Understanding FA20DIT Technology: cam-torque AVCS, EGR, TGV, DI

[arghx7]

EDIT: A thread discussing tuning TGV & EGR delete is here: http://forums.nasioc.com/forums/show....php?t=2743792

I'm creating this thread to explain and open up discussion about the new or modified technology of the FA20DIT engine. For simplicity's sake I am focusing here on the stock FA20DIT as found in a 2015+ WRX running stock tune with no hardware modifications. There is also a section about cold start combustion on TGV deleted 2016 models.

Here is the general order of topics I had in mind, that I will introduce over time:

1. new cam-torque actuated AVCS phasers with mid lockpin position

2. Atkinson cycle operation with AVCS

3. AVCS phasing for scavenging during WOT spool

4. High pressure Cooled EGR operation

5. TGV operation on the FA20DIT

6. Direct injection related topics (fuel pressure, injection timing, that kind of thing)

7. Appendix: Cams and valve timing, Combustion quality metrics and combustion enhancement

FA20DIT Cam Torque Activated AVCS with Mid-Lockpin Technology

A cam phaser (called AVCS for Subaru, but the technology can be also be called MIVEC, VANOS, VVT, VTC, VCT, i-VTEC, all sorts of sometimes confusing names) moves the centerline of a camshaft in order to change the valve events. More information on cam centerlines is available in the Appendix. The FA20DIT has cam phasing on both the intake and exhaust cams, but the phasers themselves are much closer to what's found on the BRZ engine than on an older EJ. [font=Verdana]First, here is a description showing the conventional AVCS phaser on an EJ255 when it is advancing the cam:


Source: 2005 Legacy GT Service manual

The conventional phaser uses oil pressure, essentially the energy from the oil pump, to move the vanes. When the engine shuts off the intake cam locks at the most retarded position. This is important to remember because the FA phasers don't lock in that way, which I will come back to.

The FA20DIT uses a cam-torque actuated phaser. This type of phaser still uses oil, but instead of using the hydraulic energy of the oil pump, the oil sort of “sloshes” in the chamber as the cam rotates, using the momentum of the cam for phasing:



Source: Smith, "A Cam-Torque Actuated Vane Style VCT Phaser" 2005, SAE 2005-01-0764

This image intended for BRZ service gives more explanation of how the cam-torque AVCS operates when advancing the cam:


Source: BRZ Service Manual

So the engine oil uses the "sloshing effect" inside the chamber to transfer energy from the cam motion to the phasing operation.

That leaves one more technology to discuss which is important for understanding the AVCS maps inside the ECU: the mid lockpin. As you will see later, the stock intake cam AVCS actually uses "negative" cam advance during low load operation, retarding the intake cam from its default position rather than advancing it. This is because the locking pin is not at the most retarded position:


Source: BRZ Service Manual

If the intake cam AVCS locked at the most retarded position, the engine would have trouble starting due to such a low effective compression ratio (please see appendix for explanation on effective compression). The next section will show some datalogs and maps, discussing AVCS operation for Atkinson Cycle with low effective compression ratio.

* Registered users of the site do not see these ads.

[arghx7]

“Someone told me that each equation I included in the book would halve the sales. I therefore resolved not to have any equations at all...” Stephen Hawking, A Brief History of Time

I’m going to throw a lot of pictures in here because I think that’s the easiest way to learn. Many of them are not my original work.

Atkinson Cycle Introduction: PV Diagrams

First, here's a brief explanation on Pressure-Volume diagrams.


Source: Asmus, "Valve Events and Engine Operation," 1982, SAE 820749

The X axis is the piston position and the Y axis is cylinder pressure. The top diagram shows the compression and expansion loop, which is what generates torque. The area under this curve is called the Indicated Mean Effective Pressure. The lower diagram zooms in on the “Pumping Loop.” The area under this curve is called Pumping Mean Effective Pressure. The more area under this curve, the more we are wasting energy working against a closed throttle, working against exhaust backpressure, or working to move gases in and out of the cylinder. So remember this simple rule:

Make the Power loop big and the pumping loop small and your engine operates more efficiently.

At part load, one of the main purposes of AVCS is to make the pumping loop smaller. To that end, there are two part-throttle AVCS fuel economy strategies I want to discuss here.

