Graph fueling filters & gear load

[JimMueller]

I've been using this filter for my VE error filtering:

([6310]=0 or [6310]<16) AND ([17.156] > 30) AND ([17.156.avg(500)] > 30) AND ([17.156.avg(-500)] > 30) AND [5.242]>176 AND [2111.156].slope(3000)}<3

And MAF:

([6310]=0 or [6310]<16) AND ([17.156] > 30) AND ([17.156.avg(500)] > 30) AND ([17.156.avg(-500)] > 30) AND [5.242]>176 AND [50080.50]>2500 AND [2111.156].slope(3000)}<3

However, both are tuned in open loop and IIRC my early LS1 PCM does something different with fuel trim cells in open loop? Are these the ideal filters to use in open loop for VE and MAF tuning for me? If not, recommended changes?

Also, I spent a little time on a load bearing dyno about 2 weeks ago. My pinhole exhaust leaks had me going in circles so I'll go back after they are corrected. But something I experienced, and the owner verified, is that the AFR changes in different loads/gears. I don't really pay attention to that on the street. I'm trying to tune this for HPDE events (20 min road racing sessions), so I'd want optimal acceleration across wide RPM ranges in multiple gears. How much time is spent in each gear would vary per track. What's the recommendation on how to optimize the error when I desire optimal acceleration in multiple gears?

[kingtal0n]

different gears will have different optimal timing condition.

The slower the engine accelerates (i.e. numerically higher gears) the less timing is desired

You MUST tune an engine to its MINIMUM timing requirement, so always tune it in the numerically HIGHEST gear that it will see extended load.
For example if you are tuning a car that may see some long over-drive pulls in 5th or 6th gear, you MUST time the engine and watch EGT for those long pulls.
Another example, 4l80e car never run WOT in overdrive, so they ONLY need to be tuned in 1:1 gear.

Ex3. If you tune aggressively for 1:1 pulls and then load the car on the street in an overdrive gear for extended period (say .68:1) The engine will experience high pressure spikes, eat rod bearings, beat the rings, possibly blow the headgasket or a chunk from a piston as the fuel heats and air heats up during a long pull with low engine acceleration.
------

The main issue is rate of change of the engine, followed by fuel behavior. As gasoline heats up during a long pull it will want less and less timing. The same with IAT if the car has a so-so intercooler or difficulty maintaining a 100*F~ IAT the rise in temp will upset the fuel eventually.

Gear to gear tuning is partially IAT and Fuel (component) temperature dependent (the temp of fuel vapor is dependent on the surrounding engine parts), For example if the EGT is very high and the piston and valves are very hot, it will be a risky situation for gasoline even at low output, even at part throttle it can be dangerous because of all residual heat input.

Likewise, if the car has GREAT cooling toys, huge intercooler, lots of recirculating fuel, aluminum heat sink intake parts, well controlled EGT, water injection to control combustion temp heat input, etc... Then you could use more timing for those intermediary gear changes because you have the confidence that the gasoline won't become an issue halfway into the racing event.

[kingtal0n]

Annnd you might be wondering why I posted about timing when your question was about fuel. I skipped over fuel because its a lot longer to explain and has very little relevance.

But just in case.

On N/A cars there is nowhere for the map to go other than 100KPA at WOT. That means whether you run 1st 2nd 3rd 4th or 5th gear the mass of fuel injected will be more or less the same.

So when you take each gear and notice a different a/f ratio it may or may not have something to do with the actual a/f ratio, especially if you check the log and see identical injector on-times and so forth.

How to identify whether the difference is due to rate of change (engine load) of the engine?
In a true rate of change challenge where the A/F ratio adjusts itself going from one gear to the next, the tell-tale factors are:
1. The engine is always at a higher load and a leaner a/f ratio (NEVER richer with more load) i.e. 4th is leaner than 3rd
2. the engine has a large-ish camshaft capable of deriving VE improvements over a narrow RPM window, MAINLY due to stock-size Ports (head/intake ports).
3. the alternator voltage REMAINS CONSTANT
4. there is no turbocharger

---Overview---
The quickest way to explain this is, for engines with stock-size head ports (or smallish) the port velocity is very high, or it can be very high. High port velocity is good because it means the engine can respond quickly. However it can also 'choke' the flow to cylinders sometimes, from being too small of a port.
So lets say you run the engine through 3rd and 4th, and notice the engine is leaning out in 4th near the end of the run.
FIRST You look at your alternator voltage and make sure the real reason isn't simply an injector-delay related (small influence but worth checking).
THEN you would ask yourself, "does my engine have enough time to fully fill every cylinder in 3rd gear, due to the fast rate of change? "
To answer this question you can put the vehicle on a dynojet and compare 3rd with 4th gear.
If you get LESS torque in 3rd gear than 4th, and a leaner a/f ratio in 4th, and injector duty is the same for both gears, you can be somewhat assured that the engine VE (and compression) is increased in 4th gear. In other words, the a/f is actually leaner, the engine makes more torque in 4th than 3rd.
With that situation your only option is to tune for 4th gear. You can't "get the best of both worlds" without risking damage in 4th.
The same thing goes for 5th. If you switch to 5th gear and plan to EVER run 5th this way, it too may lean out if you tuned for 4th gear.

WHY THIS HAPPENS:
In 4th gear the engine rate is slower, so this is more time to fill each cylinder, if the cam/head agree well enough. In other words, its a sign that the combination of parts is suitable to fully fill a cylinder (more torque) but only when the engine's rate has been slowed down enough (numerically higher gears).

