Knock Retard hell! Desperate for help

[sssnake]

GHuggins,

Thanks for the help. I think I am beginning to understand the goals with EOIT. Injecting on the back of a closed intake valve gives better atomization at low rpms (at some risk of fueling instability). Injecting while piston is at maximum velocity helps atomization/mix at higher rpms. So I look at the Z tune you posted. Your end of injection at high RPMs is about 580/220 camshaft degrees. Which translates into 110 crankshaft degrees. So +20 BDC. For the idle portion of your EOIT you seem to end injection around 385 and 412 camshaft degrees. This is pretty close to the stock 410 (at idle). I assume there is very little overlap in that cam. This is me trying to parrot back what you said in terms I understand, If I missed something in this let me know.

But here is the rub. If you look at stock data for the LS7. The Boundary is 520 across the board. So the latest fuel will be injected is 520/160 camshaft degrees or 80 crank degrees. This is obviously before BDC. The thing that just keeps making less and less sense is that from the other thread where DSteck offered up his calculator he mentioned both ECT and RPM values are applied as negatives (EOIT=Boundary-ECT Adj.-RPM Adj.). All of the stock data for RPM Adj (Normal EOI Target vs RPM) that I have seen increases at the maximum RPM end of the table . This would be putting the EOIT further away from maximum piston velocity not closer. What am I missing?

BTW - for some reason I keep believing these adjusters should be applied as positives. BEAR WITH ME - If the valve is cold then letting the fuel sit on it is not going to atomize the fuel much better. So I would think you would want to move the injection timing toward maximum piston velocity so the piston action can swirl the fuel. Then as the valve warms you transition back to letting the fuel sit on the valve. If that is true and the ECT values are additive rather than negative then my logic holds. Otherwise what does the larger delay at cold idle do. Let more fuel leak through the exhaust valve. That doesn't seem to help with anything other than warming cats. There are already other measure for warming the cats. Then as you approach maximum RPMs you want to get the EOIT closer to maximum piston velocity because at this point your injection window is long enough to let the fuel hit the warm intake valve AND be swirled by the piston action. This only holds if the Normal EOI Target vs RPM values are additive. If they are subtractive then you just moved the end of injection away from the max piston velocity area (around BDC).

[GHuggins]

The injection timing on the Z28 is injecting right around mid overlap of the cam at idle... To my understanding via the spreadsheet the rpm and ect tables are both adders not subtractors - In other words it moves the soi sooner. Colder the engine it injects sooner = more fuel possibility of going out the tailpipe = quicker cat light off and longer to sit on an all ready warming intake valve... Your also missing a very important part of the equation - ms of injector pulsewidth. In the higher rpms I try to move as high as I can to focus "more" of the fueling event "around" ppv and bdc of piston stroke. AND yes your right - soi is way before intake valve opens... Bigger the injectors = the more you can fine tune this...

I don't yet know everything about the "best way" of doing this. Just know what I've been seeing and doing? AND the way you "attack" it can differ vehicle to vehicle depending on customer wishes and the cam being used... Close to neg overlap and you can put it back on the closed intake valve and not have problems... I've tried setting the ect table up different ways, but I've always wound up more times than any with a stall condition after firing because of it. I always thought it was because the fuel was just going straight out the exhaust on the high overlap cam setups? This is why I went to flat lining them... I've got a really BIG cam blown car I've got to tune starting tomorrow with cold starting popping in the exhaust problems (not atomizing the fuel) - I'll see if it likes being back on the closed valve or more around BDC. SO far I've only had any luck out of it putting idle timing after the exhaust valve closes on this particular car...

[GHuggins]

Just trying to help clarify things as now I was second guessing myself....

[ECM] 13338 - Normal EOI Target vs. RPM: The target angle before the boundary that the normal injection pulse should finish. (I take it that this "subtracts from" or advances the injection angle)

[ECM] 13337 - Normal EOI Target Adder vs. ECT: Adds to the Normal EOI target (? Not really sure ?)

[ECM] 13336 - Injection Boundary: The latest possible crank angle that the injection pulse can finish for fuel to be delivered to the cylinder. (Self explanatory - latest possible injection point allowed - ecm will advance injection timing depending on injector ms to keep fueling spray time withen this "wall" angle)

This is the work of someone else on here pictured below that I used to go by... This is done by someone different than the person that came up with the spreadsheet I use, so I figured the spreadsheet was correct... The Adder vs ECT table has me scratching my head now? I know the pre-mentioned cammed motor did not like having injection timing anywhere near a closed intake valve - was wanting 40% more fuel. Still need to play with it more to see if having "it" at bdc benefits anything...

[ecotec88fiero]

Just trying to help clarify things as now I was second guessing myself....