Early Overlap Strategy – EJ255

The first is the “early overlap” strategy. Here is a diagram (not with actual Subaru valve lifts) showing that type of cam phasing:


Source: original image

You can see the piston still rising as the intake valve opens, meaning that exhaust gases will be pushed back into the intake port in low load operation. This creates internal Exhaust Gas Recirculation (EGR). The inert gases help reduce NOx emissions, but they also dilute the mixture in the combustion chamber. By diluting the mixture, the throttle valve needs to be opened more to supply fresh air. With the throttle open more, the manifold vacuum decreases (manifold absolute pressure increases), which in turn reduces the pumping work. However, we are also closing the intake valve earlier and increasing effective compression (see the Appendix on this). That requires more throttling, which can offset the pumping benefits of the overlap.


Source: Leone, "Comparison of Variable Camshaft
Timing Strategies at Part Load", 1996, SAE 960584

An EJ255 has port fuel injection and intake AVCS only. This AVCS map from an EJ255 2006 WRX shows what we could assume to be the early overlap strategy, especially at 2000-2400rpm:


Source: OEM Subaru Map extracted from Cobb Software

Late/Late Atkinson Cycle – FA20DIT

For part load fuel economy in that same 2000-2400rpm range the FA20DIT uses late intake valve closing Atkinson Cycle to reduce pumping work, and late exhaust valve opening for higher expansion ratio. The late intake valve closing reduces the effective compression ratio and also creates that dethrottling effect. The late exhaust valve opening increases the expansion work.


Source: Leone, "Comparison of Variable Camshaft
Timing Strategies at Part Load", 1996, SAE 960584


Source: Mokhtari, "Combustion System Design of the New PSA Peugeot Citroën PureTech 1.2 e-THP Engine," 2014, Aachen Colloquium

And here are corresponding AVCS tables from a 2015 WRX (note the negative intake cam phasing):



The negative values are there because the intake cam is being retarded from its mid lockpin position in order to reduce effective compression ratio more than would be possible if we just parked the cam at the “0” position. Now look at the exhaust AVCS table:


Source: OEM Subaru Map extracted from Cobb Software

In that same area where we see negative numbers for the intake valve, we see large positive numbers for the exhaust valve. A negative number for the intake valves corresponds with retarding, and a positive number for the exhaust also corresponds with retarding. So we have a “late/late” (late exhaust valve opening, late intake valve closing) approach to fuel economy improvement by dethrottling an introduction of internal EGR at low loads.

Here’s a log showing the Atkinson cycle in action on a stock 2015 WRX (these logs were supplied to me, it is not a personal vehicle).


Source: Original Image

Again, we know it's Atkinson Cycle because of the large number of degrees of exhaust cam phasing (pink line) and the negative number for the intake cam phasing (green line). There is gradual accelerator pedal movement (shown here in the darker blue requested torque trace) until the tumble generation valve triggers open (step change in light blue line). EGR valve is closed because the ambient temperature was very low at the time.

As we have now established, the Atkinson cycle uses late intake valve closing for very low effective compression ratio. The intake cam uses “negative advance” because the “0” position is set with high enough effective compression that the engine can reliably start. The image below illustrates the mid position of the phaser (right part of diagram), the lockpin position exhaust phaser (solid red line), and the lockpin position of the intake phaser (solid blue line).


Source: Watanabe, "The New Toyota 2.0-Liter Inline 4-Cylinder ESTEC D-4ST Engine" 2014, Aachen Colloquium

[arghx7]

FA20DIT AVCS at WOT

Direct injection allows more aggressive use of scavenging to help spool the turbo. Recall that scavenging occurs when the pressure at the intake port is greater than the pressure at the exhaust port during overlap. Fresh air blows right through the cylinder, evacuating residual gases. This helps with knock relief and also increases mass flow to the turbo.

On a PFI engine such as an EJ, scavenging is limited due to the raw fuel that will be thrown out the exhaust valve. With a DI fuel system the correct injection timing will begin spraying fuel after the overlap period is over, so that this “short circuiting” does not occur.

Depending on how the stock ECU controls fueling, scavenging could result in a lean AFR in the exhaust gases. "Lean" here refers to leaner than lambda=1 or a stoichiometric AFR somewhere in the 14's depending on the ethanol content of the fuel. Other times the term "lean" can mean simply "leaner than I feel it should be, but still at lambda~1 or richer." That's not the sense being used here.

