Decay time for wall wetting | AEM
11 posts / 0 new
Last post
Def
Offline
Last seen: 1 week 4 days ago
Joined: 09/14/2014 - 10:21
Decay time for wall wetting

Is there a way to specify a decay time for wall wetting? I'm using both RPM vs. Throttle rate and Throttle% vs. MAP rate to try to get more granularity on my transient fueling. I feel like I'm getting pretty close, but my 3S-GTE seems to want a really long period of wall wetting, well beyond any MAP/throttle transient. 

 

For Megasquirt, you can put in a decay time which is fixed (which doesn't suit all cases), but it allows you to choose a middle of the road decay time before the engine reaches steady state. 

 

For my 2nd gen 3S-GTE, it seems like the average transient period is in the 0.3-0.4 s total duration range, with a linear decay from peak to zero of about 0.2 s. With throttle + map rate, I can spread the amount of transient fueling out (usually MAP rate decays faster than throttle rate based on my filtering), but I'm still left with a lean spike at the end of the transient event. 

 

Here's a quick screenshot to illustrate what I'm seeing, you can see that the transient fueling generally gets lambda1 AFR down to lambdatarget AFR, then the transient throttle/map rate decays to zero (wall wetting1 and 2 are lower gradient lines), but then lambda 1 AFR trends leaner from lambda target AFR for a brief period, then finally stabilizes to where steady state VE reflect actual VE (I tried to richen up the VE map to account for some of this transient error, but the error is too large to account for slightly richer VE values than steady state). 

 

I'm not looking for specific tuning advice, just more than I can't find a way to keep any sort of transient fueling active long enough to account for the actual engine transient period. The 2nd gen 3S-GTE has 8 individual intake runners leading to a common giant intake port in the head, so maybe this leads to a high amount of fuel wetting on the large port, but I am definitely running into an issue of keeping transient fueling going long enough via the tools in Infinity Tuner.

AEM_NS
Offline
Last seen: 12 hours 10 min ago
Joined: 06/06/2014 - 09:31
The short answer is no

The short answer is no because the wall wetting function in Infinity isn't time based.  You seem to know what you're doing so I won't go too deeply into it but I will say two things.  First, have you tried to just give it more?  You should be able to accomplish a really good delivery blend of enrichment fuel between dTPS & dMAP to the point where you get more later in the transient period.  Knowing that wall wetting has a decay behavor to it, more wall wetting fuel should last longer.  Second, can you even feel that "lean spike"?  I'm guessing you can't and that the engine wouldn't run any differently with that minuscule lean tick (I personally wouldn't even call it a lean spike) or if AFRs were spot on with AFR target.  Are you digging this deeply into it because you have a real engine tuning issue to resolve than results in poor performance or are just trying to get the two traces to line up with each other?  

One last bonus point - are you sure that you don't just need a tad more base fuel in that spot?

Def
Offline
Last seen: 1 week 4 days ago
Joined: 09/14/2014 - 10:21
The rest of the tune is

The rest of the tune is generally good (I've been tuning EFI setups for 18 years, Infinity setups for ~6-7 years). 

 

The base fuel is about 3-5% richer than steady state in all the 30-100% throttle areas to try to correct this issue (you can see newVE trend below current VE after steady state), it's even worse if I actually put in "correct" base VE values. 

Wall wetting isn't really a time based function at all, it only looks at the first derivative of either TPS or MAP inputs. You can see that the MAP rate (Wall Wetting2) input has decayed away well before the readout marker, while TPS Wall Wetting1 holds on longer with 85% TPS filtering, but it still decays away and gives a lean spike. As you'd expect, it only makes things much richer during the wall wetting event, and does nothing after it. There is no time based component holding any sort of wall wetting beyond what is commanded due to MAPrate/TPSrate inputs. 

I had an Infininty 506 on a couple of interations of Nissan SR20DETs/SR20VE+T's, and didn't experience as much of an issue with transient throttle fueling, so it's probably down to the very 80's intake design in the gen 2 3S-GTE... but it's still something I'm struggling to solve since the engine NEEDS more transient fueling duration than I can give it. 

It's not an academic exercise of having the two traces match up, as I'm having a hell of a time with initial spool up knock on my 3S-GTE, and it directly coincides with near stoich AFRs and ~8-10 psi. I've pulled timing way back, but on spool up I keep getting this extremely lean spike that I'm running out of things to try to get the transient fueling correct. I have commanded a bit more wall wetting, but it still snaps up 2+ full points above target AFR as it quickly decays away. So all I get from more wall wetting is going down to ~10 AFR on initial spool up, then an immediate lean spike after, then getting to steady state fueling. 

FWIW, a Megasquirt PNP2 could do this with a 0.3 s duration on its Advanced Enrichment fueling model (that EMS had its own set of problems, but did get the transient fueling down pat). The most I can really get out of the Infinity with input filtering etc. is about 0.15 seconds of duration for transient fueling.

AEM_NS
Offline
Last seen: 12 hours 10 min ago
Joined: 06/06/2014 - 09:31
Maybe I'm missing something

Maybe I'm missing something but the log snap shot you've shown doesn't really correlate to your description.

In your initial description your concern is a lean spike after wall wetting enrichment ends (2).  In your most recent description, you say you have knock during spool up with AFRs are near stoich (1?).  Since you didn't post a full log to review I can only guesstimate your AFR scaling but I'm guessing the target before going WOT is near stoich and your high load target is probably like .75-.78 Lambda so the lean spike at 2 is definitely not very near stoich.  Actually it looks like the spike at 2 is roughly the same as where the cursor is so call it .88 Lambda - does that sound about right?  A negligible lean out after enrichment ends is one thing but the lean spike at initial tip in should be fixable.  

