Subaru WRX FA20 DIT Tuning – Initial Findings

We have been involved with tuning direct injection turbocharged performance engines for over a decade – so naturally it was time apply our knowledge to one of the most iconic sport compact vehicles on the road today: The Subaru WRX.

subaru-wrx-2018-stratified-2

Initial impressions with the car are that unlike some of its competitors it feels very much like a white canvas. There are no sound symposers, pops, burps or fancy torque vectoring differentials. The full time AWD system adds great traction and the low center of gravity of the drivetrain along with suspension tuning make the car feel very stable. A good sign when you plan on adding power.

subaru-wrx-stratified

Out of the box, it is not impressively fast or punchy. This white canvas feel to the car (a bit of a throwback) is excellent in the sense that it gives the enthusiast a lot to build on an individualize. Our plan is to do just that – build on the character of the car and add enjoyment to the driving experience.

Our first modification (after 160 miles of ownership) was a COBB Accessport. This is the tool of choice for the other platforms we calibrate and it is very well featured on the Subaru platform.

subaru-wrx-cobb-accessport-tuned

We dug right in and built our calibration on the otherwise bone stock 2018 WRX. Below you will see a power comparison between Stock, COBB Stg1 91, and the Stratified tune on the same 91 octane fuel.

subaru-wrx-dyno-stratified-vs-cobb-vs-stock

The interesting thing is that stock (as well as with the COBB tune) the car ran so rich that it actually hesitated under wide open throttle operation. The OEM MAF calibration was around 10% rich using our ethanol free fuel. This is something to keep in mind if your fuel has low ethanol content and if you feel hesitations during acceleration – even stock. Leaning the curve slightly as well as adjusting the MAF and utilizing the wideband O2 under WOT are effective ways we used to address these issues.

You can see that the COBB OTS tune is quite conservative picking up power in the midrange with mild boost increases over stock.

Aside from running a leaner (yet still very much safe) air-fuel ratio which improved driveability and response, we found that even on 91 fuel we were able to add repeatable performance and improve response over the COBB OTS Tune. We did this by increasing boost, adjusting timing, refining boost control, and tuning variable valve timing.

Fuel quality plays a large role in limiting performance beyond this point. We will be working with ethanol shortly as well as adding parts to the vehicle. Stay tuned as we continue the journey with the Subaru DIT and provide custom tuning support for the FA20DIT engine.

The Stratified Team

Focus RS 9000 Mile Checkup

With our RS coming up on 10000 miles and a year of development and (hard) driving, we thought it would be a good idea to borescope the cylinders and see how things look. This is a 2016 September build and the car has definitely been put through its paces on both the track and street.

The good news is that things look great! Spark plugs show the correct heat range (we run a step colder) and even burning across all the plugs.

Digging deeper into each cylinder we see that the pistons have the usual carbon buildup. We also see how effective the bowl design is at directing fuel and preventing it from being projected on the cylinder wall. You can see in the image below that the fuel spray is well contained within the center bowl (and this cleans it of carbon buildup). Keeping fuel off the cylinder walls prevents them from being washed down and oil dilution.

focus-rs-piston-top

Below is a video of one of the bores. The cylinder head gasket is the line in the center of the image. We checked all cylinders and none showed any signs of deterioration at this level. The black staining you see on the piston deck is from carbon deposits and the piston rings not coming up high enough to scrape it off the cylinder wall. This is normal.

Here’s to 10000 more! We will keep you all posted. CheeRS!

Focus RS Head Gasket Failure Mechanism

A lot has been made of the Focus RS Ecoboost 2.3 head gasket failures. It seems to be an issue that has occurred in both modified and unmodified vehicles. I am not certain what percentage of vehicles are affected. However we have spent significant time looking at the mechanics of the failure and wanted to share this with our customers and community.

The Focus RS uses a multi layer steel (MLS) head gasket. This is a very capable method for sealing the block to head interface. This multilayer gasket is enhanced by an elastomer that further helps seal critical areas.

mls-head-gasket

The Focus RS 2.3 Ecoboost engine block is an open deck design. This means that the cylinders are not attached to the outside of the block structure at the sealing surface of the head gasket. This space is filled with engine coolant.