So lean on the wideband does not necessarily represent what’s happening in the cylinder. To re-iterate: certainly during spool or low speed/high load, just because the wideband says you are lean does not always mean the engine is actually running lean in the cylinder. It could be skewed due to fresh air blowing through during overlap.

We can see this on the stock tune. Here is a WOT pull in 2nd gear:


Source: Original image

Between 1 and 2 seconds we see a lean spike, which corresponds to movement of the cams. You can see it again in this 6th gear pull:


Source: Original image

Toyota’s new 2.0 liter direct injected turbo engine for the Lexus NX uses an onboard algorithm to correct the stock wideband reading according to the amount of scavenging (there is no evidence at this time that Subaru is using such an approach):


Source: Watanabe, "The New Toyota 2.0-Liter Inline 4-Cylinder ESTEC D-4ST Engine" 2014, Aachen Colloquium

[arghx7]

High Pressure Cooled EGR Operation

The FA20DIT uses cooled high pressure EGR for improvement of fuel economy. We call it high pressure EGR because it relies on the high pressure before the catalyst (or turbo in this case) to drive EGR to the intake manifold, which is after the high pressure side of the compressor. High pressure cooled EGR is typically used at higher unboosted loads. Low pressure cooled EGR on a turbo engine draws exhaust from after the cat and routes it in before the compressor inlet. It is typically used at medium-high boosted loads, like accelerating at part throttle boost, but is often not used under WOT acceleration.


Source: Takaki, "Study of an EGR System for Downsizing Turbocharged
Gasoline Engine to Improve Fuel Economy" 2014, SAE 2014-01-1199

You can see the EGR pipe bolted to the back of the intake manifold on an FA20DIT:



In the old days you would see an EGR valve using uncooled high pressure EGR to reduce NOx emissions. Now that same effect can come about with hot internal EGR from creating overlap with AVCS. That’s one of the reasons why EGR valves went away in the mid 2000s. So the main fuel economy benefit of external EGR over just using internal EGR comes from the cooling effect of the EGR cooler. Under certain conditions introducing cooled inert gases can be more effective and efficient than just introducing hot inert gases through overlap.


Source: Takaki, "Study of an EGR System for Downsizing Turbocharged
Gasoline Engine to Improve Fuel Economy" 2014, SAE 2014-01-1199

Unlike using AVCS for internal EGR, an external EGR system relies on pressure difference and a valve to drive EGR flow, which makes it harder to use at really low loads. You can see the EGR in action from this datalog where the accelerator pedal is slowly applied (red requested torque trace). At the lowest loads the engine runs full Atkinson Cycle operation. We know this from the negative value for AVCS intake cam (green trace).


Source: original image

As requested torque increases, the intake cam phases to close earlier and raise effective compression ratio. The EGR valve (black line, lower chart) opens and EGR flows according to the pressure difference between the exhaust manifold and intake manifold. This improves cooling loss, pumping loss, etc. You can see the engine running with only a very slight amount of vacuum (maroon trace, top chart) due to the dethrottling effect of the EGR. As requested torque increases further, the EGR valve closes and boost increases to greater than atmospheric pressure.

One of the side effects of introducing cooled EGR is that it slows the combustion inside the cylinder introducing these inert gases (hot internal EGR does that too). The stock FA20DIT ECU has some various tables to advance the spark timing to account for the slower burn, including this one:


Source: OEM Subaru Map extracted from Cobb Software

The resulting additional spark advance seems to be a bit aggressive, at least in the datalog below, probably for the sake of fuel economy. It relies on knock feedback (dark green line, lower chart) to retard spark if needed.


Source: original image

The FA20DIT also uses the tumble generation valve (TGV) to speed up and stabilize the burn in the combustion chamber, in order to tolerate this cooled EGR (and the Atkinson cycle/internal EGR) for better fuel economy. That topic will be covered in the next section.

[arghx7]

TGV Operation on Stock FA20DIT Engine with Stock Tune

The fuel economy approaches on the FA20DIT such as The Atkinson Cycle and cooled EGR slow down and destabilize the burn, potentially causing misfire or rough running. The whole point is to degrade the quality of the burn in the name of fuel efficiency. That’s where the TGV comes in. The video below shows how a TGV allows a direct injection engine to mix fuel in the cylinder much better. Starting at about 0:25 the TGV closes and the fuel mixes better in the cylinder, speeding up and stabilizing combustion.

