Knock Retard hell! Desperate for help

[GHuggins]

First off colder plugs would help, BUT really shouldn't be needed with this build... Wouldn't hurt anything to go (1) heat range colder while still using the factory plug gap. Tighter gaps and much colder would probably just lead to misfires while idling or while cruising... My experience anyway...

Injection timing... This is how I've learned to do it. Going to be long and possibly confusing? I'll try to post as many resources as I can. First off nearly all of the tools I use for this are not mine, but work very very well for this... First you'll need to print out the "blank" degree wheel I post and keep in mind that this will have all 4 strokes of the motor on it. In other words all 720 degrees of crank rotation...

Now go to http://wallaceracing.com/Calculators.htm and get acquainted with all the different calculators on this site Primarily you'll be using this one --- http://wallaceracing.com/cam-deg-calc.php Guess I could of just posted that one, but I like all of the calculators on the main site... On your cam card you should have advertised durations at .004"... Don't use the .050" specs... Input the .004" duration specs into the calculator to get your opening and closing points of the cam...

Now go to https://www.lmengines.com/piston-speed-calculator/ Input your engine specs into it to get your "peak piston velocity" specs in degrees of crank rotation. I personally always use the engines idle speed as the rpm figure. Usually something like 750... Peak piston velocity of most ls motors is usually right around 73 degrees...

OK, so summing things up. We should now have the peak piston velocity of your engine and opening and closing points of the valves... SO we've got somewhere around a portion of the equation anyway?

Now plug all of these numbers into your degree wheel. I use print outs and pen just to better visualize things and make them easier when going back and forth to the EOI calculator which I'll post next... You should wind up with a degree wheel something like the one posted only with a 73 degree peak piston velocity and different opening and closing specs...

[GHuggins]

Now your going to need at least one dyno pull or wot pull in something like 3rd gear data logged for your injector pulse width numbers... But this is only part of the next equation... Your also going to need your injector offset and short pulse adder straight out of your tune. Go by the voltage that your battery dropped to during your wot pull for the offset table then what ever your injector ms is for the short pulse adder... Now use the following gen 4 calculator to come up with your new boundary and ect table numbers... Again this spreadsheet isn't mine, but works great on coming up with the right numbers...

Basically your goal here is to put "soi" injection timing above 2500 or so rpms right before peak piston velocity and bdc of the pistons stroke and eoi at these rpm points before the intake valve gets within 20 or 30 degrees of being completely closed... This assures best cylinder turbulence to atomize the fuel and will nearly always require fuel to be pulled back out due to better and more complete fuel burn...

Below 2500 or so rpms you will want to keep "soi" just before the exhaust valve has completely closed if not put it back on a closed intake valve. Use fuel trims or air fuel ratios while in open loop for this. Then smooth out your boundary table so there are no big jumps in injection timing. Makes for a smoother running engine throughout the rpm range.

You'll be retarding the injection timing doing it this way. Unless there's nearly no overlap "kinda like yours" this is usually what I see to be the best route to go. Your almost always better off to retard it in the higher rpms no matter what cam your using... In your case you might indeed see better fueling below 2500 rpms with it advanced...

[GHuggins]

Now even with all that being said if your running any sort of stock or maxed injector setup, you'll never be able to get "soi" anywhere near bottom dead center of the pistons stroke or as far as that goes anywhere near peak piston velocity... SO in these cases about the best you'll ever be able to do is retard it 15 to 20 degrees in the higher rpms and this will be for nothing else than to better optimize the engines fuel burn.

Secondly as Carson pointed out even with stock cams, when you throw something like a blower or turbo into the equation the added airflow and pressure going into the cylinder can and will nearly always blow the fuel vapor out the exhaust valve because after all it's now pressurized and can easily overcome the atmospheric pressure in the exhaust system. Retarding the injection timing in these instances will almost always save you fuel under boosted situations...

