FI/E85.... things to watch out for
 

I thought I'd put together a post highlighting some of the things i've struggled through while configuring my car for FI and E85. These are just things I've discovered on my own through experimentation, and i could be doing it wrong and not even know it, but my car runs well so it can't be that bad lol.

- Get the injectors right first before you go modifying anything else. If you just go and install your whole turbo kit and injectors all at once you now have to tune injectors against a maf that isn't known-good. It is much better and more straightforward to just install the injectors first, get the tune dialed in for them, then proceed with installing the turbo kit. I got around this by reinstalling my stock intake and tuning injectors against the stock maf, but it could've been done better had i known ahead of time.

- Don't just go hacking away at the maf scaling. If you catch yourself wanting to make large corrections, look elsewhere for a problem first.

- collect lots of data. the more you have the more effective the various tuning spreadsheets and tools are.

- Just because your car idles perfectly with low trims doesn't mean your ports are scaled anywhere near properly. The car idles on 100% direct injection, so simply sitting there at idle and watching trims won't help you.

- When fine-tuning the maf, you might see an area around 1.2-1.6v that is more erratic with more scatter than the other areas before and after it, and it won't respond predictably to changes. In my case, this was caused by discrepancy between the accuracy of the port and direct injection configuration. if one is more 'off' than the other, it causes a swing in the afr/corrections. they both need to be right, basically. this matters more if you're on e85 and have rescaled the direct injection as well as the port.

- if you're trying to tweak injector latency, you can set the whole map to 100% port injection and then collect your logs, filter and process the data to eliminate most of the noise introduced by direct injection.

- when you switch to e85 and have to scale the direct injection, introducing another variable into the whole mess, you can set the map to 100% direct injection and do the same for closed loop to get an idea where it needs to go.

- it seems that having a 'correct' i.e. true injector scaler number is one of the more important factors to drivability, presumably because the ecu is using it for a bunch of stuff we don't know about. it can be completely wrong and the car will still seem to run fine, you're just fighting against the problem. this is where process separation between maf and injector scaling is important.

- when you change the injector scaler, remember to change all the tables defined in ms (ipw) along with it. tip-in enrichment and cranking ipw are the two i can remember.

- no matter what i do i can't fix an issue where the car goes rich on deceleration. my best guess at this point is that there is another map in ipw like the ones mentioned above or something that isn't defined. luckily it doesn't seem to effect much except an occasional puff of smoke coming to a stop.

- tip-in tuning is a bitch. it also effects drivability a lot.

Note that none of this has anything to do with tuning so to speak, just getting the car running and configured well. I wouldn't even attempt to give advice on things like timing maps, as i'm only beginning to work on that side of things. Anyway, hopefully that'll help someone out or start a conversation.

James, thanks for sharing to the community. Everything you have posted about has to do with tuning. There are many additional aspects to tuning than just AFR, Timing, Boost. Looks like you are discovering some of these aspects. Keep digging and you'll have your car running great!

- Bob

Just more amazing information that I am soaking up. How much can flow can the DI system keep up with? Is it capable of flowing e85 while boosted without the PI?

i guess what i meant by configuration vs. tuning is that everything i'm describing is deterministic. there is a well-defined idea of 'perfect', a target to aim for and math to get you there. there is no math that will tell me what afr my engine will like, or where mbt is (actually maybe there is, but it's above my head lol).

so maybe it's tuning, just the less voodoo-ish parts :).

i haven't pushed it more than just driving around tuning closed loop, but i was able to push it into open loop with acceleration up to 4.5-5krpm without leaning out. that's not WOT, i didn't chance it :). if you wanted to be safe you could turn the ports on above 1.0 load and still accomplish the same thing in closed loop.

here are some screen grabs of the results after fine tuning everything on e85. sorry for the huge size, skitch + retina display = fail.

the final maf curve:
https://www.evernote.com/shard/s251/...ls-Desktop.png

and fueling error chart:
[IMG]https://www.evernote.com/shard/s251/sh/9b61c2b1-376f-4e19-bee1-75c69a7***42/c3067a29698b39ca1e40d4e44235cab7/deep/0/Windows-7---Parallels-Desktop.png[/IMG]
sorry i didn't collect much data for this run, just enough to get a pic. with more data it fills in the gaps and the worst error is -1.1%, average -0.2%.

injectors are dw900's and the pump is a dw265. it could be tweaked further, but it's good enough for me. still experimenting with timing and turning up the boost. at 11.5psi it'll break traction when the boost hits in third, what more could you want? for now.... :)

Yup, this is very often overlooked and why i plan to do fuel pump and injectors while i am still NA

Very very true, more often than not you can tweak the individual scaling (port vs di) to get your trims closer to zero (or tweak both together) but you need to pay attention to which system is active. For example.. cold cranking is 100% port, cold high idle is combined port and direct, cold kick down idle is 100% port hot idle is 100%DI.