[ECM] 13338 - Normal EOI Target vs. RPM: The target angle before the boundary that the normal injection pulse should finish. (I take it that this "subtracts from" or advances the injection angle)

[ECM] 13337 - Normal EOI Target Adder vs. ECT: Adds to the Normal EOI target (? Not really sure ?)

[ECM] 13336 - Injection Boundary: The latest possible crank angle that the injection pulse can finish for fuel to be delivered to the cylinder. (Self explanatory - latest possible injection point allowed - ecm will advance injection timing depending on injector ms to keep fueling spray time withen this "wall" angle)

This is the work of someone else on here pictured below that I used to go by... This is done by someone different than the person that came up with the spreadsheet I use, so I figured the spreadsheet was correct... The Adder vs ECT table has me scratching my head now? I know the pre-mentioned cammed motor did not like having injection timing anywhere near a closed intake valve - was wanting 40% more fuel. Still need to play with it more to see if having "it" at bdc benefits anything...

This is how I thought it worked for the most part.

Boundary is the absolute latest angle the ECM can FINISH spraying fuel.
The RPM and the ECT adders basically subtract from boundary angle which then sets the EOIT for that RPM and coolant temp and the ECM calculates that start of injection from injector pulse width and this EOIT.

If you look at a stock calibration file for an E38 Camaro ECM you can see how this would make sense.

For example the ECT adder table shows at low ECT, the adder is 250 and as the engine warms the injector moves closer to 100. A cold engine needs more time for the fuel to evaporate so you would inject earlier in the cycle and as you warm you can inject later and end up with the same evaporation.

The RPM adder table starts at close to 0 and increases to close to 100 degrees as RPMS climb. I have a harder time explaining why you inject earlier as RPMS climb but it may have to do with intake velocity and how it affects fuel atomization. If you inject earlier as intake velocity is increasing, does this gives you better fuel atomization? Also you have less time to mix the fuel as RPMS climb so maybe you need to inject earlier to give the fuel a better chance to mix even if the intake velocity is higher.

Obviously with a bigger cam, most of this goes out the window and you need a dyno and some time to figure out what works best. I haven't played with the RPM adders on the dyno yet but usually just shift the ECT to help prevent short circuiting.

GHuggins obviously has played with this a lot on the dyno and that is the only real way to get it dialed in and see which helps power production which is all the validation I would need along with some street driveablity validation.

[GHuggins]

OK, that's what I thought and I think I have been mis-saying all of this leading to the confusion... When I've been saying "adding" I actually should have been saying "subtracting from"... Basically it's just like the above pic and ecotec describes... BUT from my own findings I think the OE injects sooner as the rpms climb to better vaporize on the valve to meet emissions. Port velocity isn't really there until ppv - It is, but isn't kinda thing The "popping" cammed motor also liked the soi to be just before the exhaust valve closed as all of the other cammed motors have... Problem was I wasn't factoring in dynamic compression where the current 11.2:1 motor was actually running at a pumping compression of 7 or 8:1 (give or take for altitude compensation) - IVC point isn't until 87 degrees abdc... After optimizing the injection I added in some serious timing = currently no more popping = me currently happy... Now off to finish dialing it in and revaluating after cool down

[GHuggins]

And just to help clarify things. The aforementioned motor did not like idle injection timing at ppv or bdc. Air just moves too slow at lower rpms. PPV was the better of the two, or at least seemed to be and didn't require "as much" extra fuel to be added...

[schpenxel]

Whatever happened to the simplicity of firing all the injectors at once? lol

[sssnake]

From a post in the "other" forum Greg B had this to say:

OK, it looks like the info here went sideways pretty hard...

Before adjusting injection timing, it's important to understand why it's there. Most OEM applications are timed to inject fuel against a closed intake valve. This does two things; it cools the valve and aids in evaporation. When the engine is cold, fuel takes longer to evaporate, so it must be injected sooner to get enough of the liquid evaporated to enable clean combustion.

Clean combustion is the key here. If you're burning everything you injected and you had decen charge mixing, it shouldn't matter as much when you injected it. On the other hand, if you're either not burning liquid fuel in the chamber or not mixing completely, it can skew what you think got injected into the cylinder. This is why the OE calibration injects earlier on a cold engine to allow time for fuel to evaporate and mix prior to combustion. (Look at the table and you'll see the trend!)

In the aftermarket, we have the nasty habit of plugging huge cams into engines that open the door for a couple problems. One is that the lower port velocity reduces charge mixing at low engine speeds. The other is that having both valve open simultaneously (during overlap) opens the door for "short circuiting" where vaporized fuel may escape out the exhaust valve without ever being burned if there isn't sufficient port velocity to push it down into the cylinder. This can usually be addressed by delaying the injection event slightly so that the vapor isn't there when the exhaust valve is open. The potential downside is that large quantities of liquid fuel being injected into a cylinder at low port velocity may lead to bore wash at extended idle.