YouTube (Short URL)

YouTube (Short URL)

On the earlier EJ engines, the TGV was purely used for cold start emissions, so that the engine could tolerate retarding the spark to heat up the catalytic converter. Later 2008+ EJ engines had to meet stricter emission standards (California LEV II). They used the TGV more, but they still had port fuel injection (which is easier to mix than DI) and the engines used a different AVCS and internal EGR strategy at part load.

The FA20DIT has separate timing maps for TGV open and closed, and we can see major differences in low load areas. Below are the base ignition maps for a 2015 WRX with TGV open and TGV closed. The dynamic advance maps (which are added on top of the base values, depending on knock activity) for TGV open and closed are pretty close at lower load areas where the TGV is actually used. I did not post them here.


Source: OEM Subaru Map extracted from Cobb Software


Source: OEM Subaru Map extracted from Cobb Software

The higher numbers on the TGV Open maps indicate the spark plug firing earlier to account for a slower combustion speed from poorer motion inside the cylinder. We also can see the difference in spark timing between these two maps in actual vehicle logs showing the TGV valve moving from a closed (high tumble motion) to an open (lower tumble motion) state:


Source: Original image

The above log shows the opening of the TGV at part load. The ambient temperature is very low and thus the EGR operation has been disabled by the stock ECU. You can see the TGV open (light blue line in bottom chart) by the voltage switching high at about 5 seconds. Open TGV increases airflow capability of the head while reducing the tumble flow and mixing inside the cylinder. Now notice that the spark advances about 10 degrees at the same time the TGV opens. This is because the combustion slows down with an open TGV.


Source: Original image

The second log, shown above and taken from the previous section about EGR operation also shows the effect of the TGV opening. Notice at the beginning of the log that the spark advance is about 35 degrees even with the engine speed at 2100rpm and the TGV closed to speed up the burn. It demonstrates the mixture burning slowly in the cylinder with the late exhaust valve opening, delayed overlap, and late intake valve closing Atkinson cycle. With the engine mostly dethrottled (running almost no vacuum) and the EGR valve open at higher loads the engine runs 20-25 degrees spark even with engine speed below 2500rpm. Again we see about a 10 degree bump in spark timing advance when the TGV opens to compensate for the slower burn.

From these maps and logs we have now seen how the TGV works to speed up the combustion in the chamber, especially in light of the Atkinson Cycle and cooled external EGR operation.

[arghx7]

Still reserved for more direct injection discussion (fuel pressure, injection timing, etc)

So below is a BRZ injection timing map because I don't have an FA20DIT one yet. Consider this a placeholder for now. Vertical axis is engine speed, horizontal axis is airflow/rpm calculation of engine load (on a naturally aspirated 2 liter), numbers in map are crank angle degrees BTDC firing for a single injection event in warmed-up engine conditions.


Source: OEM Subaru Map extracted from ECUTek Software

Now keep in mind that the BRZ also has a port injection system, but that's not that important for this discussion. PFI injection timing has its own set of rules which I can get into if anyone is curious, but as far as I know nobody has reverse engineered Subaru PFI injection timing controls in the software. The DI injection timing trends you see below are common to many engines, but the bore washing concern on side injection engines gets replaced with a valve impingement concern on central injection engines.

I am listing general rules of thumb about start of injection timing for solenoid injectors, running a single injection event in a given cycle at a warmed-up condition. It's the most basic type of direct injection control and excludes catalyst lightoff, lean burn, or split injection for knock or particulate reduction. A Mazdaspeed 3 DISI 2.3 engine uses this very basic type of control, whereas a Mercedes AMG M133 2.0 turbo engine uses very elaborate injection controls with multiple small injection sprays within a single cycle. I am listing 3 injection timing rules and 3 rule reversals (for you Robert Green fans) showing the flip side of advancing injection timing.

Rule 1: The earlier the start of injection, the better the volumetric efficiency, as long as you aren't hitting the piston much. The fuel sprays into the air coming through the intake port, cooling it and increasing density.


Source: Davis, “Development of the Combustion System for General Motors’
3.6L DOHC 4V V6 Engine with Direct Injection” 2008, SAE 2008-01-0132

Rule 1 reversal: The later the start of injection, the more the mixture is cooled inside the combustion chamber, reducing knock. Spraying later means less of the volumetric efficiency improvement effect.