Maybe what you have is tip in knock which is typically more common in higher compression NA motors but could also manifest in a quick spooling, small turbo engine.  You should be able to get the initial tip in fueling to go rich enough to blow smoke if you wanted but the other side of the equation is obviously ignition timing.  While there isn't a default function for it, you might try making an ignition trim table that references dMAP to pull timing during that quick spool up period instead of having to reduce base timing all the time.  Might be something worth trying.

AEM_SB
Offline
Last seen: 2 months 3 weeks ago
Joined: 06/07/2014 - 14:29
At low and medium RPM it's

At low and medium RPM it's probably worth spending a few minutes adjusting the fuel injector timing / phasing / end angle. Try changing the value for the FI_TimingX [degBTDC] table (the table is in the Injector page of the layout). Good numbers to try are +100, +50, 0, -50, -100, -150 and -200. Firing the injectors at certain crank angles may do a better job mixing fuel with the air going into the engine. The most noticeable difference tends to be transient response, because injecting a large amount of fuel early is better than injecting too late (after the air has already gone into the cylinder and/or the intake valve has closed). Some experienced tuners have found injector timing/phasing/end angle rarely makes a difference and others have found it can be worth power on certain setups, I'm interested to hear if this helps for the car you're working with.

Hope that helps,

SB

Def
Offline
Last seen: 1 week 4 days ago
Joined: 09/14/2014 - 10:21
Thanks for the detailed

Thanks for the detailed responses - lots to unpack here. 

Phase 1 leanness does not seem consistent, and I think it's a recent change of adding a 0.040" orifice to the MAP sensor line and reducing its filtering value - with no orifice and high filtered I had a horrible lag on throttle inputs that caused prolonged lean conditions. I think really quick MAP changes will show a small lag in MAP sensor response. But this is at generally low MAP values and is only observed for a few cycles before the MAP sees the change in pressure. The MAP signal at WOT is super jumpy on this engine at WOT, so it needs something to slightly calm it down. Choose your poison type decision here...

Phase 2 is the primary issue that causes intermittent knock, even with low timing values. I'm reading off the 3S-GTE distributor, which is driven through the timing belt. I think random timing belt induced timing scatter + this momentary lean spike (it peaks at ~13.7:1 AFR in this trace at ~8 psi of boost. In the words of Cherynobyl, "not great, not terrible." That plus timing scatter = knock on my old tractor of an engine. 

 

I don't have a short log to share, so I won't try your patience with a 60+ MB log. 

Generally I think I've found the relative happy place for this engine (0.040" MAP orifice, the right combination of TPS + MAP sensor filtering, good transient fueling), but I do still occasionally see this later lean spike with a very fast throttle input. If the throttle input last another 0.1-0.2 secs, everything is golden all the down to steady state AFR, with almost perfect transient fueling. The fast throttle stabs give this concerning lean spike that are also causing some real knock events - which is why I'm concentrating so much on them. 

 

 

Thanks again for the help guys, I'll come back in with logs if I can't solve it, or if I do find something that works for me, I'll post it as a future learning opportunity for everybody. 

cutlassefi
Offline
Last seen: 1 month 13 hours ago
Joined: 05/13/2016 - 06:06
AEM_SB wrote:

AEM_SB wrote:

At low and medium RPM it's probably worth spending a few minutes adjusting the fuel injector timing / phasing / end angle. Try changing the value for the FI_TimingX [degBTDC] table (the table is in the Injector page of the layout). Good numbers to try are +100, +50, 0, -50, -100, -150 and -200. Firing the injectors at certain crank angles may do a better job mixing fuel with the air going into the engine. The most noticeable difference tends to be transient response, because injecting a large amount of fuel early is better than injecting too late (after the air has already gone into the cylinder and/or the intake valve has closed). Some experienced tuners have found injector timing/phasing/end angle rarely makes a difference and others have found it can be worth power on certain setups, I'm interested to hear if this helps for the car you're working with.

Hope that helps,

SB

X2. Have you tried using Fuel Trim 1 or Fuel Trim 2 tables? You can set one of the axis to either "MAP Rate" or "TPS rate". If you set the graduation in small increments you MAY be able to accomplish what you're looking to do..I've been fooling with these tables off and on with some success for other issues.

Def
Offline
Last seen: 1 week 4 days ago
Joined: 09/14/2014 - 10:21
I am using both MAP and TPS

I am using both MAP and TPS rate. It seemed referencing TPSrate vs. RPM and MAP rate vs. throttle% gave me some ability to control over those two conditions since the wall wetting is not constant across both of those parameters.

 

 

I need to tweak fuel injector timing once I get the car back and going. My 30 yr old 3 piece Work wheels were leaking like crazy so the car has been down getting that fixed. 

cutlassefi
Offline
Last seen: 1 month 13 hours ago
Joined: 05/13/2016 - 06:06
Gotcha. But what are you

Gotcha. But now what are you going to do to address cold running fueling?

Def
Offline
Last seen: 1 week 4 days ago
Joined: 09/14/2014 - 10:21
Cold running? Like the

Cold running? Like the LambdaAfterStartTrim? Typical table there, and the Honda F20C1 "base cal" I used seemed to have reasonable values for that. 

Low coolant temp transient conditions are such a small amount of the total driving that I'm not too concerned with them. If I were tuning for cold start emissions - I'd care much more - but I'm not. 

cutlassefi
Offline
Last seen: 1 month 13 hours ago
Joined: 05/13/2016 - 06:06
Gotcha, makes sense.

Gotcha, makes sense.