31 Dec 2017

Happy New Year everyone!

Drove along the beach and saw people laying out in bathing suits – it was 59 degrees. Maybe they’re from somewhere that’s 50 degrees colder so this seems warm?

I decided I didn’t like the rigid mounting of the exhaust pipe. I don’t know how much the engine moves around but even if it’s a little, it’s flexing the weld which will eventually crack. A half-bracket was welded on and safety wire looped around so the pipe can move around when it needs to, while at the same time keeping it from resonating and thrashing about.

With the worst of the tuning hopefully done, I couldn’t help but add something admittedly a little childish. Let me explain: normally when the driver takes his foot off the gas, the ECU shuts off fuel. Trouble is, I found a thread where someone asked on a car forum, “How can I get my car to do that cool crackling and popping on deceleration?” How did I not know about this? Disabling fuel-cut is just a checkbox away, and retarding ignition timing at high vacuum (which happens only when coasting in-gear) is easy. This causes any residual fuel to be ignited so late in the cycle that it’s still burning as the exhaust valves open. Done right, it’s not loud but just as described above, a crackling and popping. If the timing is retarded too much, flames may even come out the exhaust, which is kind of a strange thing to want to do because, A. You can’t see it, and B. The cops can. I guess I’ll have to provide a video because right now, I don’t think I’m emitting flames but haven’t confirmed. 🙂

BTW, burning fuel in the exhaust can be taken to an extreme, at which point it becomes “anti-lag”. This is where fuel is purposely dumped into the exhaust where it burns and keeps the turbocharger spinning, which is all fine, except that it’s both loud as hell, sounding like a machine gun, and  it’s not good for the turbocharger. Not going to happen on my car.

30 Dec 2017

I had this week off but for many it was a partial work week, making otherwise-enjoyable test drives less so, so it was a good time* to deal with the inconsistent throttle issue.

Turns out I had two spare throttle bodies but one was a mirror image of the correct part, probably coming off the original Honda CRV donor engine. Anyway, I removed the one from the engine and stripped it down, leaving only the throttle angle position sensor (TAP) in-place. Playing with it by hand showed no stickiness or inconsistencies, so I was betting one of the removed parts had caused it – I would have lost the bet! The TAP sensor was temporarily reconnected to the ECU and the throttle shaft moved by hand. Sure enough, closing the throttle 10 times resulted in 10 slightly different readings, some differing by more than a full percent. Strange; the throttle plate, shaft, and bearings looked fine and felt fine, no slop, no stickiness, yet it just wasn’t repeatable, and cleaning didn’t help either.

The first thing transferred to the spare throttle body was the K-Tuned Hall Effect TAP sensor, and the readings returned to within 0.1% of zero every single time. Close it slow, close it fast, didn’t matter, it was very repeatable, proving that the TAP sensor was good and that the throttle body wasn’t.

Bolted everything back together and while in there, also cleaned the idle air valve – a known issue with Honda K-series engines when dirty. Next, my tuner buddy, JR, recommended setting the mechanical throttle stop at just under the target 1000 rpm idle, like around 950 rpm. The idea is that regardless what the ECU tries to do, idle cannot drop much bel0w the target. This does not rule out a decelerating engine with “momentum” blowing through this limit as well, but that’s a separate situation.

After much flailing about, I learned out that the Honda idle air valve is partially open at all times when controlling the idle of an OEM engine, no surprise I guess. The way around that was to remove its mounting screws and physically rotate it to close off the valve ports, which lowered idle speed to around 800 rpm, then used the set screw to push idle back up to 950-ish rpm.

After that, the TAP sensor was recalibrated for full open and closed, plus setting the idle air valve’s operating range. I’ve seen several recommended ranges but after again flailing around, settled for “-3%” at target idle.

After all this, went for a few test drives and it’s now close to an OEM-quality idle. There’s still a bit more tweaking to do but it’s refined enough that I can work on it when I want to rather than needing to.

*In the back of my mind was a concern that someday, this “slightly-sticky” throttle body might just decide to stick open at, oh, 50%. Having that much power unexpectedly pushing the car could end very badly. One time at Autoclub Speedway, I got a hint of what that could be like when braking for a turn and catching the edge of the gas pedal on my shoe without realizing it. Nothing bad happened but only because there was no wall or cars ahead of me as I “drove” straight off. It was a lesson I fortunately didn’t have to pay for (kind of like what I learned by overhearing a woman ask another woman “when are you due?”, and hearing her say “I’m not pregnant”). Anyway… a stuck throttle is mildly terrifying, especially in a light car with this engine. The problem was solved by trimming back the gas pedal to move it further from the brake, but still close enough to do heel-and-toe downshifting.