Rule 2: The earlier the start of injection, the more likely the fuel will hit the piston (causing smoke and particulates) or hit the cylinder wall, diluting the oil and washing the protective film off the bore.


Source: Davis, “Development of the Combustion System for General Motors’
3.6L DOHC 4V V6 Engine with Direct Injection” 2008, SAE 2008-01-0132


Rule 2 reversal: The later the start of injection, the less likely the fuel will hit the piston or bore wall, reducing the chances of oil dilution and bore washing. In the image below, the X axis shows start of injection in degrees BTDC Firing. A higher number means earlier. To reduce bore washing and piston impingement, the SOI needs to stay away from the steep part of the SOI “cliff” where impingement occurs.


Source: Yi, “Development and Optimization of the Ford 3.5L Ecoboost Combustion System” 2009, SAE 2009-01-1494


Rule 3: The earlier the start of injection, the more time the mixture has to form. This stabilizes combustion (see appendix) at high speed/high load and very low loads, where more time is needed for mixing.

Rule 3 reversal: The later the start of injection, the less time the mixture has to form. This can cause torque fluctuations and misfire, especially at high speed or low/speed low load.


Here is a general guideline: 300 degrees BTDC is a good baseline value for most engines at most speed and load. It's not too early, not too late. As you get to high engine speeds it may be a little more prone to misfire because more time in terms of milliseconds is needed for maxing, so advancing the timing in crank angle degrees is preferable. 300BTDC is also not fully optimized for very low load operation, where combustion speed is slow. It may also produce more impingement at low speed full load than a more retarded value. It's a good starting point though.

Homogeneous Combustion Cold Starts vs High pressure Stratified Combustion Cold Starts

The 2016 models are prone to startability issues when TGV's are deleted. In order to counter act that, stratified combustion mode for warming the catalytic converter has been disabled. Recall that cat heating stratified combustion has some of the following characteristics:

Late injection in the compression stroke, usually in conjunction with an earlier injection(s) in the intake stroke

Very retarded spark to warm up the catalytic converter

A large volume of airflow being drawn into the engine, on account of the retarded spark reducing engine torque

High rail pressure to improve atomization of fuel

The logs overlaid below show what happens when a 2016 with TGV deletes is run in standard stratified mode (dotted lines) vs disabling the high pressure stratified start (solid lines):



Take a look at the AFR and the rail pressure setpoint control. The black dotted line shows a high pressure stratified start, with rail pressure setpoint (dotted red line, a bit hard to see) much higher than in standard homogenuous mode, the solid lines. AFR isn't jumping all over the place, partly because you don't have severe misfires affecting combustion.



Look at the blue dotted and solid lines. The solid blue line is lower, meaning less injection duration, and the black line is lower, indicating the aforementioned lower rail pressure. That means the fuel mass delivered is a lot lower. On a port injected engine, rail pressure varies much less, but on a direct injected engine the rail pressure has a much bigger effect on the mass of fuel delivered. Part of the reason for less fuel delivered is less airflow into the engine.



Notice that, in homogeneous mode, the engine load is lower? You also see that the throttle doesn't have that spiking. The spiking is anti-stall control because of the misfire. The lower overall level of throttle opening is because less airflow is required to maintain the engine speed. Why would less airflow be required to maintain the engine speed in homogeneous mode rather than stratified mode?



The spark timing is much much more advanced in homogeneous mode! The ECU isn't trying to heat up the catalytic converter (be careful trusting that "cat temp" line, as it is a model, and these cars aren't stock). More spark timing means more advanced combustion, meaning the combustion is more efficient. Less airflow is required to make the torque for maintaining the high idle speed.

So for cold starts:

Homogeneous mode - 1 injection in the intake stroke, more advanced spark timing, lower rail pressure, less airflow, more stable combustion

Stratified mode - multiple injections usually, at least one of which is in the compression stroke, very retarded spark timing, high rail pressure, more airflow, less stable combustion that is prone to misfire, especially when the tumble motion in the combustion chamber has been altered.