The head and block are secured to each other using torque to yield bolts (which do not require re-torquing the head fasteners after the engine is heat cycled). In spite of this there is inevitable movement between the two sealing surfaces. The two types of movement are vertical movement (which is what causes head lift and a sudden gasket failure) as well as lateral movement. This lateral movement is of most interest in this situation. Minute lateral movement causes what is called gasket fretting/scrubbing. There are two mechanisms that cause this lateral movement:

  1. Thermal deformation. As the engine is brought through a range of temperatures (within normal operation), the expansion and contraction of the head and block cause this lateral movement between the two interfaces. The movement caused by this is however low cycle, as the engine is warmed up and cooled down relatively few times compared to the next mechanism.
  2. The most relevant mechanism for lateral movement is due to cylinder firing. When a cylinder fires, it applies pressure on the open deck cylinder structure. This in turn causes this minute movement between the two interfaces scrubbing them at a high rate.

The result of this scrubbing is material loss from the head gasket, block, or head. Over time, the seal between the block and head can fail. The area where it will fail first is the area with: the highest movement, the highest contact pressure, and the least sealing material available.

The area of highest lateral movement for the Focus RS is between the cylinders.

focus-rs-high-head-to-block-movement

The area of highest contact pressure is around the cylinders (because this is where the combustion pressure is highest and needs the best seal). This is by design.

focus-rs-contact-pressure

The area that will inevitably experience the highest wear rate will be between the cylinders. Manufacturing discrepancies, vehicle use, temperature fluctuations and other factors will influence whether a failure will occur on a vehicle and if so when.

focus-rs-high-wear-rate-head-gasket

There are two methods to reduce this scrubbing wear. One is to stiffen the block structure. When rebuilding an engine for example, using the closed deck 2.0 Ecoboost block is an attractive option. However, Ford could not use this block in production and I am certain there are good reasons for this; likely related to emissions certification.

The second method, and what is being implemented as a fix is to redesign the head gasket.

Below is an image of a failed Focus RS gasket. A few things to note:

  • There is a very high wear rate of the elastomer (black section) between the cylinders. This is where the gasket has failed. You can see that it is almost completely missing in these areas.
  • Notice how thin the sealing area is between the cylinder overall.
  • Looking carefully you will notice that the are two holes in gasket between the cylinders marked on the image below.

focus-rs-failed-head-gasket

Now let’s have a look at the block. You can see there is a very smallcoolant hole on the exhaust side. This hole lines up with one of the holes in the failed head gasket. You can also see a trail of coolant marking on the block surface leading to the second hole. This shows there is no elastomer between the cylinders on the original head gasket.

focus-rs-block

Looking at the head, you can see that the first hole on the head gasket lines up with a coolant passage. The second hole, however does not line up with any coolant passage. Once again, the head shows that coolant was found in the area between the cylinders but this was a closed path. Overall the cylinder sealing area was very narrow and the head gasket failed after a number of miles.

focus-rs-cylinder-head

When the car was brought to the dealer, it was repaired with the head gasket shown below. There are some very important changes in the area between the two cylinders on the updated gasket.

  • The second deadheaded hole is no longer present
  • The entire area between the cylinders is now coated in elastomer. This will increase the sealing surface area and better distribute the clamping force in this critical area. This will in turn reduce the fretting/scrubbing failure of the head gasket.

focus-rs-head-gasket

The latest gasket part number is G1FZ-6051-C and this is what dealers are replacing failed gaskets with on vehicles that experience this failure. I am not sure at what point the factory engines have received this latest gasket part number. The only indication I was given was that this part number was available to service centers sometime between 3/27/17 and 8/10/17. Again, this says nothing regarding when this change was made to vehicles coming out of the factory.

I hope this clears up some of the confusion around these head gasket failures and what is causing them. The symptoms of failure are:

  • Loss of coolant.
  • Misfires, especially on a cold start and rough running.
  • Coolant fouled/wet spark plugs.

