Playing around with the 2.7 software (ECU and TCU)

[RTB]

Well, if there is any improvement, it isn't much.

I still can't let go of the idea that they must do at least something good at low loads. A lot of trouble has gone into designing tuned intakes over the years on various vehicles, so seems reasonable there should be some level of benefit from higher velocity air creating that slightly improved VE at low RPM. Even on our non-tuned inlets....

So I left the flaps stock below idle. Closed from idle to 1400, then ramping open to 1800 RPM. (This RPM zone is used in a light load gentle acceleration for me, or light load low speed cruise. As soon as there is any real load, I keep the RPM up above ~2000 RPM to ensure adequate oil film stiffness) They will however go full open anytime the IQ is more than 30, as for me this is firmly in the acceleration range, so the boost is up and I'm thinking less restriction is better.

The other thoughts were that I don't want them stationary and potentially seizing. Also wanted to minimise continuous movement to reduce wear. Stock maps have the swirl flaps almost always moving. So now for me they move if I am in a gentle acceleration, or traveling at low speeds with light loads, but varying RPM. Full open on the fwy so no cycling as the speed changes there.

If you are a true optimist, then you might say there is a teeny tiny slight improvement in economy, but I've only done 2000 k's since the change. No long trips since doing it, mostly just across town, around local streets, or down to the boat. I've gone up to 265/65/17 AT's for the last 9,000k's too. The next time I head down to work will be a better indicator to see what full open does. Long freeway runs there.



Current Swirl Flap commands



or in 3D

[RTB]

Well, just did a tank after a work trip.

510 km of fwy which included about a 1.5 hours stuck in Westgate fwy traffic lucky to get out of 2nd. 301 km of 50, 60 , and 80 zones around home doing short trips.

So 811 ks out of the tank. Most ever I have gotten. That's 811 real k's too, not 811 fake k's from a factory speedo calibration with a 4% optimistic error.

So EVIC ended up showing 7.0 l/100 by the end of the tank (6.6 after the work section of the tank). At the pump, the tank was actually 8.4 l/100. Lowest numbers I have had to date. Especially on a tank which included fwy and local driving.

Only changes since last work trip are swirl flap mods per below, and some torque response tweaks to better match the TCU mods.

I really do wish we had an 8 speed box. Pity I'm not set up to mod the car enough to fit one in. I'd say an 8 speed would make it a whole new car.

[JeanLuke]

I have read somewhere that the swirl flaps are about improving the combustion process at low power. They effectively close off, or restrict, flow to one of the two inlet valves in each cylinder. (Jean-Luke: Special category is stating the bleeding obvious!)

I am not so sure the flaps are related to fuel consumption but more to particulate and nitrous oxide emissions.

My linkages came detached from the swirl-vane motor, so I have zip-tied them fully open. Access to the linkage means removing the starter motor. I figured I'd fix the linkage if ever I need to remove the starter. With the vanes fully opened (as opposed to before I zip-tied them they were fully closed) I did notice a drop in "launch" from idle. Otherwise, no difference, except the plumes of black smoke at high load stopped.

[RTB]

Yep, pretty much what I am seeing here. Mine does launch quit well, easily first off the line at the lights with normal part throttle. Just got the best economy ever, but that is an N of 1, so time will tell.

Regarding the fuel injection. Looking through some of the stuff MuzzWA shared, it suggests our engines have a dual injection (pre injection and main).

I can't see any of this in the ECU, and the main GC manual only talks of a single injection cycle.

Can anyone confirm if the 2.7 in the WJ has a single injection cycle, or uses a pre/main injection?

This comes from one of the MB OM600 series engine training docs. (Thanks for sharing MuzzWA).



From the GC manual and what seems to line up with the ECU.

[RTB]

Started looking at the Start of Injection (SOI) stuff.

Earlier in this thread I had a quick look at the SOI maps and it all looked a bit weird. There's a bit of a hole in the middle which had me thinking.

Well, I threw the relevant maps into an excel sheet which calculates all the map outputs in the same way that the ECU does. This has created a far more reasonable looking output for the real SOI.

Injection Quantity and RPM go into the rail pressure map to output a rail pressure. Rail pressure and RPM go into the SOI map to create the start of injection in degrees of crank rotation BTDC. Injection Volume and Rail Pressure go into the duration map to output the injector open time in degrees of crank rotation. From this we can calculate End of Injection crank angle (EOI) which is kind of important.

The base SOI map with the non linear data.



The actual SOI once all relevant variables are calculated. IQ on the top, RPM down the side. Degrees of cranks rotation BTDC in the middle.



I have the same outputs for duration, EOI, and rail pressure, so have more reading to do now on what works best for economy, and what works best for power. Very different from the old days of just giving the distributor a touch more of a turn and a mixture adjustment on the old XC Falcon and advancing the timing. Also seems making a basic knock detection system isn't that hard, so might consider that maybe. Seems our detonation frequency is 6.514 kHz. I doubt I will be brave (or stupid) enough to do anything that will require that though. Can't see me doing much to this map really.

SOI below just in graph form.



Thoughts?? Does it look like I have it all wrong, or does it seem reasonable?

[JeanLuke]

I do not think that first input map is in degrees BTDC. I can't see injection starting ~70 degrees BTDC.

Are you sure that is not in milliseconds or microseconds?

[RTB]

Yep, fairly sure. That map had the wrong scaling offsets from WinOLS in it. The SOI map I used to create the real timing map used a map with the correct offsets applied. Below is the WinOLS map with the correct offset put on it. Everything to do with injection is based on degrees crank angle as this is what the ECU measures in real time from the pulse.

There are some good videos on Youtube by vagecumap and phil cooper explaining the offsets for it.

Edit:- I've seen various comments on diesel injection timing, more specifically End of Injection. EOI should be no more than 20ish deg ATDC at most, in normal conditions. Does this sound reasonable.

Edit to the Edit:- Maybe the youtube teachers' information is not relevant to our WJ... See later posts.