If you see a change in AFR resulting from injection timing, it just means that you didn't get a complete burn in one case. The object is to completely evaporate and burn whatever fuel you inject, regardless of what's going on with the aircharge or port velocity. This was one of the exercises we did in my last advanced GM tuning class with a 23x+ camshaft in a GTO. Once you find clean combustion, it's a lot easier to dial in the VE (or VVE) tables, THEN you can address idle airflow targets.

OK this completely agrees with ecotec88fiero's comments. I can accept that this works for addressing variations with valve temperature. But that leads to why the RPM table moves the boundary closer (lower degrees) at higher RPMs. The only thing I can think of is the higher volume of fuel needs more time to atomize and moving the boundary closer provides more time for this to occur. Also, as these events are measured in degrees not time then at faster rotations you are afforded less time for an equal number of degrees. But this gets us back to why don't we leave the fuel on the back of the valve as long as possible at idle. The only thing that I can think of is port velocity. If I am not mistaken the port velocity will be lower at lower RPMs for a cam with more overlap. If this is the case you may not want the fuel to sit on the intake valve because the added obstruction of an intake valve in the lower lift position may be detrimental to getting all of the charge into the cylinder. This is only conjecture but I am groping for a logical answer to how this is tuned and why.

The real answer is that I need to get the wideband hooked up and get out and log.

[07GTS]

i think EOIT is very engine/cam/injector specific where i have 57 deg total overlap and i delayed from ECT adder at 110 down to 80 and thats it works the best ( if i go other way and set it to 120 i loose 15% out the exhaust ) i also keep the RPM adder to slope to 100 deg at 4k with that at any rpm i have good response and smooth fueling but if i delay it to PPV or BDC then i get a strange lean spike in a wot pull and the engine dosnt feel as responsive to throttle, with the BOUNDARY 520- ECT 80- RPM 100 = 340 deg EOIT that finishes right at my IVO 331.5 deg and that seems to work the best for mine

[GHuggins]

i think EOIT is very engine/cam/injector specific where i have 57 deg total overlap and i delayed from ECT adder at 110 down to 80 and thats it works the best ( if i go other way and set it to 120 i loose 15% out the exhaust ) i also keep the RPM adder to slope to 100 deg at 4k with that at any rpm i have good response and smooth fueling but if i delay it to PPV or BDC then i get a strange lean spike in a wot pull and the engine dosnt feel as responsive to throttle, with the BOUNDARY 520- ECT 80- RPM 100 = 340 deg EOIT that finishes right at my IVO 331.5 deg and that seems to work the best for mine

They are indeed all different. I try not to hit anything around bdc until 2500 rpms. It seems that velocity in the cylinder doesn't pickup enough until then... I will taper it up starting around 1500. AND more than cam data goes into the equation. Like you pointed out with the (engine) portion of the equation... Such as whether or not it's boosted. If boosted retard it quicker... Boost overcomes the atmospheric pressure in the exhaust rather quickly and even stock cam motors will allow the fuel vapor to escape the combustion chamber.

You've also all ready retarded yours 30 degrees. That should be more than enough for that cam...

[GrannySShifting]

Exactly what I was saying. To be clear though you can't zero all of the predictive coefficients (at least I dont think you can). You CAN zero the MAP, TPS, and corrected values. This is what I am currently running with much better results than stock. Not completely there yet but getting much closer.

Attachment 55538

Where are all those tables located?

[schpenxel]

Airflow-->dynamic

[GHuggins]

Something I've noticed with the prediction coefficients (since sssnake was pointing it out) is that if it's zeroed or altered in any fashion you have to wait the "15 minute" relearn period before it will correctly calculate the cylinder air masses. It will show tons of spark knock, even though there is none and it will show a much lower cylinder airmass. Timing after the "perceived" knock timing pull will actually be the wot correct timing anyway - confusing After the usual 15 minute relearn the cylinder airmass calculations go back correct with no knock at all...

Something to think about if your seeing funky things with the prediction coefficient tables....

[itsdaveonline]

Interesting. So if i were to flash my tune and which also has zero'd prediction coefficients, I should run the car for 15 minutes before logging data for VE tuning? What about the other way around if i fail SD and tune the MAF?

Maybe im over-thinking this. I forget the exact functions of cyl air mass, again not at the same level of tuning you guys are... I know it's a calculated number. Is it specific to just one or both airflow models (VE & MAF)?.

