Pictures of my inlet ports....

[silverRS4]

Good direction everyone. Misfires are usually attributed to excess fuel or fuel vapor. Some RS4 cold-start issues were due to failed injectors that were seeping fuel into the cylinders under the residual pressure in the fuel rails. The cars with failed injectors also had carbon buildup (discovered when the manifold was removed to replace the injectors). This is a rare instance though. The hypothetical engine, and many other hypothetical engines like it, had cold-start misfires but the injectors passed leak-down tests. In other words, the injectors were not the direct source of incorrect fueling.

Less timing retardation is simply less correction in a closed-loop system. Base timing is not "advanced". The target has not changed. The end result in the hypothetical engine with clean valves are conditions in which the actual timing is closer the target. The target being the ignition angle determined by the engineers to be the best angle for peak torque (since we are speaking of WOT here). I really do not know how else to say it. Yes, all engines are the same, but all open-loop (look-up tables only) and closed-loop (feedback with correction) ignition systems are not. A few degrees of less correction is far more important than a few grams per second of airflow when considering engine output, that is why I asked. I also asked the question (to which I do not know the answer) simply to refocus the knowledge base.

2manytoys - my hypothetical engine is not pinging or knocking. That is a serious problem. Has Audi driven the car and witnessed it doing so under light loads? Your "conditions" have created a situation where there is no longer enough correction. Either the closed-loop system is failing or there is something very wrong with the fuel or the engine.

[PetrolDave]

But I'm curious...

Would you object then if a manufacturer managed to build a DI engine that didn't deposit so much carbon in its manifold?

That'd be presumably a waste of time would it?

No I wouldn't object. I remember my Dad having to take the head off the engines on his cars and de-coke them every 10,000 miles or so - we don't have to do that any more. Similarly any design improvement that reduces the deposition rate can only be a good thing for longevity of the engine.

I don't understand why anyone would not consider such an engine a better product?

I've never said, nor even implied, that I hold such a view.

I wonder if the Veyron suffers?

I bet it does, it's a "feature" of all DI engines from all manufacturers - unless a solution has been developed for the Veyron that is simply to expensive for other applications (and I've never heard nor seen anything to suggest there is).

[S2tuner]

The Veyron doesn't have direct injection.

5B0906012
J623 ME911
W16 8.0 1004
0261S02025
1037391140
A00.000 P0000

ME9.1.1 with no D means no Direct injection.

[2manytoys]

2manytoys - my hypothetical engine is not pinging or knocking. That is a serious problem. Has Audi driven the car and witnessed it doing so under light loads? Your "conditions" have created a situation where there is no longer enough correction. Either the closed-loop system is failing or there is something very wrong with the fuel or the engine.

Yes, absolutely. From memory it took Audi almost 10,000kms I think before they discovered it was carbon buildup.

As the pinging got worse they started to take the problem more serious (at least they could hear it themselves). When I first told them it felt low on power they just tried octane booster and said drive it for 1000km. In fact, they said it felt ok.. WTF. Ultimately it wasn't the timing being pulled (even though it was the main contributer to the power loss) that sent them down the carbon path, it was the air volume logs that did.

Just to be clear, if the carbon problem is left too long you'll probably start to get pinging. I think most people catch it early and start to get the problem addressed. And believe it or not, I'm a pretty relaxed guy so I wasn't pushing them as hard as someone else might (given I didn't know what it was, and neither did Audi at the time... maybe the lack of RS4's in Oz didn't help)

[mark758]

OK give up the times man, the mag got ~10.7 (RS4 vs C4S)
that's one of the better times avg is ~11.1 or so
40k miles and never cleaned? nice

what's yours?

Both runs were in the 10.9s....have to confess I wasn't driving though

[ArthurPE]

OK give up the times man, the mag got ~10.7 (RS4 vs C4S)
that's one of the better times avg is ~11.1 or so
40k miles and never cleaned? nice

what's yours?

Both runs were in the 10.9s....have to confess I wasn't driving though

that's fast!...< 11 sec is great...
40k miles and never cleaned....
and you must be good with the gears/clutch

I just had a 'rapid ride' in mine, nice twisty back road, up/down, tight turns, high speed turns, straights...the whole gambit...
the car never ceases to amaze me

[SR71]

Yes, all engines are the same, but all open-loop (look-up tables only) and closed-loop (feedback with correction) ignition systems are not.