WHAT ELSE CAN HAPPEN:
Because the majority of modern era-combustion tech (head ports, valves, etc...) is *VERY GOOD* We rarely see this kind of discrepancy between 3rd and 4th gear and 5th gear, etc...
In other words, there is a MORE Likely explanation which has nothing to do with a/f ratio, in a majority of cases.
1. As the wideband heats up (too near exhaust temp), it read more RICH (typical in-car wideband). This would give a richer A/F in 4th than 3rd, which should NOT happen in reality
2. Alternator voltage dropping during a longer high rpm run can lead to a leaner a/f ratio at the same duty cycle as numerically lower gears
3. Ignition timing... has an influence....

Ignition timing is what I based the "tuning theory" on above in the first post. Because it is by far more important to locate your "happy timing" area than it is to nail the a/f ratio using a typical wideband.

The rule for normal in-car widebands and the rule for tuning in general is to always tune a bit too-rich, in the longest gear that will be used as a long gear.
So if your target is 12.0:1 and your longest gear is 5th, you would tune on the street so that 5th gear was approx 11.5:1.
Then, when you hit the dyno, the a/f in the car will stay 11.5:1 but the dyno will tell you something like 12.2:1 and you will actually be slightly 'lean' and have to add some fuel.

Timing on the other hand, has both "dyno setting" and "street + longest gear setting" and then everything in between.
Street timing will be less than dyno timing because a typical vehicle can accelerate a dyno roller faster than the vehicle can accelerate on the actual street.
Remember the rate of change of the engine influences air and spark which can have an influence on a/f ratio.

Again, I am assuming fueling is 'negligible' since we always shoot for some rich a/f ratio number, UNLIKE timing which we never know where it will land unless we've tuned a bunch of "those setups".
When you first use the dyno the timing should be the least, remove timing until engine backfires or torque is lost significantly, and EGT begins to rise rapidly. Then add some timing back to bring down the EGT and prevent misfires. That is the starting point.
Then you add timing from there, 1 or 2* increments, watching torque, graph smoothness (Use smoothing=0), a/f ratio, EGT.
-Once the torque graph starts to become rough, its a bit too much timing. Try to keep the graph very smooth with smoothing=0
-Only keep the added timing when it results with significant torque increase and EGT drop. (more than 1.5 to 2% additional torque)
i.e. 10ft*lbs from 1* of timing across the board on a 400ft*lbs of torque engine, coupled to a 70*F drop in EGT is GREAT.

-As you keep adding timing, it will always give more torque, due to the spiky/wavyness influence of the torque curve. In other words, MORE timing will always INCREASE the appearance of engine torque on a dyno, until it outright destroys the engine. This is why we STOP ADDING TIMING once the MAJOR improvements are made (1.5 to 2% increase per degree of timing).
In other words:
-If you are adding timing and the EGT does not change, REMOVE the timing
-If you are adding timing and do not see major gains (1.5 to 2% usually) in torque per degree, REMOVE the timing

Last, timing can influence a/f ratio. Or rather, timing can adjust what the wideband tells us. Remember the wideband isn't measuring fuel, just O2, so the effects of timing, and high exhaust gas temperature/pressure, play a role in what the wideband can report. For example a very retarded timing results with small misfires will show lean in theory. But it could also show up as rich depending how the fuel is burning as it enters the exhaust system. Remember that fuel can react in the exhaust OR cylinder, and that wideband position and any exhaust leaks have an affect.
For another example, if you find that advancing the timing is leaning out the a/f ratio during a pull, it may be that more fuel is able to burn in the cylinder and that the exhaust temp isn't very high yet. In other words, timing adjustments don't make a/f adjustments, but they can still adjust the a/f ratio reported by the wideband.

also
-A high overlap camshaft in it's peak VE range can also show some LEANING because of O2 ingested during overlap (suction of the exhaust, pulling O2 from the intake).

rood ruck

[JimMueller]

A lot of good info there, thanks! OK, my takeaways:
1) My control module voltage was stable 13.6-14.2 on the test pulls
2) I did not compare A/F in the different gears due to not knowing about it
3) Both in car widebands are in the same vicinity that O2S1 sensors are located (in the collector)
4) I generally try to tune 2-4% rich
5) Cam is relatively small for the N/A 416ci displacement (232/242 114+4, only 2* overlap), unported OEM LS3 heads
6) For VE, I'm guessing tune the RPM columns for the longest gear I would expect to use in that RPM column?
7) For MAF, would I separate the frequency into sections, and for each section, tune it for the longest gear as well?
8) What equipment is needed to measure/log EGTs? Does it matter which cylinder(s) are monitored?
9) What about the filters, can they be improved for this vehicle?

[kingtal0n]

1:1 final gear is fine, no need to test in overdrive unless alot of racing is going to be taking place in overdrive.

Tune the engine so it can run rich and stay well healthy for long gears, this will trickle down effect to protect shorter gears

IF its Naturally aspirated I wouldn't worry about the EGT. Thats just for theory and understanding.. until you get to the turbocharger experience where EGT becomes integral

Filters can be used many ways. To "improve" can mean many things. First, you must define a goal or target. Then sample the data at a high rate to increase signal resolution. Increase filter sensitivity and so forth as needed. Define your purpose and decide whether the ECU data can supply that kind of purpose

[JimMueller]

By 'improve' I mean using the best combination of filters to ensure the fueling error reported in the VE & MAF graphs is as accurate as possible. Filtering out scenarios that should not be factored into the averages.