If you experience these symptoms, take the car in to Ford for the fix. Otherwise, enjoy the RS! It is not the only car in history to experience difficulties with head to block sealing. High output, low emissions, small displacement engines like the RS are a challenge to engineer and can have some teething issues; but at the same time we can all agree that fun behind the wheel was certainly well engineered!

The Stratified Team

Photo credit:

Brewer, T. and Chen, X., “Cylinder Head Gasket Fretting/Scrub Mechanism Investigation and Analysis Procedure
Developments,” SAE Technical Paper 2017-01-1091, 2017, doi:10.4271/2017-01-1091.

How the Knock Detection System Works in the Ecoboost

The Ford Ecoboost knock detection system is one of the more complex systems out there. A lot of the vehicles using the Ecoboost engine are rated to use fuel that ranges in octane from 87 to 93 octane. Because of this, Ford built a system that is very powerful and adaptive. It actually learns your fuel quality while you drive. All the EB engines we work with use this adaptive timing system. This is great for us calibrators as we can build failsafes and adaptability into the tunes – but it has to be setup correctly.

The knock detection system starts with the sensors (microphones) that pick up engine noise and are mounted to the engine block. If the engine noise falls within the correct frequency and window of operation (around the time the spark plug fires), the ECU registers this as knock. When knock is registered, the ECU looks at the Knock Response table pictured below. Based on the intensity of the knock it subtracts timing advance from the Ign. Corr. Cyl X Parameter (which you can view and datalog) on a per cylinder basis. On each knock events it also increments the Knock Count Cyl X parameter which can also be viewed. Both of these parameters are reset to 0 once you let off the throttle, shift gears, or experience a throttle closure.

Keep in mind that the following ECU tables shown have OEM values. We adjust these to better suit our performance criteria, but the principles remain the same.

ignition-reduction-table

The amount of timing being pulled for each knock intensity and engine speed is completely under the control of the tuner who builds the tune. You want to make sure that you pull enough timing per knock event to maintain engine safety and stop the knocking. Here’s how this looks in a graphed datalog of a single cylinder during a pull.

ecoboost-knock-count-and-ignition-correction

These negative ignition values are always subtracted from the current Ign Corr. Cyl X variable for each cylinder. This is important to understand because the Ign Corr. Cyl X variable also ADDS timing when the system does not “hear” any knock and you are on the throttle.

The rate at which timing is added and how much total timing is being added is also adjustable in the tune. Ideally the tune should walk timing up at a reasonable rate that is not so fast such that knock is induced and not too slow such that performance is left on the table.

One thing that has to be understood is that all engines will knock at some point regardless of tune and fuel used. The combustion process is a chaotic process and sometimes you get some irregularities. This is inevitable. You want to tune out heavy and sustained knocking, but completely eliminating all knock is not something that is realistic – especially in a street engine. Handling knock correctly is the key. Light knock will NOT damage an engine especially if timing is reduced via a reduction in Ignition Corrections when it happens. LSPI (Super Knock) and bouncing off a Rev Limiter have a MUCH more damaging effect on the pistons and rods.

The beauty behind this knock control system that both adds and subtracts timing is that it also LEARNS the quality of your fuel over time. The learning variable is called OAR (Octane Adjust Ratio). This variable is like a fuel trim but used for ignition and load/boost targets. It can range from -1 to 1. Below is the table that allows the ECU to learn and adjust OAR. You can see from this table that in order for this system to work, ignition timing and the Knock Response table must be set such that the ECU can ADD and SUBTRACT enough timing to make the Ign. Corr. Cyl X parameters positive as well as negative during a pull. With an adaptable tune you are likely to see Ign. Correction numbers that are both positive and negative at times.

how-oar-learning-works-ecoboost

Finally, this OAR number is not just another pretty face. It affects the timing AND load/boost tables the car will be targeting. An OAR that is closer to 1 will remove both timing and boost reducing future knocking. An OAR closer to -1 will add timing and boost due to good quality fuel and increase performance. It is normal for OAR to fluctuate slightly depending on your fuel quality and driving conditions (5th gear pulls are more likely to knock compared to 3rd gear pulls and OAR will adjust to reflect this). Below is the timing table that is multiplied by OAR in the OEM calibration and added to the overall ignition timing the ECU runs. In this case an OAR of -1 will add 4* timing and an OAR of 1 will remove 4* of timing near redline. OAR dependent boost/load limits are setup via a blend of several tables not shown here.

oar-multiplication-table-ford-ecoboost

It should be clear now that NEVER seeing any negative corrections is an indication of a tune not setup to use this learning system to its full potential. This is not ideal on a street driven car as it means that it will not adapt to keep the engine safe when fuel quality or driving conditions change nor will it extract more performance when the situation allows. 