26 Dec 2017

First off, happy holidays!*

A few months ago, a small crack was found on the exhaust pipe. I knew I was pushing my luck thinking a pipe supported solely at only one end by a vibrating, shaking engine would ever last, and nope. Anyway, the crack kept growing and today I had the choice of ignoring it and enjoy driving it, or fixing it, but not both. Nuts.

The cracking wasn’t helped by the fact that there was some extra weight out at the end of the assembly (due to quickly adapting the old exhaust to the new turbo setup). No doubt that at certain engine speeds, it was resonating and probably shaking like crazy. In an effort to fix that, a flex coupler was added in addition to rigidly supporting the outlet end; we’ll see how it goes.

Note: I looked more closely at the close-up picture of the crack and see it propagated a lot further than I thought… more welding!

*My poor wife made the mistake of wishing someone a “Merry Christmas” and was informed that “We don’t celebrate Christmas and are just shopping!” Good thing they don’t expect presents…

10 Dec 2017

Drivability is now much better, though one loose end is when slowly transitioning from no gas to slowly accelerating. The “handoff” from essentially idling to accelerating still needs some work, as the transition still has a bit of a hiccup.

Anyway, my brother and I did a car show at our old high school. It was a typical hot rod show where few people do the work on their own cars and everyone walks around looking at 50’s-60’s muscle cars that after a while, all start looking the same. Worse, about the only people there were car owners, with virtually no crowd of onlookers. There was one notable exception to the cookie-cutter muscle car: the guy started with a USPS Postal truck, removed the body, then added his own frame and paneling, and calling it a “33 Riley”. The engine’s from a Nissan 280Z and fits the car’s character perfectly. He made everything himself, cut, bent, and riveted the paneling, added big drum brakes, complete with real wire wheels. He “wins” in our book, though we didn’t stick around for the expected win of no doubt some $150K muscle car that’s hardly ever driven. The Rikey builder said he’s put about 70K miles on it since finishing it.

9 Dec 2017

Drivability tuning continues.

While maintaining a constant low speed at zero boost, lambda (air/fuel ratio) swings back and forth quite a bit between about 0.80 and 1.2. A side effect is feeling an oscillation in the car, like someone’s pushing and letting off the gas about once a second.

The fact that it was oscillating reminded me that closed-loop lambda control has its own PID loop. Comparing the values with an old tune showed that the tuner had set the (P)roportional constant about 4X what I had and the (I)ntegral constant at about double. I suspect he just set it to something in order to get on with the main tuning. Well… that’s fine, but it leaves the car with yet another drivability quirk. It’s harmless, but it’s stuff like startup, idle, and partial-throttle performance that gets noticed far more often because 99% of the time that’s where the car operates. If I’d taken the car for a test drive, we’d have easily been tuning the rest of the afternoon, but since he doesn’t fit on the passenger side(!), I don’t think sticking around would have helped.

I’ve talked to other tuners and they said this is where all the time goes – drivability issues. Customers get all excited about loud pulls that produce dyno curves for bragging rights, but it’s the little, nuanced variables that complete the tune and make the car act, what, like a real car?

8 Dec 2017

The little fire turned into a beast which is still completely uncontained; a couple coworkers are waiting to see if their homes are still standing. Last night our pond skimmer worked overtime collecting tree debris. Imagine that as burning embers driven by 30-80mph winds.

In car news, it’s fairly straightforward tuning a fuel and ignition map, what’s not is cold-start and idle. It takes a lot more time because it’s so subjective, plus cold-start can only be worked on for a few minutes before the car has to cool off.

AEM’s ECU versatility allows changing many parameters, some by a lot. For example, the idle PID control loop terms appear as “X.XXX”. Annoyingly, they don’t give a rough idea where to start, instead sticking with “every engine will be different.” Well, yes, and no. For example, is a good starting point 9.000, 1.000, 0.456, 0.045, or 0.005? If you start with 0.005 when the majority of engines need 1.000, it wastes time. It’s like being asked to guess a number between 1 and 10,000 without even a ballpark starting point, or even the state where the ballpark is located…

I bring this up because idle is currently running an “I” term of 0.010. Today I found a post where someone got AEM to reluctantly cough up a starting point. The response was “… cars with aggressive setups and big injectors may need a ‘P’ term of zero and an ‘I’ gain of 0.1-0.2 at idle to prevent oscillation.” So, maybe what I thought was okay could stand further improvement. Increasing “I” to 0.05 didn’t show a marked improvement, but it needs to be driven, but now it’s time to pick up the wife at LAX. Told the neighbors that if things get insane around here before we’re back, to rescue our dog.