Plus when a turbo's involved it will usually cause the turbo to spoil 400 to 500 rpms sooner making in some instances an extra 100 or more hp in the lower rpm areas. I've fixed turbo cars breaking the tires loose at any speed problems (more or less un-drivable) by doing this and this alone. Turns a non functioning and pointless car into a missile Downside - well the turbo spoils up quicker putting you into PE easier and cost mpg... The extra fun makes it worth it though

Well this is how "I" do injection timing... I know some others do it this way, but as you've all ready seen there are tons of "personal" opinions and exceptions to this. After "ALL of this" it will boil down in the end to what ever shows you the best fuel usage and by this I mean what ever shows you that you need to pull fueling out. In other words where ever you put it for "peak efficiency".........

[GHuggins]

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?

There's also another table in the gen 4's for controlling how much fuel is added or taken out to promote fuel swings for the cats - it's the proportional - limit table - lowering this helps to control rpm surges caused by fuel swings from the O2's switching...

[itsdaveonline]

So that sounds easiest. I'll just throw some properly gapped TR6's into the motor, giver 'er hell and monitor KR readings. I already pulled the plugs once and they did not appear to show any signs of detonation. I'll post up some pics of the old plugs in here anyways.

If i get KR with TR6's that has got to be a false read given the setup, tune, and ultra conservative timing.

In regards to the injection timing ill have to start with getting my cam specs. all i have is @.050 specs But that by far seems like the most comprehensive explanation on how to actually tune it.

[itsdaveonline]

Ghuggins, you have helped a TON here. One last question and then I have probably exhausted your expertise and time.

Where are these tables your talking about? I see only two options under Closed loop proportional, neither quite fit the bill on what your describing.

2015-12-20 22_24_56-VCM Editor - C__Users_David_Desktop_CTSV tunes_Finish Line_nc tunes_11-3-201.png

[schpenxel]

You may only have access to one of the tables.. My E40 is the same way.

[GHuggins]

Go to hptuners.com and under contact - tech support, email Bill and see if they can't add ECM 12470 to your Tune. Usually if they can't add them it's typically an 08' and missing the proportional table... He'll need you to send a copy of your tune to them for clarification usually. At least you've got the proportional if they can't add the delay. You can always really dial it back...

It's the closed loop proportional - airflow mode table for clarification...

[schpenxel]

I hope they can. On mine they couldn't add it

Though they did add 12480, 12481, 12486 for me

[itsdaveonline]

Which ones are those if you dont mind me asking? I'm emailing bill now to see what we can do here.

[schpenxel]

Closed loop proportional Airflow Mode / O2 sensor / Limit

[sssnake]

Dave,

I have had similar issues. The other guys are likely helping but I think there may be more to it than those changes. I have a blown 07 that kept showing random knock. A LOT of the occurrences were when I was coming OFF of the throttle. With the transition to the new Beta I don't have a great shot at hand but the one below should let you know what is happening to me and likely to you.

2015-12-21.png

What happens is that when mixed mode operation is enabled the prediction coefficients filter the airflow measurements/predictions. In my case you can see in the third panel of the log MAF Cyl Air and Cyl Air (unfortunately during this time the Beta didn't work for Fuel Cyl Air and VE Cyl Air) diverge during the transient. From the way the knock is occurring and the way the models are diverging it appears the blended model is too slow for a FI car. One of my undertakings during the Christmas break is to get these to match more closely and see what happens in terms of knock.

Again these guys are NOT telling you something wrong. I have tried as much of their suggested COAs as possible with a mixed bag of results. The closed loop integrator delay is evidently not available on our cars (I have already requested). I have tried to adjust the closed loop proportional airflow mode with some benefit but not to the knock issues I was seeing. I believe until the larger dynamic airflow issue is fixed the benefits of adjusting the amount of fuel required for closed loop fueling oscillation is insignificant. I believe the same is true for EOIT. BUT THESE GUYS ARE MUCH MORE EXPERIENCED THAN I AM.

I may just be spinning my wheels and may be doing the adjustments in those two areas incorrectly but the sometimes large differences between MAF measured airflow, VE airflow, Fuel Airflow, and Dynamic Airflow (these are the analogs to the cyl airmass values I chart) are the major cause of the issues - at least in my car.

[itsdaveonline]

Sssnake - good stuff. I know you have helped me in the past too. Interesting theory with the blended / mixed mode potentially causing issues. Have you tested messing with the disable / re-enable RPM's for dynamic airflow? That's too bad we wont get the delay function.