See above, only DI idle after warm up, you can actually here when it flips over from port only idle


Yup, as i mentioned above, anyone working on scaling need to log di time, port time, and or the injection mode... you have to get the two systems balanced before you worry about maf scale too much.

Good tip, i may try that when i do the pump and injectors

you probably need to adjust the top row (lowes g/s row) of the DI scale table, decel will probably be under DI control and at a lower maf g/s than idle so you should be able to reduce the gdi flow rate some there. Get some logs of the rail pressure and maf g/s during decel and you will see where you can try to adjust, will probably be low rail pressure as well as low g/s/. There are also some overrun fueling tables now in romraider you might look at.

Cam timing has a lot to do with the tip in area. I found that running too little retard on the exhaust cam (bigger number = more retard) in conjunction with too much advance on the intake side would induce more tip in knock and in some cases steady state part throttle knock like going up a hill in a higher gear right around the .6 to .8 load range... you can try bigger numbers in the exhaust cam map and smaller numbers in the intake cam map in the problem area, just go 5 at a time on each map till you find the sweet spot... also be sure to ramp smoothly.. ignition, fuel, cams, etc

Ohh and sorry i didn't get back to you about your base map man... major life changes happened and i got busy as heck moving to socal. But I'm getting settled in now. If you need anything give me a shout :burnrubber:

Great tips btw:clap:

Keep this in mind: When your exhaust cam is near its most advanced point, and your intake cam is near its most advanced point, you have early overlap. By early overlap, I mean that as the piston moves upward and the exhaust valve is still open, the intake valve begins opening. This can cause backflow of hot residual gas.

The backflow occurs when the pressure in the combustion chamber is higher than the pressure at the intake port when the valve is opening. This often occurs in the cruising area of operation. In some situations the backflow is a desireable thing, but in other situations it just results in increased knock sensitivity and combustion problems.

http://www.ft86club.com/forums/attac...1&d=1386640616

So this is out of the BMW M5 service manual, but it's a nice illustration of a cam profile diagrams used during camshaft development and cam phaser calibration. If somebody had the proper equipment they could develop such a chart for the FA20 stock cams. Looking at the x axis units: we're used to thinking in terms of BTDC firing, like a spark timing of 20 degrees BTDC. The "0" number in the diagram is TDC intake, or 360 BTDC firing. This numbers at the top there, 115-125 , 60-55, are the centerlines. When we phase the cams the centerline positions change, but the profile of the cams do not.

The area I circled is the spitback area. The other extreme is between 0 and 60 degrees, when the exhaust is fully retarded and the intake centerline is locked at the default position. In that case, the piston is descending when the overlap occurs, and you don't get spitback into the intake port. Again, there are times you would want this and times you wouldn't, and it interacts with what hardware is one the engine.

great info! this should be a sticky

awesome! thanks for the info, always much appreciated.

i actually want to argue one part of my first post: the 'Get the injectors right first before you go modifying anything else' part.

doesn't having complete control over which fueling mode is active at any given time allow for independent scaling of the maf and port injectors even if they are replaced at once?

put better, couldn't you just turn off the port injection and scale the maf off of the di (which you know is really close to correct), then turn the ports on and scale them against the now-known-good maf? come to think of it, this is exactly what i did when creating the e85 map, only starting with the di already scaled for the ethanol content.

I suppose you could do that but we have control over di flow rate using a 3d table port injection we only have a 1D scaler value. Doing the maf off the DI only would allow you to create a lot of problem spots for port injection.

It may be better to do it the other way around, run di only, adjust the injector scale and latency to get as near zero trims as possible, then scale the maf, then bring DI back online and tweak the di scaling using the 3d table so it matches up with the maf and ports.

@arghx7, awesome info as always :thumbsup:

Right, and this aligns with what i was seeing... advancing both cams in the problem area makes the problem worse, retarding both makes it better (less prone to tip in knock). With our cam maps, larger numbers on the exhaust side = more retard larger number on the intake side = more advance.

So, of for example the cam maps read 25E 30I in an area you are pron to tip in knock, try 20E 25I and see if it improves.