Tuning Fuel Injection-my observations (long and rambling)

[elacruze]

Beginning with the 2011 Hot Rod Power Tour, we drove about 4000 miles in 6 weeks starting in Florida to Michigan and back home through the Appalachian Mountains and the Carolinas. We saw conditions from 105* at Sea Level to 65* @ 3800 feet of altitude.
For new readers here's the thread to the engine build;

http://www.dodgecharger.com/forum/index.php/topic,73529.0/all.html

Due to a failed ground inside the ignition module we didn't get any tuning done after ~400 miles until I replaced the module in Michigan at the halfway point. Also the distributor is curved for my old .509 cammmed 440, and is way too aggressive for this engine and I didn't want to start fooling with it away from home since I was able to make it drivable if not optimal. Base timing is only at 5*, and still pings when the timing comes in at 2200 rpm. Heavy springs on the way...

First thing I realized is that diving into tuning EFI with no assistance from an experienced hand leads to...a very steep learning curve. EFI is not like a carburetor, where you take it out of the box and in almost every case the car will run and drive and you just have to do basic tuning. EFI has no built-in physics; it only commands what it's told, and even with pre-programmed data it can be so far off that going around the block may take a while.

EFI is broken down into parameters, or tuning segments; you need to get your head around them and relate them back to carburetors (if you know carburetors). These segments all act at the same time, so tuning the right one can be very troublesome until you know which affects what.
Basically there are 3 main fuel tables-a Main Map and two Acceleration tables.

The Main Fuel Map, which has numbers indicating how many milliseconds the injectors are commanded to be open, at the cross-section of RPM and Manifold Pressure. This is your steady-throttle flat road map. It's tuned when neither RPM nor Manifold pressure is changing. This is easy to tune, and the least important for drivability.
Next is the Throttle Position Sensor Acceleration table, which does the same job as the accelerator pump in your carburetor.
Third is the Manifold Pressure Change Acceleration table, which also does the same job as your accelerator pump...confusing at first.

The TPS acceleration table has % inputs, you tell it how much more fuel to add based on how far the TPS moves. Pretty simple, it prevents the mix going lean when you open the throttle.

The MAP sensor table also has % inputs, you tell it how much fuel to add based on the rate of change of manifold pressure, which is not necessarily dependent on TPS change.

The difficult part of understanding all this, is that you can't separate these three while operating the engine unless you're on a perfectly flat road where you can have long segments of uninterrupted steady throttle. Anytime you move the throttle or have inclines which affect engine load, either MAP or RPM changes, or both which adds fuel from the MAP Acceleration table. So, the best way to approach tuning is to tune the MAP fuel table first, if the car will go down the road at all on the base fuel map. You can drive straight and flat either way too rich or very lean, but you can't accelerate without adding enough fuel. By the time you understand all this, a big piece of the Universe will have become transparent, and you will be rewarded with an engine far more drivable and powerful than you thought possible. My engine, though with timing issues, idles dead smooth at 1000rpm (and with 18* initial timing very smooth at 750rpm) We ran it so lean on the main fuel map that we got completely off the ability of the oxygen sensor to read it, that should be about 17:1 A/F ratio with no reduction in power, and surprisingly running so lean made the engine much more responsive and also ran cooler!
Unless you see this data on-screen, you can't know how lean your engine likes to be at low loads-most people have their carbs set way too rich at steady state.

One other trick not in the user's manual is that you must understand that the MAP and TPS tables are in % (percentage) of Main Fuel numbers-so when you lean out the main fuel map, you're also leaning out the Acceleration fuel, which can lead to you thinking wrongly that reducing Main caused you to 'go lean'. When reducing Main Fuel, you must consider adding % of Acceleration fuel.

More; TPS fuel added is very short-term, the added fuel goes away very quickly (how quickly is programmable) Consider that you move the throttle to accelerate quickly, but the engine takes time to reach your desired RPM-during that time, the MAP change table is adding fuel. Thus the order of fuel goes like this;
Main>TPS>MAP>Main. For example, when you pull out to pass then return to your steady speed.

This may be clear as mud, but it's only the beginning of this thread-there's a lot more to discuss, and although I had a good working knowledge of carburetors before this EFI experiment, I know how much about tuning carbs I couldn't know before and this information is applicable and useful to carburetors as well.

[firefighter3931]

FI tuning sounds like a lot of work ! 

Hopefully it runs just as good or better than a well tuned carb 



Ron

[elacruze]

FI tuning sounds like a lot of work ! 

Hopefully it runs just as good or better than a well tuned carb 



Ron

It already runs at least as well. The stuff you can't get online is the instantaneous throttle response, and the depth of feedback from having a real-world interface instead of third-party investigation (spark plug condition and dyno with gas analysis) although those indicators are not to be dismissed.