I'm not sure I understand this statement. Could you expand?

Is there anyway to tell from VAGCOM whether the car is operating in open or closed loop mode?

Looking at http://www.ross-tech.com/vag-com/m_blocks/ gives us an idea of the measuring blocks we can look at...is there a correlation, say, between the "Operating Condition" (Group 60, Block 4 on cars with E-Gas) and whether or not the ignition is operating in open or closed loop mode?

In for example, the figure below, a "Full Throttle" run, is the value of Group 3, Block 4 (Timing Angle) merely the value in an "open loop" lookup table dependent on revs, temperature etc etc? i.e., if the values of <beep> in Group 20 are all zero, it is fair to assume the car is operating in open loop mode?

If this is the case, then it is very curious how, for example, my car and Rob's, pre-modifications could run such different base timing? What variables were so different between the cars that you'd get 10 degrees of difference?

I'd always thought that, regardless of whether the car was running open or closed loop, knock control feedback was "always" being done, and that this might have an effect on base timing even if the values of Group 20 were zero....

Of course, I appreciate that this means the car never really runs "open loop" with respect to knock control.

To my mind, ignition <beep> due to knock control is easy to understand...if the piezo's detect certain frequencies, back off the timing.

However, if there is no knock sensor input into base timing, this puts another complexion on things...

[silverRS4]

Yes, all engines are the same, but all open-loop (look-up tables only) and closed-loop (feedback with correction) ignition systems are not.

I'm not sure I understand this statement. Could you expand?

That entire paragraph was directed at the person who is generalizing the RS4's response to less than optimal ignition scenarios to be like all other engines. I do not know what the exact disconnect is, but his understanding of how a closed loop system (in a RS4 specifically) operates is not quite right. The RS4 timing control is strictly closed-loop. If you consider this, the understanding of your logs should be very straight forward and many of your questions answer themselves. I should point out that "base" timing can't be directly logged. If I understand you correctly, you are taking base timing to be the target that the closed-loop system is trying to achieve. The timing angle logged with Vagcom is the timing that is actually occurring. Base timing plus retardation value equals actual timing angle. So the only way to see what the base (or target) is would be when the retardation is zero for all cylinders, then base=actual timing angle. For instance, it would be pretty safe to say the target ignition timing angle in rows 19-28 of your log is 33.8 degrees. This occurs only when the retardation value is zero for all cylinders, so either the intake air temp is fairly low, and/or you are using 98+ RON. There are (4) sensors, each monitoring two cylinders, but retardation can only be logged one bank at a time. So between precise timing control in each cylinder and constant lambda control, the RS4's engine operation is extremely accurate. But it's also very sensitive, which is the underlying point of my hypothetical question.

The only odd thing I see in your log is a peak of only 152 g/sec of mass flow. At sea level, I'd expect it to be near 160. All corrected logs I've seen of clean RS4's show a mass flow of right around 160 at 8000 rpm. You mentioned once about swapping mass air sensors with another RS4 owner. Did you try that?

[ArthurPE]

all modern timing systems are closed loop, ie, knock sensors...they are a reset, or secondary control variable, a safety, not the primary variable...
there is no 'disconnect
ALL engines operate on the same principle...
90%+ of ALL ECU's are made by 3 companies, who also do the programming, obviously in collaberation with the engine design team...
the maps are derived/fine tuned on an engine dyno, after initial values are arrived at by computer simulation/historical data...

the point is, how would depostis affect timing? by what physical mechanism?
and are we talking about
detonation
pre-ignition
mis-firing
???

we've lumped all these conditions into some kind of timing control issue...
again, if we are talking about mass flow using the cars sensor, 10% resolution is the BEST you will get, in other words, if you are comparing 152 vs 160, it's not valid...

