreed. He restarted it and sure enough, the smoke dissipated. We were able to successfully complete the remaining two hours of tuning without incident. So what the heck happened that it magically fixed itself? (I work in Field Support and things that magically fix themselves are always a mixed blessing.)

Oil coming from the turbo meant that it was unable to drain. The drain line gravity-feeds into the pan, which is how virtually all turbo engines are set up. With this engine though, it has a dry sump system, where scavenge pumps vacuum oil from the pan and return it to a remote tank. The combination of gravity-return and the vacuum should always insure all the oil is returned to the tank, but for some reason it wasn’t (in addition to the hose being a large -12 line). Also, keep in mind that this exact setup has worked without issue for two years – so what changed? I have two theories, both involving the oil return hose:

Theory 1. The oil runs about 80-85 deg C (and the turbine housing, which is right next to it, gets way hotter). Both likely soften the rubber (steel-braided) return hose a little – or maybe a lot. Because the dry sump pulls a vacuum on the crankcase, could it have sucked the hose flat? Possibly related, all the other hoses on the inlet side of the dry sump pump have coiled springs inside to prevent collapse… but I didn’t think it through, that the oil return hose needs one as well, or it needs to be switched to a different type of material that isn’t as temperature sensitive.

Theory 2. Perhaps I inadvertently build the return hose with a “flap” in it (a known issue when constructing AN hoses, where the cutter on an AN-type end fitting can sometimes go into the hose slightly off-axis, resulting in it slicing off a slab of rubber hose material, which can potentially hang out in the oil flow. Maybe, combining the already-soft rubber and the vacuum allowed the flap to lift up and obstruct the hose? An inspection showed nothing, but with one end being a 45-deg fitting, I couldn’t see through it. A piece of welding wire was used to probe the hose assembly but failed to find anything. Still, one way or another, drain hose failure seems most likely given the evidence, so it was replaced with a Teflon hose assembly. I can see why people with big budgets use it; it’s noticeably lighter due to its smaller OD, even though the ID is the same (see picture below). This should absolutely rule out the return line as cause and if it happens again, the turbo seal is then suspect, but how does such a thing magically fix itself?

One thing I don’t know is where the oil was burning. Smoke came out the tailpipe, but it could be burning in either the engine or in the turbocharger turbine housing. One nagging thing is, what caused the engine to not start? (I forgot to ask if Daniel reloaded an earlier tune in order to get it to start, so maybe there were two separate things going on.) Anyway, knowing for fact that the oil did come from the turbo is a relief (relatively speaking) so I think upgrading the drain hose is the best first step.

As an aside, with the engine running on the dyno, it gave a chance to see how well the intercooler ducting works. The inlet-side end tank was nearly too hot to touch, but the exit was practically room temperature.

Daniel said that this engine can reliably make 700 hp as-is, just turn the boost up – I can’t imagine. Shaun Church, the owner, said that they’d installed superchargers on Ariel Atoms, bumping power to around 500hp, but noted that their lap times typically increased due to them becoming a real handful. He also said that even over slight rises, they were lifting the front wheels off the ground, and asked if I plan to add a wing. Coincidently, my brother and I are currently considering just that.

I’ll work on the obligatory video next.