With the o2 sensor tuning, i realize there is a good change it wont help knock retard, but after light reading it could still definitely benefit my overall drive-ability. I adjusted my min airflow a while back to fix most of the idle surging but im thinking the rest of the light surging will be fixed with the proportional airflow adjustments.

As far as EOIT im reverting back to stock for now as opposed to running it advanced. Calculating the correct one is another project! But at least i have the steps now. I never noticed any difference either with EOIT adjustments in the past for what it's worth.

Here's one suspicious spot i found, but apparently I wasn't logging cyl airmass (DOH!) It looks like the swing for the o2 sensors maybe changed the fueling too quickly or made it go a tad bit too lean. Just theory anyhow...Now i get the joy of touching up VVE & MAF again, but I cant imagine EOIT and proportional airflow will make a huge difference to adjustments on VVE & MAF.

knock2.png

[GHuggins]

Dave,

I have had similar issues. The other guys are likely helping but I think there may be more to it than those changes. I have a blown 07 that kept showing random knock. A LOT of the occurrences were when I was coming OFF of the throttle. With the transition to the new Beta I don't have a great shot at hand but the one below should let you know what is happening to me and likely to you.

2015-12-21.png

What happens is that when mixed mode operation is enabled the prediction coefficients filter the airflow measurements/predictions. In my case you can see in the third panel of the log MAF Cyl Air and Cyl Air (unfortunately during this time the Beta didn't work for Fuel Cyl Air and VE Cyl Air) diverge during the transient. From the way the knock is occurring and the way the models are diverging it appears the blended model is too slow for a FI car. One of my undertakings during the Christmas break is to get these to match more closely and see what happens in terms of knock.

Again these guys are NOT telling you something wrong. I have tried as much of their suggested COAs as possible with a mixed bag of results. The closed loop integrator delay is evidently not available on our cars (I have already requested). I have tried to adjust the closed loop proportional airflow mode with some benefit but not to the knock issues I was seeing. I believe until the larger dynamic airflow issue is fixed the benefits of adjusting the amount of fuel required for closed loop fueling oscillation is insignificant. I believe the same is true for EOIT. BUT THESE GUYS ARE MUCH MORE EXPERIENCED THAN I AM.

I may just be spinning my wheels and may be doing the adjustments in those two areas incorrectly but the sometimes large differences between MAF measured airflow, VE airflow, Fuel Airflow, and Dynamic Airflow (these are the analogs to the cyl airmass values I chart) are the major cause of the issues - at least in my car.

Have you tried zeroing out the coefficients all together? Then tuning in the VE table to correlate with the MAF? This should take care of the 2 or 3 second lean transition delay...

Dave as for the EOI. It makes a BIG difference on fueling... Of course that's solely based on cam specs... Do your injector calculations, dial in the closed loop proportional as much as possible then redo MAF and VE. AND when tuning the VE make sure to zero out the MAF table and 100% fail the MAF... Your VE table shouldn't have those ramps and dips...

If it is truly only on decel which I was under the impression that it wasn't then try installing a vettes decel idle table for your year model...

[itsdaveonline]

Roger that. Is there any benefit to still do VVE / SD tuning well over 4000RPM though? The way i understood is the PCM should be using strictly MAF data @4000rpm and beyond based on the dynamic airflow RPM enable / disable settings.

That big "ramp" is actually in the stock tune too, but you'd only technically hit that ramp if the MAF failed and you still decided to bring her to 6000RPM for some reason lol. I will clean up the VVE table either way, not trying to argue just for the sake of arguing. But just want to learn haha

[sssnake]

GHuggins,

I don't only get knock on decel but it it worse there (likely have pulled the timing too low in those sections to avoid knock which actually makes it worse). I was going to zero out the predictive but really what I would prefer to do is have it work properly. My MAF signal definitely is noisy (lumpy response even with an airflow straightener) so I do want to both filter the MAF and use a predictive model. Looking at the MAP and TPS slope over subsequent measurements should help get things much smoother BUT if nothing else works I will defeat them altogether. However, IME that seems to cause issues elsewhere.

On the EOIT settings. From my reading this seems to be more important with high overlap cams as they seem to differ from stock the most. Is that a somewhat accurate generality? The reason I ask is that I am seeing more issues with idle than I expected with a blower cam. I will re-read but the last time I went down this road I think I came to the point where I needed my cam card which I don't have...

