Question about EFI cold start AFR

General engine tech -- Drag Racing to Circle Track

Moderator: Team

BLSTIC
Expert
Expert
Posts: 869
Joined: Wed Dec 05, 2018 7:14 pm
Location:

Re: Question about EFI cold start AFR

Post by BLSTIC »

I think it's worth investigating further why it idles higher with the same IAC commanded when cold. Does the valve have a position sensor that you are reading, could that be reading different when cold, or could it actually be opening further with colder temperatures? Is there some other dumbass thing like a coolant operated valve you've forgotten about?

I hunted a 'vacuum leak' for a couple of hours before finding out my Prelude engine had a second idle speed controller that was entirely coolant controlled hidden under the throttle when the entirely electronic one was the one in plain sight.

*edit I should clarify that this is largely because I doubt the air going through the same hole when cold is significantly more than when it's hot (it's 13% difference in mass, is that enough to increase idle speed 25%? What's your manifold vacuum doing under the two conditions?), and this is the first I've heard of an engine with the same ignition timing and idle bypass running faster when cold. It isn't the first time I've seen a bigger hole when it's cold though.
Belgian1979
Guru
Guru
Posts: 4576
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: Question about EFI cold start AFR

Post by Belgian1979 »

IAC is determined by the number of steps between fully closed and fully open (in my case 0-255).
The IAC valve is controlled by closed loop (PID).
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

BLSTIC wrote: Wed Nov 04, 2020 8:41 am I think it's worth investigating further why it idles higher with the same IAC commanded when cold. Does the valve have a position sensor that you are reading, could that be reading different when cold, or could it actually be opening further with colder temperatures? Is there some other dumbass thing like a coolant operated valve you've forgotten about?

I hunted a 'vacuum leak' for a couple of hours before finding out my Prelude engine had a second idle speed controller that was entirely coolant controlled hidden under the throttle when the entirely electronic one was the one in plain sight.

*edit I should clarify that this is largely because I doubt the air going through the same hole when cold is significantly more than when it's hot (it's 13% difference in mass, is that enough to increase idle speed 25%? What's your manifold vacuum doing under the two conditions?), and this is the first I've heard of an engine with the same ignition timing and idle bypass running faster when cold. It isn't the first time I've seen a bigger hole when it's cold though.
I don't think there's any other valve there.

I tried tuning the fuel to stoich at idle both engine hot and cold and, with the same IAC opening % fixed with an override, it runs happily when warm but stalls immediately when cold. This, plus the fact that it's seeing 1.5x bar boost every now and then, makes me believe it's not a vacuum leak. The MAF values are also sensible, slightly higher when cold but not dramatically so. Instead, I believe that I've just had the hot idle tuned to stoich and cold idle tuned to best power and the difference is larger than I expected.
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

Played more with the cold engine in the early mornings.

The pattern emerges that the engine needs a lot of fuel for the first minute or so to run well. After that, it wants to race out of the gate with low IAC opening % and lambda 0.85 mixture when the coolant is still cold but some of the internal parts have warmed up. Then when the whole engine with its long pipes etc. have warmed up, it needs more air with a higher IAC opening % and lambda 1 mixture.

In terms of fuel, the hot cranking pulse width is 2.25ms, it idles well with 2.2ms pulse width, but wants over 4ms of cranking pulse width when cold and then prefers to taper down fast to about 2.6ms and then slowly from 2.6ms to that 2.2ms.

I guess there's no mystery here.
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

So would it be fair to say that a fully hot engine should crank with about lambda 0.95 and then taper to about 0.95-1.00 lambda at idle and a very cold engine should crank at about 0.60 lambda, then taper fast to about lambda 0.80-0.85 for idle, and finally taper as the engine warms up to lambda 0.95-1.00? Would this be a good rule of thumb for future reference in port fuel injected engines?
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
Belgian1979
Guru
Guru
Posts: 4576
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: Question about EFI cold start AFR

Post by Belgian1979 »

imo there is no reference. Each engine wants something different.
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

Belgian1979 wrote: Sun Nov 08, 2020 3:38 pm imo there is no reference. Each engine wants something different.
So no useful guidance for a generic port-injected gasoline V8?

