bugblatterbeast

Edmundu passed some words of wisdom to me a while back about AVCS.... To put it simply, don't spend too much time dialing it in on the dyno. You will get some gains with the first few adjustments, then the returns get smaller quickly. You can spend a lot of money on dyno time trying to get the absolute best AVCS map and find that you were chasing a shadow because you didn't account for all the confounding influences during the test. Just get it close, then back it out a little. You go downhill real fast with a little too much AVCS.

I see. I do have a WBO2. I was just curious to see if anyone had tried CL targeting all the time. Most of my data logs have the two sensors in agreement, so long as the pressure in the manifold is low. When the exhaust pressure is high, the stock sensor reads extra rich (as expected). My hope was that it would be possible to eliminate discontinuities when transitioning from CL to OL. I've been carefully adjusting the primary fuel maps to match the OL/CL transition points. The method works but is tedious. I was hoping to find a short cut...

has anyone considered/tried moving the front 02 to the rear 02 location + keeping the ECU in CL targeting at all times? The Near WBO2 should be pretty accurate if it doesn't have to deal with the large pressure variations pre-turbo

yes. I've tried experiments with the AVCS map for economy. you can get good economy with 0 AVCS but you need a bunch of timing advance. you can get slightly better economy with higher avcs but less advance (I did the experiments over a 8 hour drive where the settings were altered every min to randomize the experiment). With the AVCS hole you get all sorts of weird drivability issues as you transition in and out of it. Another thing to note is that the economy goes downhill fast if you have even a degree or two too much AVCS.

is the ECU OL all the time with all 0s? I thought load/rpm thresholds had to be exceeded before the transition would happen. my datalogs indicate that the ECU is CL even with all 0s. It just doesn't delay the transition to OL when you stomp on it

if you set them all to zero won't the delays be matched? on another note, can anyone explain why there seem to be two fuel economy peaks for the lower load cells? I've found that I can get very good economy with 0 AVCS and a bunch of advance OR moderate AVCS and much less advance (~5 deg less). My running theory is that the 0 AVCS setting is self EGRing and the extra timing is needed because of the charge dilution whereas the moderate AVCS point has better VE and compression. So far the economy peak is at 6.8L/100km at 105 kph

If you run out of space, I believe there is a way to make additional changes to the basemap using enginuity. Alternatively, once you figure out what you want, you could just use enginuity to generate the entire basemap.

that's the funny thing. I thought I had picked a good spot (top of bellmouth). I think it was just bad luck combined with my attempt to log the startup transient. I figured the water couldn't make it up the up pipe on startup but I was wrong. I think it is safer to simply start the vehicle without letting the sensor heat up.

If we are going to be picky about the details, both the heater and pump cell need to be closed loop regulated, not just on . I agree that the sensor is OEM in a bunch of other cars. That doesn’t mean it is appropriate to run it full time in our application. When the sensor was designed into a bunch of OEM applications, the engineers went about measuring the operating temperature of the sensor tip under quite a few conditions. They also made sure the sensor was not installed aft of any large dips in the piping. Our application is probably ok but may have issues in the winter because of the large amount of piping hung below the sensor. I’ve already lost one sensor to cracking of the ceramic element. I’m guessing water pooled in the headers. When I started the car, I didn’t fire it up immediately so the sensor tip had a chance to heat before it was nailed by a slug of water. Based on your location, I assume you don’t have MMT laden fuel (has been banned for use in the US for years now…. but not for production and export). We get a lot of O2 sensor failures north of the border because of MMT deposits. WBs are particularly sensitive due to the little metering hole on the ceramic substrate. It only takes me 2 mins to yank the sensor so I’m not averse to doing so.

due to NAFTA the gas we have here often has high levels of MMT (thanks Ethyl corp ). I try to minimize sensor aging by only using it during tuning.