[sssnake]

when tuning the VE make sure to zero out the MAF table and 100% fail the MAF

I forgot to ask, this recommendation is only if you are looking lamda/AFR to adjust the VE, correct? What I have done in the past is tune the MAF and VE simultaneously by comparing MAF Cyl Air and VE Cyl Air vs Cyl Air Mass times Lambda (correct the cyl air mass by lamda and then trim both MAF and VE as necessary). I don't believe there is anything wrong with this approach but I am definitely willing to listen.

[GHuggins]

I forgot to ask, this recommendation is only if you are looking lamda/AFR to adjust the VE, correct? What I have done in the past is tune the MAF and VE simultaneously by comparing MAF Cyl Air and VE Cyl Air vs Cyl Air Mass times Lambda (correct the cyl air mass by lamda and then trim both MAF and VE as necessary). I don't believe there is anything wrong with this approach but I am definitely willing to listen.

When you come right down to it, there's probably a million different ways to do the same thing. Is my way the best. NO. BUT it works best for me and my end tune result...

I have tuned "rushed or via email" VE tables without killing the MAF for others. Only problem is that it takes several runs as the VE is directly relating to the MAF transfer function. SO if the MAF is off, then so will the VE be off... As you get the MAF and VE both tuned in closer and closer the VE table will start looking more and more like it should (smooth without dips or ramps). I find it most accurate to 100% fail the maf and tune the VE table fully... I will post the current one I'm working on from a Z28 blower car. I've only done up to 2200 rpm tuning on it so far...

Your way isn't wrong. In fact I think some of the more well knowns on here do something similar, but via math equations and so on are able to tune in the air flow model on one or two passes...

As for the stock tunes having the ramps and humps and so on... This is only my opinion, but if you look at how tunes have progressed over the years it kind of backs it up. GM went through a big "change" where they got rid of their engineers and got in new younger guys... If you look at the older tunes compared to the year model tunes in which these guys started you will see big differences such as big humps in the VE tables "I believe this is offsetting the predicted coefficient issues", you'll also see PE tables that look like T-Pees to make up for air models being off, timing maps that jump all over and cause little "fish bite" sensations while cruising. These are just speculations and opinions, but there is some proof when you start looking... Now you look at the much newer tunes. Lots more smoother. VE tables without jumps or ramps. Somewhat decent timing maps... This is in "my opinion" why you see the big ramps in the VE table in the older stock tunes...

If you figure out how to tune the predictive coefficients then please share as I'm not sure anyone has quite yet...

Here's a supercharged ls7 I'm working on... I try to keep all of my air maps smooth as it helps with drive-ability.

What kind of issues have you had with zeroing out the predicted coefficients? Not sure if I've experienced any or not now?

I use a combination of the O2's and my air fuel sensor to tune these tables via my own equations and methods I've come up with. Not sure if it will show as usually my scanner config just shows not supported on these setups for other people, but I'll post it...

[sssnake]

Only problem is that it takes several runs as the VE is directly relating to the MAF transfer function. SO if the MAF is off, then so will the VE be off...

I ask this question not to pee in your cheerios but so that I understand what is going on. How is VE directly relating to the MAF transfer function? I thought GMVE was an airflow model for the engine that was independent of the measured MAF. I realize in mixed mode of operation the dynamic airflow is a mixture of the measurement (MAF), the model (GMVE) and the predictive stuff. I thought MAF and VE were independent.

[GHuggins]

You kind of all ready said it. When the maf is still in the picture, it uses both. In other words the maf and transients (predicted coefficients ) that will play into fuel corrections. A good way to show this is to kill the maf and see how much the fueling actually changes from when it was tuned with both.

Now with all that being said and you may be doing it this way, I know there is supposed to be a way to back calculate the ve table straight off the maf and then I think there's even a way to get both the maf and the ve table straight off of calculations. I'm not smart enough to come up with these equations and doubt even those who have would post them.

Don't worry about peeing in my cheerios. If there's a quicker and better way to do something, then I'm all for it ; ) I would just need you to explain and show like how I tried to explain the injection timing