7 Dec 2017

Did some more local test drives; idle is slowly being improved.

During a drive, saw some smoke and headed to a local mountain for a better view. The picture is looking straight north so you can see how Santa Ana winds blow west,  opposite the normal direction. This one was later named the “Lilac Fire” and unfortunately started in a canyon running east-west, a perfect recipe for disaster. People living in SoCal don’t lecture people regarding fire danger; we all know Santa Ana’s don’t play favorites and any year could be our turn. Pretty sure this fire was responsible for the second emergency warning we got on our phones in as many days.

On the way home, “Baja Dude” in an enormous off-road truck tried to show off his gentleman tackle (as Top Gear would call it). We were in heavy traffic moving about 26 mph, so I’m wondering, “huh, really?”, with no choice but to ignore him. The group of cars we were stuck in proceeded to the next light, where a guy in a BMW SUV completely moved over on me – first time that’s happened but it’s inevitable. Driving something so low makes Midlana a candidate for such things and as long as it’s planned for (think: motorcycle) it’s not a big deal; just touch the brakes and let him slide over – 12″ in front of my nose. What was surprising was Baja Dude suddenly accelerating up even with the SUV, honked his horn and yelling at him. At the next light, the SUV owner got out and walked back to me, very apologetic. I appreciated the gesture and said “it’s okay, I’m used to being invisible in this”, but couldn’t help but notice him checking his rear bumper for damage…

6 Dec 2017

Headed off to Borrego Springs, mostly because the timing was right. The first picture is part way down the twisty road into town. Just short of the horizon is the Salton Sea. My brother drove out there a couple weeks ago and said the road was in such terrible condition that he didn’t recommend it; another time perhaps.

Poking around on Google Earth lead to making this the destination for this trip, the “Sand Dragon” that lives near town. This thing’s got to be about 300-feet long and note how it continues in the background on the other side of the street. The metalwork and creativity were very impressive. There are a bunch of other sculptures as well but I was worried I’d get high-centered on the dirt access roads.

The last picture is Midlana in front of an odd little building off by itself at the base of the twisty road leading back. Anyone recognize it? It was in the 2000 “Within” and “Without” episodes of “X-files”. In the show it was a school but unsure what it is/was; seems to be closed up and deserted. When I was a kid I remember it being a real estate office.

I tried recording the drive up the hill, but alas, again foiled by slow cars. I didn’t feel strongly enough about driving back down the hill and waiting for a clear spot in traffic because the Highway Patrol regularly runs up and down the hill was another reason to just let it go.

Idle definitely needs more work; it’s still a little erratic, sometimes stalling, and sometimes gets into an up-and-down surging cycle. Other times it idles perfectly…

5 Dec 2017

We’re having some of our “Santa Ana” winds that come up this time of year. I did a trial drive today and it was too bad I wasn’t recording – those of you far from the American South West would get a kick out of me having to avoid tumbleweeds blowing across the road, yes, just like in the movies. Anyway, the further east I headed the stronger the winds got. The winds are a funny thing, where west of the mountains and depending upon altitude, there’s nothing at all. But head up into east/west-running passes and it can be 30-50 mph just a few miles away. When driving, it’s more about the blowing sand than the wind itself, both for it getting into the engine and my eyes. As California deals with every year, the winds create the ever-present fire hazard as those north of us are dealing with right now. Anyway, if the drive falls through, maybe I’ll spend time in the garage refining the idle control a bit more.

4 Dec 2017

Have this week off (well… until Thursday, when I have to clean up the house, then pick up the wife on Friday). That leaves three clear days, to be used wisely.

Today was used to do test drives to deal with the remaining drivability issues: a stumble with getting on the gas, even from idle; a weird surging when coasting in-gear, and an irregular idle:

The stumble was cured by increasing wall-wetting, which cleared it right up. It might actually be a bit too much, because if I rev it off neutral a few times, I can smell gas. Easy enough to fix and drivability has been restored.