If every engine wants something different, how precisely should we measure whether the engine is getting what it wants? The leanest engine that the engine runs on, up to lambda one, or some other criterion?
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
RW TECH
Guru
Guru
Posts: 2387
Joined: Thu Nov 24, 2005 9:08 pm
Location: DETROIT, MI

Re: Question about EFI cold start AFR

Post by RW TECH »

Rick! wrote: Mon Nov 02, 2020 6:45 pm Those ran power mixture type AFRs to create driveability until the heated air intake took over.
This is pretty much the answer. As a starting point, you should consider .82-.85 lambda for idle after initial cold start, tapering towards whatever your engine is going to want for warm idle lambda. Sometimes idle wants to be .87-.90 and other times closer to stoic, depending on cam, etc.

Don't chase numbers so much, go after idle RPM and MAP stability, of course with most idle vacuum possible.
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

RW TECH wrote: Mon Nov 09, 2020 10:56 am
Rick! wrote: Mon Nov 02, 2020 6:45 pm Those ran power mixture type AFRs to create driveability until the heated air intake took over.
This is pretty much the answer. As a starting point, you should consider .82-.85 lambda for idle after initial cold start, tapering towards whatever your engine is going to want for warm idle lambda. Sometimes idle wants to be .87-.90 and other times closer to stoic, depending on cam, etc.

Don't chase numbers so much, go after idle RPM and MAP stability, of course with most idle vacuum possible.
Thanks.

I am going to run this close to stoichiometric at hot idle, because of three reasons. First, the camshaft overlap is very modest and I don’t believe much fuel or air gets blown directly to exhaust during the overlap. To the extent that unvaporized or incompletely burned fuel goes to the exhaust, it’s during the exhaust stroke, and that mostly happens when the engine is cold. Second, I want to keep open the option of turning on the narrow band closed loop fuel control in the future. Third, and most importantly, it runs idle very smoothly at lambda 1.00 — it just needs more air to do so than with a somewhat richer mixture. (With more camshaft overlap, this might not be as good of a plan.)

The other thing I am trying to do is set the hot idle AFR to be as constant as possible at all loads and rpms that the engine is likely to ever see at idle. There are all sorts of lags in the fuel system. I’ve unwittingly tuned the fuel table to create a cyclic idle behavior. Because of the lags, this just needs to be as constant AFR as possible, regardless of the level and change of the load and rpm.

In contrast to fueling, the ignition responds with minimal lags to rpm and there I have (and will continue to) set up the idle ignition curve as a natural proportional controller with ignition advance declining in rpm over the rpm range relevant to idle.
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
Belgian1979
Guru
Guru
Posts: 4576
Joined: Sat Mar 26, 2011 11:34 am
Location: Belgium - Koersel

Re: Question about EFI cold start AFR

Post by Belgian1979 »

RW TECH wrote: Mon Nov 09, 2020 10:56 am
Rick! wrote: Mon Nov 02, 2020 6:45 pm Those ran power mixture type AFRs to create driveability until the heated air intake took over.
Don't chase numbers so much, go after idle RPM and MAP stability, of course with most idle vacuum possible.
=D>
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

One of the things that I am noticing idling the car is that fuel is heating up. I’m basing this on the temperature of the fuel pressure line in the rear.

I did the fuel heating computation.  The Bosch “044” pump may be contributing to fuel heating, but it probably isn’t the whole story. Unless of course I made a mistake in my computations.

Suppose we have a half a tank of gas, or 43l.  For us to get gasoline boiling or about 5% leaning in terms of mixture lambda, we’d need to heat the fuel by about 50C from a typical starting point. This is because the volumetric thermal expansion coefficient of gasoline is (950/1000000)/K (or C) at 20C.

43l of gasoline weighs about 32.25 kg.  With specific heat of 2.22 kJ/(kg*K), we need 3580 kJ of heat energy.  Bosch 044 draws 11A at 13V pumping against 4bar, which by recollection of high school electrical formulas is 143W or 0.143kW.  kJ/s = kW, so even if all the electrical energy consumed by the 044 pump would be used to heat the fuel (maybe not too far fetched at idle), it would take 7 hours ((32.25*50*2.22/.143)/60/60 hours) of idling for the pump to raise the fuel temperature to 50C.  So it can’t be the hearing by the 044 pump alone that leans out the mixture by 5% in normal use.