Btw im getting my tune dialed in alot better. Far better than what the professional "tuners" the car has been to before. It's taken a long time, new wideband, new battery, better grounds, more reading, more tuning, and more driving haha

[schpenxel]

After you flash a gen 4 you pretty much need to wait 10-15 mins of run time before using any data. You'll tend to get some weird/useless data the first few minutes. Seems to be less of a problem in open loop. There's a thread called "rich after flash" that goes into more detail.. but yes, try to give it 10-15 mins of run time before you start recording (or at least don't use logs created in that time frame)

I know some things like intake valve temp reset and have to work their way back up to where they'll normally be, but I'm not sure beyond that why they behave so strangely at first

[Michael_D]

As for the O2's in the headers... Go to Open/Closed Loop - Closed Loop Proportional - Airflow... This will need to be dialed in off of your air fuel readings while in closed loop. What your looking for here is how much of a swing you have on your air fuel reading while your O2's are switching. Use something like a constant 2000 cruising rpm while on a flat level road. You'll keep "decreasing" this until your air fuel no longer swings more than .3 points. At least that's what I like to aim for... This will also slow down your fueling corrections, so don't go too low.

For the integrator delay - airflow.. You'll want to increase this to "fine tune" your air fuel swings...

With long tubes I like to start with something like multiplying the proportional airflow table by .8 and the integrator delay table by 1.25 then dial things in better from there. Sometimes the .8 and 1.25 is nearly perfect at least for most kooks setups...

This is how I do things. Perhaps not the best or even correct, but it works... As far as I know anyway?

This thread has a lot of great info in it.... I was not expecting to see so much on injection timing and O2 calibration.

GHuggins - How far would you suggest lowering / raising the PI tables, maximum amounts? I just multiplied Proportional by .8 and Intregal by 1.4. Didn't help get bank to bank fuel trims closer in my trouble areas, but closed loop accelleration is smoother. The engine would kinda surge when mildly accellerating.

My problem area is 1800 rpm, while in fifth gear. That's when I feel the engine missing, and that's also when bank to bank trims have the furthest spread, B1 always lean at this engine speed/load. RPM goes up or down, the trims get closer.

6-30-16-O2 setting change-pm.hpl

[GHuggins]

Michael, unless I was looking at the wrong portion of the log, bank to bank at 1800 was almost identical? Of course I am tired - just got in the hotel room and all and it's 4am my time, 1am this places time...

BUT to answer your question, I have heard of guys going to multipliers of 1.9 on integrator and getting better results and almost zeroing out proportional and getting better results - I don't recommend going that low on proportional because it seems like fueling "gets stuck" and can take forever to rebound if it does at all...

If you like you can send me your tune with build specs and I can take a look at it - I actually took the first vacation I've been on in years this weekend, so no guarantees, but I'll try to take a look

[JBZ]

I actually went the opposite dirrection with proportional 1.1 and integral 1.2 or so. What i saw with the lower than factory p values was it was taking to long for the swing to reach peaks and valleys. When i finally increased the p value i saw a steeper mv jump (very vertical line) between rich/lean and response seemed to be better. Granted every set of headers are different in 02 locations so what works for me doesnt for thd next guy. Keep in mind the porportional gains are how much fuel is injected to reach rich/lean setpoint (for lack of better terms) so too much seemed to over shoot and too little seemed to not reach it. All of which for me anyways would result in random misfire count changes and some potential surging at idle or cruise, but that mostly was attributed to integrator delays being to small.

Micheal, I started a thread about header tuning with o2 placement changes and Ghuggins commented so pretty good stuff in there about these gains. And i agree this thread was very informative, not what i expected when i read it for thr first time.

[GHuggins]

Yes, I've had better luck increasing proportional slightly myself as well

JBZ, did you ever have any luck dialing in individual cells of the proportional or integral modes? Wondering if Michael doesn't have this problem?

Michael a couple of other things you can look into after dialing in the proportional and integral - try making you min air less and playing with your etc scaler

[JBZ]

Yes, I've had better luck increasing proportional slightly myself as well

JBZ, did you ever have any luck dialing in individual cells of the proportional or integral modes? Wondering if Michael doesn't have this problem?

Michael a couple of other things you can look into after dialing in the proportional and integral - try making you min air less and playing with your etc scaler

Yes i did, was swinging to slow last time you commented on my thread and i lowered the i delay in the lower cells but got idle surging...not bad in normal dynamic airflow mode but when i had maf failed for vve tuning it got much worse. I have since increased the i delays again to 1.2 of stock other than idle area (0-8) as my car never sees that cell. I need to go back to my airflow table and readjust it all to the left again...had just been focused on other areas lately. Seeing as how everything damn near in cl references that airflow mode i think it carries more weight then we think. But all and all the values i have keep it pretty stable. No surging anymore. Still want to mess with proportional o2 error. Thats something i put back to stock values but like airflow mode think that it carries more weight than previously expected.