28 Sept

To quote Yoda, “There is another.” While I fiddle around with ECU and dash issues, congratulations are in order for Jim Langan, who has Midlana #2 sitting on her wheels! Jim chose to go with the traditional Lotus Seven front fenders, which look great. There are currently about half a dozen builds under way, which you can check out on the Midlana Forum.

26 Sept

With the exhaust leak fixed, we need a new problem, and alas, one turned up. The AiM MXS digital dash has been working fine, except for a strange error reading MAP. The MAP value is sent over to the dash via the CAN bus, along with engine speed and several other parameters. They all work fine, but the dash claims that MAP isn’t the usual 22-24 kpa at idle, but instead, something lower. After maybe 30 minutes of run time, it’ll even display 0 kpa, which is a perfect vacuum… not likely. Meanwhile, the ECU sees a correct value for MAP, which is the exact same value being sent to the dash. That’s the old problem…

The new problem started several days ago, when oil temperature started to lie. Unlike MAP, oil pressure is a voltage, read by both the ECU and the dash. As soon as ignition is turned on (like first thing in the morning), oil temperature should be about room temperature, and it is – so says the ECU. However, the dash claims that it’s 51-52 degrees C, which is nonsense. However, as oil temperature rises, it seems like the dash starts reading the actual value once it’s above its false 51 degrees C floor.

Neither problem can be explained by a wire being loose. In the first case, since MAP is in the CAN packet, all other values should be messed up as well, but they’re fine. In the case of the oil temperature, if the wire was broken, the reading should go to max, and if the voltage is shorted, it should read zero. It does neither, and if it was shorted, it would affect the ECU’s reading as well, but it’s fine.

As a hail mary, I updated the firmware in both the dash and ECU – no change. To add insult to injury, the engine now gets limited at a lower rpm than before; I hope it’s my test tune and not some mystery feature of the new ECU firmware. It’s always something.

21 Sept

Odds and ends.

Replaced the exhaust manifold gasket; yup, not hard to see where the old one was leaking!

Finally took care of something I’d been meaning to finish up for a long time. A spare valve cover had been modified to remove the oil filter and dipstick (both being redundant due to the dry sump). Also shaved off the HONDA logo, which will be replaced by a MIDLANA logo. However, before the newly powder-coated part could be installed, all the blasting media had to be removed… there was more than I expected. I really like how it changes the engine compartment; it makes it look much more integrated and clean.

Finally, the air filter scoop was replaced with another of the same.

13 Sept

So close. The crack in the #4 primary was real, but turned out to not be the main leak. That ended up being down below the flange where the gasket material had blown out. I don’t blame the gasket itself; I knew that I’d been pushing my luck by reusing it several times, so a replacement and spare are on order.

Thought I understand PID closed-loop control, there’s still a lot to learn when controlling the engine. More early-morning testing is slowly dialing in the boost control loop, gradually getting rid of the overshoot while still maintaining fast response. Just as important as determining the P(roportional) and I(ntegral) terms is limiting the integral value to keep it from “winding up” the loop and causing overshoot. A little is okay, but I was seeing 20-30 kpa, which isn’t acceptable; it’s now down to about 10 kpa and I’m trying for a bit better.

Reintroduced throttle-position-dependant boost, which makes the engine act more like a V8 and less like a boosted 4-cylinder.

Also ordered a replacement hood scoop for the air filter. While I plan to someday make a proper carbon unit, the cheesy vacuum-formed unit had melted somewhat and cracked. Time for another temporary replacement.

11 Sept

Went over the exhaust system and found the probable leak hiding in plain sight, on top of the #4 primary at the cylinder head exhaust manifold flange. Maybe I’ll remove the header before welding, maybe not, depending on whether the TIG torch can be maneuvered into position. As noted in the Midlana book, one reason for not coating a header is for exactly this reason – if it’s coated, it can’t be repaired. It looks as though it’s been there a while, gradually growing. The header wrap probably didn’t help, either.

Regarding the oil O-ring failure last week. After it was repaired, the dry sump tank was topped off, with the thinking that it was probably about half a quart low. Added one quart, still low. Added a second quart, still low. Added a third quart, and it still wasn’t quite full. That was rather shocking, because it meant that either I’d let it get low previously, or I’d lost a lot more oil than expected. Regardless, it potentially suggests that the dry sump system may have saved the engine.

In other news, I got called for jury duty. In California, that happens as often as once a year. For the past two times, I never got called. The time before that, I got called but was excused (saying your an engineer usually displeases one lawyer). Before that, I was on one jury and the case lasted just a couple days. This time though, like watching a iceberg looming on the horizon, it was announced that there was a big case coming up that would last for weeks (the iceberg). Groups were first called for more normal cases, leaving a smaller pool of us, then an enormous group called for the major trial… (iceberg, dead ahead)… and boom, I got selected. As many of you know, it can’t be discussed until it’s over, but I’m pretty sure there’ll be a story to tell.

The downside is that it’s going to extend far beyond what our company pays for jury service, leaving me the fine choice of either using up vacation time, or “LWOP” (leaving work without pay). Such is the cost of doing one’s civic duty.

7 Sept

O-rings showed up. Looks like Summit Racing took a lesson from Mcmaster, shipping small items in big boxes and charging for it. (The weird picture formatting is due to them being from my cellphone.)

Drilled the Canton remote oil filter adaptor so that their fine swival design doesn’t swival the nut off again. With the safety wire added it should take this item off the issue list for good – bitch.

Having solved that, it left a vacancy for the next bit of drama: what sounds like an exhaust leak, but only one cylinder. It’s only audible under hard acceleration, so it should be a pretty short list: loose spark plug, leaking exhaust manifold gasket, or a cracked header primary. Since there’s another heatwave underway here I didn’t bother opening the hot engine compartment, but hopefully it’s nothing major.

The good news is that with the weaker wastegate spring back in, the closed-loop boost control is finally working well enough that I can stop messing with it.

Lastly, I’m finally moving into the new century and have been posting photos to Instagram (@midlana1). Sometimes I post impromptu pics there that I don’t post here, so please follow me there!