The surging when coasting in-gear was due to fuel-cutoff being disabled, leaving the ECU to fiddle with AFR even while coasting. Enabled fuel-cutoff with a combination of MAP and throttle position – solved.

The irregular idle… ah yes. This took the most time to mess with the last time, and even then I wasn’t entirely happy with it – “happy” would be a rock-solid idle that’s no different than an OEM car; frankly I’m not even sure if it’s possible. I worked on that an hour or so and it’s definitely better, steady-state at least. The PID loop needs more work because if idle is disturbed, the control loop isn’t quite there yet, but it’s now much more drivable. We’ll call it a work-in-progress.

One thing that’s going to have to be addressed is the throttle body. Though the throttle plate shaft is tight with no apparent free play, it’s inconsistent as far as where “0%” is. Open and close it 10 times and 10 slightly different versions of “0” are reported even though the plate is physically closed. With the engine off, if I very delicately let off the gas, I can get the throttle plate to sometimes stick open at 5-15% (it is the assembly sticking because the throttle cable is loose). It’s not as bad as it sounds because when the engine’s running, the vibration allows it to completely close, but that can’t be counted on, as in “it’s all fine until it isn’t”). Cleaning it unfortunately didn’t fix it. Not sure whether to buy another used unit and take my chances, or buy a brand new OEM unit just so I know what I’m getting. [Then again, maybe not… nearly $800 from discount sites, and they come complete with stuff I don’t need. All I want is the housing, throttle, and shaft.] Anyone have one in good condition? These things are getting valuable – when I get this one off I’ll see how it can be repaired. A “fix” might simply be another spring, but it would have to be an OEM type spring, a helical type due to how the shaft rotates.

Before anyone mentions it, I don’t really want an aftermarket assembly because they typically run oversized throttle plates for hp bragging rights. That’s fine for drag racing, but when finesse is needed, having a smaller throttle plate makes it easier to modulate power mid-corner.

Anyway, the plan is to drive out to the desert one day this week.


3 Dec 2017

Two steps forwards, one step back.

The tuning changes made a large and positive difference in power delivery. The tuner also changed the idle control, deceleration settings, and wall-wetting (compensation for fuel that ends up on the walls of the intake instead of the cylinders) along with many other things. Unfortunately I was itching to beat LA’s rush-hour traffic and declined his offer of a test drive… I should have done so. Now I get to do it myself via trial and error like before. I’m hoping I can reuse some settings from earlier tunes to save time – fingers crossed.

Before that was started though, the intercooler was removed, inspected, and cleaned. There was surprisingly little oil in it, with practically none at the outlet end and zero at the throttle plate. It “may” have been pushed through by the later dyno pulls, but it still raises two unpleasant questions:

  1. Where did all the smoke come from?
  2. Why didn’t the car start?

I think much of the oil likely flowed past the turbo seals (due to the theorized blocked drain line) and into the turbine housing. The housing was still hot  from the previous pull, which reminds me of something. While it was smoking like crazy and yet not starting, as soon as he’d stop cranking the engine, about 2 seconds later, there’d be a soft “pop”, and an extra puff of smoke would come out the tailpipe. I’m guessing it was engine oil/gas lighting off in the hot turbine housing. This happened 3-4 times and always after he stopped cranking, so there was definitely oil pooled in the turbine housing, so, I think this part of the mystery is solved.

The second question is harder since oil in the turbine section shouldn’t affect whether the engine starts or not. The tuner confirmed that it eventually started on its own, he didn’t touch the tune. Problems that just go away… ugh.

Anyhow, the next thing to solve is how engine speed takes a nosedive when barely engaging the clutch (like leaving from a standstill). It may well be insufficient wall-wetting. We’ll see.

In other news – don’t remember if I mentioned it – earlier this year my brother got fed up with his GoPro, replacing it with a Sony action-cam. I can’t say enough good things about it, and it’s mostly because the Sony, well, it just works – every time. I’m sympathetic; my GoPro 4 (with the latest firmware) still had demon-possessed moments. Starting a second video by itself, resetting itself, and my favorite, acting exactly as if it’s bricked, then works days later like nothing was ever wrong. I have to give GoPro’s Marketing Dept props for still selling even with pissed-off owners. Point being, if you’re looking for an action cam for Christmas, consider the Sony. Here’s my brother’s Virginia City Hill Climb video recorded with it, which has both mechanical and electronic stabilization – it shows.