[arghx7]

Appendix A: Cams and Valve Timing


Source: original image but based on Asmus, "Valve Events and Engine Operation," 1982, SAE 820749

The chart shows on the horizontal axis a full 720 degree crank angle cycle measured in degrees ATDC firing, with top dead center (TDC) firing occurring at 0 & 720 degrees, and TDC intake occurring at 360 degrees. So 180 degrees would be the end of the expansion/beginning of exhaust stroke, and 540 degrees would be the end of the intake stroke/beginning of compression stroke. The vertical axis is valve lift. The red line shows a lift vs crank angle curve for an exhaust cam, and the blue line shows a lift vs crank angle curve for an intake cam.

The centerlines for the exhaust (“ECL”) cam and the intake cam (“ICL”) is the distance in crank angle degrees between TDC and peak valve lift. An intake cam with centerline 110 would have peak lift 110 degrees ATDC intake (or 470 ATDC Firing). An exhaust cam with centerline 110 would have peak lift 110 degrees BTDC intake (or 250 degrees ATDC Firing). Cam Phasing (AVCS) phases the whole lift profile and moves the centerline relative to the default locking position (left or right in the image). It doesn’t change the peak lift, the duration, or the shape of the lift profile. Now here are some camshaft parameters to understand. These aren’t my own personal ideas but rather they have already been written about by acknowledged experts in valve timing.

Effective compression ratio – this is the compression ratio based on the closing timing of the intake valve. The engine may have a nominal compression ratio of say 8.0:1, but the effective compression ratio depends on the intake valve closing timing. The closer the closing timing is to BDC compression (540 degrees ATDC), the higher the effective compression ratio (up to the nominal/geometric value).


Source: original image

For purposes of this discussion, closing timing before 540 degrees (BDC compression) is early intake valve closing. Closing timing after 540 is late intake valve closing. You might hear those terms used in slightly different ways though. When we utilize very late intake valve closing for fuel economy we call it Atkinson cycle.

Effective expansion ratio – This is similar to effective compression ratio, except it relates to the expansion work in the cylinder which actually makes the torque. The earlier the exhaust valve opening timing, the lower the expansion ratio. High expansion ratio could come at the expense of poor blowdown, making it harder to evacuate gases out of the cylinder.

Blowdown volume – relates to the area under the curve before BDC expansion/exhaust. Blowdown helps evacuate gases from the cylinder and relieve backpressure at the turbine inlet. Blowdown is helpful at higher speeds and loads to evacuate gases, but it reduces expansion work.

Exhaust flow volume – relates to the area under the exhaust valve lift curve
Intake flow volume - relates to the area under the exhaust valve lift curve
Overlap volume – relates to the area under the overlap of intake and exhaust valve opening

Intake valve closing volume – relates to the area under the curve after BDC compression. Intake valve closing volume improves volumetric efficiency depending on the rpm.

Notice we use the term "volume" (like intake valve closing volume). We can take an area under the valve lift curve shown in the crank angle vs lift diagram and then do some math (which I won’t get into) according to the geometry of the valves to normalize the parameter. It's a way of comparing cams between different heads. For example, a “big cam” should be considered in light of the number and size of the valves. It’s a way to do the comparison between say aggressive cam with small valves vs mild cam with big valves, or aggressive cam with 2 valves vs mild cam with 4 valves.

Some Cam Phasing Rules of Thumb

As I advance my intake cam centerline, effective compression ratio increases and intake flow volume decreases. Overlap volume may be increasing as well depending on the exhaust valve events. As I retard my intake cam centerline, effective compression ratio decreases and intake flow volume increases. Overlap volume may be decreasing as well depending on the exhaust valve events.

For intake phasing there will always be a tradeoff between overlap, closing volume (volumetric efficiency) and effective compression ratio.

As I retard my exhaust cam centerline, effective expansion ratio increases and blowdown volume decreases. Overlap volume may be increasing as well depending on the intake valve events.

As I advance my exhaust cam centerline, effective expansion ratio decreases and blowdown volume increases. Overlap volume may be decreasing as well depending on the intake valve events.

For exhaust phasing there will always be a tradeoff between overlap, blowdown volume (evacuation of exhaust gases) and effective expansion ratio.


If my overlap volume begins before TDC intake, I get hot exhaust gases spitting back into the intake port because the piston is still rising. If my overlap volume begins after TDC intake, I get exhaust gases drawn back into the cylinder from the exhaust manifold. This behavior depends on the air pressure at the exhaust port. At low loads, I will spit hot exhaust gases back, but under boost, when intake port pressure is higher than exhaust pressure, we can scavenge residual gases out of the cylinder and also help spool the turbo.