[ArthurPE]

actually, they are pretty much an 'open loop', since the primary control variables pick a value from a map...
closed loop implies to me (and pretty much industry std) that the primary variable (rpm, load, temp) etc control via a feed back signal, ie, closed loop...that is not the case
this is an open loop system with over-ride

the lamba system is not really a closed loop feedback system, it 'trims' or resets the selected map value also

a true closed loop system
a pump pressure loop
PID = 'controller'
input = Q setpoint
feedback = actual, measure flow Q
output = CV = vfd (variable freq drive) speed signal = pump speed

say you desire the pump output to be 10 gal = setpoint
pump starts up, as it ramps up Q > 10 as sensed by Q sensor
that is fed into a PID as negative feedback, compared to the set point, and an output signal is derived and sent to the VFD, ie, 'slow down', signal = smaller
if Q < setpoint, the converse will happen, signal = speed up or larger
this cycle repeats until the system is stable (if tuned properly) at setpoint

the error or output signal from the PID will be derived from P, proportional or offset error, I intergral error (summed over time) and D derivative error, rate of error/time

big difference vs knock control...
knock control is not related to set point
rpm, load, etc is, via the map
it can only <beep> or 'reset', the map value

imo the reason true closed loop knock control is not used, it's too slow, open loop positioning is much faster...

[SR71]

SilverRS4,

I think my understanding was correct.

I also think the terminology is confusing and "closed loop" means different things to different people.

If my understanding is correct, I agree with Art that the control of knock isn't strictly "closed loop" in the sense that a Control Systems engineer might use the word. There is no error term being fed back to recalculate timing in the presence of knock. All that happens is that once the knock sensors are triggered, 3 degrees of <beep> is knocked off the target to see if that "solves the problem", followed by a further 3 degrees etc etc

Still that is to some extent semantics.

As for my car, I've sold it.

I discussed with you the oscillatory nature of my measured airflow when running SW0080...the values of airflow were not monotonically increasing with rpm. I never got to the bottom of this problem, inspite of replacing the MAF, but reflashing to SW0090 solved the problem.

However, it didn't mean my car flowed anymore air, still an absolute corrected max of 155g/sec.

If you remember Anto_RS4 observed the same behaviour.

(If Art was correct about the resolution of the MAF, I would expect the maximum observed value on a run would vary by ~15g/sec but it never did... Neither did it change depending on what MAF sensor was in the car (I tried a few), and whatever software it was running...the maximum corrected value was always 155g/sec. I understand hot wire anemometers can be accurate to 1% in laminar flow?)

Rob's car flowed ~166g/sec, yet his achieved timing was almost 10 degrees less than mine, which meant on a FATS run (4200-6500rpm) my car was approximately 0.2 secs quicker over the interval than his. This was, of course, prior to having the manifold cleaned, ported and the exhaust system modified.

So, extremely crudely, ~6% better airflow (Veff), but ~5% slower (proportional to accel^1/2) means you can work out what a degree of advance is worth assuming a linear acceleration across the range...

[ArthurPE]

a true feedback/closed loop knock system would function as follows>

start with a base setpoint, say 10 deg
measure knock

001 if knock < sp, sp = sp + 1, else 002
002 if knock > sp, sp = sp - 1, else 001

if it is <> sp it would loop with no change

this way timing would always be on the threshhold of detonation, and maximum power
this is not done for obvious reasons: mostly speed of control, ie, response

timing control is open loop with a safety reset (knock)
if it were a closed loop system timing would be the feedback signal and the setpoint

control theory is a passion of mine...I know, I'm weird...
both analog and digital

[S2tuner]

Art, if you can read German, I can pass on a few things to you for you to understand better knock control on a MED9 ECU. If you don't, no point.

[2manytoys]

What I don't get; is if air flow is not that critical (according the Arthur) then why have Intake Runner Flaps. Why bother with moving them at set RPM's too?

I get the whole "make the air enter correctly so when fuel and spark happens it burns right" I also get that the RS4 is so precise that it needs Intake Runner Flaps to produce the correct burn (and max power), then any amount of carbon could mess this up. A lot of carbon is going to mess this up to a point that the burn is not going to happen as intended, hence timimg being pulled back (on knock detection).

Surely this seems logical?

[PetrolDave]

The purpose of the intake runner flaps is both to make the airflow more turbulent and to direct it to one or both valves. This helps create a small "pocket" of air:fuel mixture in the vicinity of the spark plug that has a sufficiently fuel content to ignite even with a very weak air:fuel ratio taking the cylinder as a whole.

Since the airflow is not equal across the whole inlet area, especially when the intake runner flaps are operated, carbon in places where there is minimal airflow will have little effect on the volume of air entering the cylinder - and hence little effect on power.

It's important to remember that with this engine intake airflow is not laminar, and the fuel mixture is not homogeneous.