I don't think there is enough difference in performance or mileage in an ultimate state of tune to warrant changing to EFI, but the amount of work it would take to get a carburetor to that ultimate state will be a great deal more. Also, EFI doesn't go rich with altitude, if you're driving through the mountains (although it probably makes little difference below 5000 feet) Most people don't go that far with their weekenders.

[elacruze]

Short update.

Distributor crapped out-old MP unit sproinked the pickup coil bracket loose...maybe that was part of the 'not returning to zero' start timing problem, certainly had to have some effect before failure. I made it home on the last leg, checking it with fingers the bracket just came apart.
Rather than fool with it in a hurry, I ordered a new setup from FBO with recurved distributor, no-ballast module and new coil (my old coil has...unknown origins)

I've sorted and refined the EFI too, there are so many subtle changes that make big differences in drivability-the latest was finding the combination to smooth transition from acceleration to cruising at a constant throttle position. Previously, when the vacuum caught up with engine speed and stabilized the mix would lean out suddenly which was not harmful, but annoying. With the new ignition setup I'll finally be able to optimize idle and mid-range timing and begin dialing in the top end a little too.

[flyinlow]

Interesting.      I have some questions if you have time.

What A/F ratios does it run at cruising at 70 mph? light acceleration? and WOT?

Does the fuel flow go to zero during closed throttle deceleration at road speeds?

Does the system have a limp mode to get you home if something fails?

Craig

[elacruze]

Interesting.      I have some questions if you have time.

What A/F ratios does it run at cruising at 70 mph? light acceleration? and WOT?

Does the fuel flow go to zero during closed throttle deceleration at road speeds?

Does the system have a limp mode to get you home if something fails?

Craig

I only have a narrow-band O2 sensor so I don't know the A/F ratios. We did lean the motor out until it felt weak, then richened a little-we were lean enough to go way past 0 volts on the O2. Based on the known (14.7:1 by the sensor) and the percentage of milliseconds we removed, I'm guessing we are around 17:1. That's a guess.

Fuel removal while coasting is optional with a programmable RPM cut-in. It's a nice feature, saves gas and can't putter through the exhaust either.

There is no limp mode per se, you can't take much away and still have the system know what to do. On a similar OEM  system, the 'limp-home' mode is really just open loop on the O2 sensor. If you lose a temp sensor or MAP sensor, I don't know what would happen-I suppose it may default to some median reasonable value. There is no accomodation for a 'check engine' light, nor trouble codes etc.

Don't forget that this Commander 950 system is obsolete, and more recent systems are probably more user-friendly to tune, even if they are 'dumb' systems like this one. The software uses meaningless terms for value tables that I don't fully understand, and don't seem to affect anything on my engine. The user manual is weak too, but this is all part of the reason I chose this system-it was used, cheap, and forces me to learn these things the hard way.

[b5blue]

  Believe it or not I'm following you somewhat, my Cherokee has forced me to wrap my head around TPS, MAP, and other sensor's functions. If I had my way, dealing with a RENIX unreadable injection system, I'd can the whole mess and go to a carb. setup. That's NOT a option for me as my system has a knock sensor tied to ignition and that is tied to all others through the ECU.
  My cuz. had a VW Dasher years ago with Bosh injection that crapped out. I found Omni's used a carb version of that engine early on and swapped it out. It ran super strong for many years after that. It's interesting your setup has no interconnection with ignition. 
  I'm dealing with value readings like getting .5 to 1 volt AC from the crank position sensor (I had .1 due to a crusty buildup on it.) and replacing my TPS then adjusting it to 17% of it's input voltage of 4.65 volts DC. With no memory or "fault code" provision, dealing with intermittent problems is a PIA as you must "catch" the failure and do voltage check readings then to find what is wrong. Do you have codes or readable memory to work with?
 

[elacruze]

My system *can* control ignition, but I won't let it-I don't want to add that layer of complexity to it, on this system you have to calculate some sort of delay factor depending on your ignition type/setup. My Gen II setup will be complete one-box engine management and I'll be able to map ignition completely, with individual cylinder trim if I want or need.

The only information WRT troubles is reading real-time values from the sensors, and there is a recording function in there too-none of which is useful without the wide-band O2 sensor, which requires a separate controller and another $250-$300 I haven't been willing to spend yet-and I want to dial the motor in by the numbers and seat of my pants first, so when the wide-band comes I can see what it is rather than making it what I think it should be.

[b5blue]

I see what your doing by limiting complexity. Once your somewhat adjusted well it will be easier to tweak. Don at FBO did a good job on my dizzy and ignition components. You may be closer than you think after installing it.