Appendix B: Combustion Quality Metrics and Combustion Enhancement

This appendix provides context to discussion of the TGV, EGR, Atkinson Cycle, and head design. First I will show some pressure traces of a non-knocking engine at different types of combustion phasing. Since the TGV is a burn rate/combustion quality enhancement device, the next section will explain combustion speed and stability metrics by drawing an analogy. It will show another burning enhancement approach: dual spark plugs as implemented on a modern Hemi engine. Adding an additional spark plug doesn’t restrict flow like a TGV or a high-tumble intake port, and it can improve combustion. Besides adding cost and complexity though it also takes up space in the head that could be used for valves, an injector, a spark plug, more casting, etc. There’s always a tradeoff.

Now, on to introducing three metrics for combustion: Combustion Phasing (and peak pressure), Combustion speed, and Combustion Stability.

Combustion Phasing

We’ll start with combustion phasing, which could be described as how advanced or retarded the combustion is. You might be used to thinking about advanced or retarded spark timing, but remember that spark timing only tells you when the plug was fired, not how the burning actually behaved.

We can measure combustion phasing with the location of 50% burn (called MFB50 or CA50) and the location of peak pressure (LPP). Both are in units of crank angle degrees ATDC firing. Typically your minimum spark advance for best torque (MBT) sets your CA50 to about 8 degrees, with an LPP of 11-13 degrees. This is a pretty hard and fast rule for spark ignited piston engines with conventional combustion. Take a look at the images below showing pressure traces on a crank angle basis and a Pressure-Volume diagram basis.



Source: Original Image


Source: Original Image

In an area of engine operation with a knock limitation (for that given fuel), we need to retard the combustion, and our combustion pressure looks like the “retarded” image above. Combustion can also be retarded to heat up the exhaust for lighting off the catalytic converter. Otherwise we want our combustion to be like the “MBT” pressure trace. If it’s advanced further than that, we get unnecessarily high combustion pressure and unnecessarily fast burn, which can stress components and cause more combustion noise.

Peak Cylinder Pressures

You can see above that advancing spark timing increases peak cylinder pressure, even if you are not knocking (like with E85). Other things that increase cylinder pressure are:

Preignition events
Engine load - airflow injested at a given rpm, as well as indicated mean effective pressure (area under the Pressure loop)
Higher engine speeds

Peak pressures fluctuate (see section on combustion stability below), and will vary over time. We can take a certain number of cycles, typically 300 cycles, and average them over time. That gives a mean (average) value for peak pressure at a given speed and load. We can also look at the max cylinder pressure cycle among those.

The image below shows the mean peak pressure curve for a Ford Ecoboost 3.5 engine and for a naturally aspirated engine. This is for full load operation for max rated torque curve.


Source: Kapp, "3.5L V6 EcoBoost: Democratization of Sustainable Engine Technology," 2008, Aachen Colloquium

You can see that the mean peak pressure increases with rpm, until it reaches 80 bar. The engine has been rated for 80 bar mean pressures, and the engine calibration keeps to that limit by not having too much spark timing/too advanced of combustion phasing. This engine will be able to withstand occasional cycles greater than 80 bar.

Also, here is peak cylinder pressure map represented as speed & brake mean effective pressure (basically, flywheel torque) on a turbo DI engine (non Subaru) with stock tune. Notice how the max is at high speed & load (peak power), even though peak torque is hit at lower speed:



and here is a chart showing the impact on cylinder pressure of switching from regular fuel (87 octane) to E85 on a naturally aspirated 6 cylinder and advancing the spark. Notice the 20% increase of cylinder pressure at 6000rpm just from advancing spark.





Combustion Speed

Combustion speed is the second metric to discuss, and it is the missing link between spark timing and combustion phasing. I may know when I fired my spark, but I have to know how fast the mixture burned to figure out my combustion phasing and whether I am at MBT or not. When you run an engine at low loads, and then start degrading the burn for fuel efficiency (Atkinson Cycle, external or internal EGR, dethrottling with low valve lifts), combustion slows down. Generally speaking (not always, but generally), fast combustion is good and slow combustion is bad. There are several metrics to measure the combustion speed, measured in crank angle degrees:


Initial combustion formation, also called burn delay or ignition delay: this is the number of crank angle degrees from 0% mass fraction burned to 10%, also called the 0-10% duration.