Instead, now I’m thinking that some of the exhaust manifold, turbine, or exhaust components may be heating up the fuel.  The fuel does get hot, based on the sound of the pump and temp of the line, for example.  So maybe the solution to fuel heating is additional insulation or ventilation of the fuel line sections that are close to the exhaust components?
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
Orr89rocz
HotPass
HotPass
Posts: 2123
Joined: Sun Sep 28, 2008 9:25 pm
Location:

Re: Question about EFI cold start AFR

Post by Orr89rocz »

If the fuel is a return style system then going thru the rails mounted on a hot engine it will heat up

Also, from all the engines i have seen and some i have tuned from sbc, bbc, lt1 and some ls vehicles, they all seem to want to be a touch richer cold and closer to stoich warm. Some seem to run slightly better just richer of stoich. But they all like richer at cold idle and start up. I wouldnt say lambda .85 but could be for a few seconds in very cold environments. I usually dont need to be below .88-.90 lambda
arlancam509
Member
Member
Posts: 84
Joined: Thu Nov 07, 2013 8:23 am
Location:

Re: Question about EFI cold start AFR

Post by arlancam509 »

Rick! wrote: Tue Nov 03, 2020 7:43 pm
If this thread hasn't done much, it sure has brought out a lot of practical, hands on info. Hopefully it'll be put to good use.
this thread could not have come at a better time for me. i am directly in the cold start tuning and idle for my 'brand-new-to-me' fuel injection setup.
so, yeah, this information is absolutely going to be put to good use by me.

the only problem i am having so far is my wideband is showing some odd lambda numbers compared to what i think it should. i just have to keep playing with it. :lol:
° alt+0176 °
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

arlancam509 wrote: Wed Nov 11, 2020 2:25 pm
Rick! wrote: Tue Nov 03, 2020 7:43 pm
If this thread hasn't done much, it sure has brought out a lot of practical, hands on info. Hopefully it'll be put to good use.
this thread could not have come at a better time for me. i am directly in the cold start tuning and idle for my 'brand-new-to-me' fuel injection setup.
so, yeah, this information is absolutely going to be put to good use by me.

the only problem i am having so far is my wideband is showing some odd lambda numbers compared to what i think it should. i just have to keep playing with it. :lol:

My practical observations are the following:

- It's pointless to try to tune the cold start or warm-up enrichment before the car is exactly right when in operating termperature.
- Before doing any electronic logic trickery, the car needs to idle well (stable, no hunt) with just the equivalent of a "idle air adjustment screw"
- In my opinion, the way to get the car to idle well with just an "idle air adjustment crew" is to (a) tune the whole idle area (load, rpm) to give approximately constant AFR and (b) make the ignition advance a decreasing function of rpm in the range of idle rpms. The reason why this is that fueling has inherent lags in it, while ignition is almost without a latency. Making the fueling neutral and ignition reactive creates a natural proportional controller that has low latency and thus can't create cyclic behavior.
- Once it's set up like that, then it'll idle nice hot at almost any reasonable setting of the "idle air adjustment screw". Now, you can have a PID algorithm or some other similar electronic logic to pull the hot idle rpm to whatever number you want and it's going to be a manageable controller tuning effort because the underlying system is relatively stable to start with.
- Only after all this is working great one should start tuning the warm up enrichment maps that add fuel on top of this setup that works well when warm.

Those are all opinions, not facts, as goes without saying.
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
User avatar
ptuomov
Guru
Guru
Posts: 3587
Joined: Fri Aug 07, 2009 3:52 am
Location:

Re: Question about EFI cold start AFR

Post by ptuomov »

Orr89rocz wrote: Wed Nov 04, 2020 7:44 am Also note it is very common to add 2-4 deg timing during this time and idle speed demand being set few hundred rpm higher. Warm idle 25 deg and 800 rpm, cold starts may be better at 27-28 deg and 1000 rpm until it warms up
Here’s a question. I can see why adding timing in cold weather makes sense if the ability make air flow adjustments is limited. However, just from theory perspective, wouldn’t it make sense to retard timing when cold and increase the air flow to compensate for the power loss due to late timing? This way, the engine would heat up faster. Does this make sense?
Paradigms often shift without the clutch -- https://www.youtube.com/watch?v=cxn-LxwsrnU
https://www.instagram.com/ptuomov/
Put Search Keywords Here
Post Reply