....part of the reason the WB only gets installed during tuning....

the stock O2 is reacts to the partial pressure of CO in the exhaust to get an AFR estimate. Because it is placed before the turbo, it will also react to exhaust gas pressure. for a given (rich) AFR, it will read richer if the pressure is higher. you need to becareful if you are using its reading as a leaner mix will show up richer than it really is if the exhaust manifold pressure is high.

might want to add a WBO2 trace to the log. The stock sensor trace is saturated so there's no telling exactly how rich the car is running

I'm running 11.3-11.5 at loads of 3.2 (last row of my map, which does need more rescaling) At lowish to part throttle, high RPM I drop into the 10s, drive the WGDC way up and pull timing (something I'm experimenting with right now, it seems to help keep the turbo lit during shifts.... but it could just be placebo). Just below WOT at high rpms I'm running around 11 to keep the EGTs down during continuous high speed runs

If the rotor was originally engineered to have holes through it there isn't a problem other than pad wear. On the other hand, if a blank rotor is simply drilled, you end up with stress concentrations around the holes that the original designer didn't account for. In some cases there's enough margin and the rotor doesn't crack. Other times, there isn't and the rotor does. I guess it is a matter of personal risk tolerance. Are you willing to bet that the guy drilling holes through the rotor knows more about the design than the guy who ran the finite element analysis...

I installed a 245lph pump on the weekend and found the probably reason for the difference between the scale values I got earlier and Edmundu's. Before the pump swap the injector scale I need to center the trims was 37xx. After the swap it was 36xx.... same fuel, same day. If anyone is installing a 245lph pump in an early 05 Canadian vehicle, you need to cut the 5 ribs in the holder back ~7mm. The center part needs a small trim too as the pump doesn't have the same cut-away as stock.

does anyone know off hand what the knock detection rpm and load range is set to on the Cobb 1.16 stage 2 map?

I remember seeing instructions on one of the Cobb forums on how to re-install in the event of a computer crash. The PC side SW is hardware locked anyway so I don't see them having a problem with people installing on more than a single PC.

I did a bit more testing of the DW740 injectors this weekend and found some interesting tuning quirks. I don't think the effects I noticed are limited to DW injectors, more likely they will be present with any large injector. It looks like the 1st order injector approximation (latency and scale) used by Subaru breaks down when the injectors get large enough. An additional 2nd order term is probably needed once the injectors get large enough. If I extract flow vs pw using only the 0th and 1st order terms I end up with either errors at high flow or at idle. The computer's AFR trims are able to correct the error but there is a little annoying transient between idle and higher loads while the trims are integrating. Centering the value makes the transient tolerable but I ended up adjusting the terms to keep higher loads happy and fudged the bottom end with the intake calibration. If the intake has ever been tuned on the vehicle, this quirk will never be noticed as both the intake and the weirdness at narrow PW get taken care of at the same time..... just something to keep in mind while chasing down weird fueling problems.

I've found that 5% boost below 29 MPH works ok. If you set WGDC high to hit your targets in 2nd and WGDC low to keep things in check in 4th, 5th usually comes out ok. The 5% boost pegs WGDC in 1st, which isn't much of a problem as boost never builds up fully till the end of 1st anyway.

the two values are a boost multiplier and threshold speed. below the threshold, your boost targets are scaled by the multiplier.

thanks. are there any gotchas using that method? I assume the process is 1) dump everything to a base map 2) flash with Cobb's package 3) dump an image 4) modify the tables with enginuity 5) dump the modified image back to the ecu. do I need to do anything special to the image to keep ST happy? It really sucks that there are no trusted protuners in my area

ah... makes sense now if you don't mind me asking, are you using Enginuity on top of ST to modify the tables we can't get to with ST?

do we have access to the coolant comp tables through some sort of update or patch? I'd really like to tune out the hickup I get on warm restart. When the car is stone cold it fires up perfectly. when it is hot it is fine too.

this probably doesn't apply to our vehicles, but I've found a few where it helped: 1) Bleed once, then go and use the ABS. 2) Bleed again. Sometimes a big bubble gets stuck somewhere in the abs system. on older Honda's and Acura's I'll usually open the high pressure port (use the right tool or you'll hurt yourself) and cycle the pump a few times. 3) it sometimes helps to tilt the car when bleeding 4) it also often helps to give the calipers a good tap to dislodge bubbles.