Bulk burn duration: this is the number of crank angle degrees from 10% mass fraction burned to 90%
You might also see other metrics such as 0-90%, 10-80%, 10-50%, but the above two are the most commonly used.

The below image demonstrates the benefit of a combustion enhancement approach on an engine. In this case it is the dual sparkplugs on the modern Hemi engines. You can see the faster burn with the dual plugs (blue line), the design that went into production, vs an experimental single plug design (red line):


Source: Hartman, “The New DaimlerChrysler Corporation 5.7L Hemi Engine,” 2002, SAE 2002-01-2815

This image shows improved fuel consumption and earlier MBT spark timing for a cruising condition:


Source: Hartman, “The New DaimlerChrysler Corporation 5.7L Hemi Engine,” 2002, SAE 2002-01-2815

High tumble flow such as that from closed TGV has a similar effect of speeding up the burn rates, which is important for offsetting slow burning from internal EGR (overlap through AVCS phasing), external EGR, or Atkinson Cycle. The improved combustion can also reduce knocking tendency, as the gases at the end of the combustion chamber have less time to auto ignite.

Combustion Stability

Combustion stability metrics are based on statistical calculations of the combustion pressure over some number of cycles. Unstable combustion can result in torque fluctuations (surging feeling in the car) or even misfire. Here are some factors that affect combustion stability:

-- Dilution of gases in the chamber (internal or external EGR, lean burn)
-- Spark plug related factors (plug gap, voltage, spark blowout etc)
-- Injection timing and rail pressure, especially on DI engines
-- Combustion phasing, especially spark retard for knock or for gearshift torque reduction (which is one of the reasons why torque curves can look noisy)
-- Motion inside the combustion chamber due to things like intake port and combustion chamber shape, piston bowl or squish clearance, injection spray pattern

I’ll skip the math for calculating combustion stability metrics and give some rules of thumb along with an example of the primary burn rate enhancement for a modern Hemi engine, dual spark plugs. You’ll see in the image below that at an idle condition with dual spark plugs the Hemi engine had more consistent burn with different spark timings. This makes for a smoother idle, better fuel consumption, improved emissions, etc. Here are the two main metrics:

-- Lowest Normalized Value (LNV) – 0-100%, high number is better. The lower the number, the more likely the engine will experience rough running or misfire.

-- Coefficient of Variation (COV) – from less than 1% all the way up to 10+%, and lower is better. The higher the number, the more likely the engine will experience rough running or misfire.


Source: Hartman, “The New DaimlerChrysler Corporation 5.7L Hemi Engine,” 2002, SAE 2002-01-2815

The above images demonstrate ways to evaluate the quality of combustion and they show how a combustion enhancement can improve that quality. Dual spark plugs, a TGV, a restrictive high tumble port, piston crown shape, or valve timing tricks can all enhance burning quality but always with some kind of tradeoff or downside, and usually more than just cost.

[anjuna]

excited to read all this

Sent from my HTC One_M8 using NASIOC mobile app

[wrxdit]

This is awesome!!! Thanks for very educating write up.

[turboSz]

Very nice. Always good to have a solid technical explanation

[15DGMiL]

Tldr; JK. Subscribed for education.

[nmlittlebigman]

Enrolled for this class (Subscribed).
Something tells I don't have the prerequisites

[stealth06wrx]

Awesome! I can't wait to learn more about the new motor! Great idea for a thread!

[gerbs]

This is great. subscribed

[Subperior]

This has sticky potential!

[arghx7]

Added AVCS pumping work reduction section - early intake phasing only vs late/late Atkinson Cycle

[arghx7]

Added section on AVCS scavenging behavior at WOT on stock tune.

[arghx7]

Added section on operation of cooled EGR system.

[sti nova]

This is great information! Keep it coming.

[Legba]

good stuff.

[arghx7]

Added section on TGV operation and its relationship to spark timing on the stock tune.

[dark hawk]

beautiful, thank you sir

[SlimJim8804]

Awesome stuff, thanks for sharing.

Sent from my SM-N900T using NASIOC mobile app

[maoyokhan]

Good stuff to learn from. Thanks for sharing

[DutchOven]

thanks for posting. this is extremely interesting.