Keep in mind that these knock detection systems are extremely fast to react. If knock is detected, if setup correctly, the system will very quickly pull timing to stop the knocking during that pull and adjust the tune for future pulls. Excessive repeatable knock or only positive ignition corrections are something that should be addressed in the tune – and can only be done by the person that has set the tune up. More on knock in this article. 

Now let’s have a look at a set of logs see this in action across four cylinders.

Here’s a normal 3rd gear pull with all 4 cylinder Ign Corrections and Knock Counts graphed. As you can see, ignition timing walks up when there is no knock in a step-wise fashion. When there is knock it drops down stopping further knock. If the knock events happened earlier in the pull or were of a higher intensity, Ign Correction could have become negative.

ecoboost-knock-count-and-ignition-correction-2

Here is how we can setup the same knock detection system in a less ideal manner. At first glance, it looks like this engine is basically not knocking at all. However if you look carefully, you can see that there were 4 knock events. The amount of timing pulled for each event is very little. This does two things. One, not enough timing is being pulled to stop the occurrence of further knock. Secondly, pulling so little timing will mean Ign Correction values will not be negative even though the engine could be knocking heavily and continuously. This in turn will mean the OAR will not adjust and the tune will not adapt to the fuel quality used.

ecoboost-poor-knock-handling

For this reason, watching Min/Max Ign Correction on the AP can be very misleading. Not only does it not show you how many times the engine knocked, it also doesn’t show you how intense the knock was. To determine the performance of the tune, a full datalog is always necessary or at least a good understanding of HOW the tune handles knock events. 

Happy Tuning,

The Stratified Team

Misfires and the Ecoboost, when should you worry?

I’ve seen a number of recent inquiries regarding the misfire count on Ford Ecoboost cars and I wanted to explain a little bit more about what it means, how the ECU determines when the car misfires, and when you should worry about it.

What are Misfire Counts?

First of all, the Misfire Count as a live parameter or datalogged represents the number of misfires detected by the ECU since the last engine start. ecoboost-misfire-count

In the case shown above, the COBB Accessport monitor shows that there have been a total of 8 misfires since the engine has been running WITH THE COBB AP turned on.

 What is a Misfire?

Combustion is a complex process. The power generated by an engine comes from the “power” stroke of the 4 stroke cycle. The power stroke is where the spark plug lights the air-fuel mixture inside the cylinder and you get a propagating flame front that evenly pushes the piston down. This is of course the ideal scenario. However sometimes, for a variety of reasons discussed below, the spark plug does not effectively ignite the mixture inside the cylinder. This causes what is known as a misfire. For that one cylinder, power is either low or not made at all during that one cycle and you will feel it as a short hiccup in what is otherwise smooth engine running and power delivery.

How is a Misfire Detected by the ECU?

Modern cars have a high resolution crank position sensor and multi toothed wheels. This is attached to the crank where the accessory pulley is and looks like this.

ecoboost-crank-wheel

The ECU knows when the spark plug fires for each cylinder with respect to the position of the crank wheel and expects the engine to be accelerated during the power stroke. If the crank teeth are not accelerated during the power stroke (you can also feel this sitting in the car) it will determine that the particular cylinder has misfired.

Why does a Misfire Happen?

The engine needs a few elements for good combustion. These are air, the correct amount of fuel, effective mixing between the air and the fuel in the combustion chamber, a strong enough spark, and good compression.

During NORMAL and HEALTHY engine operation misfires do happen but they are random and intermittent. For example, the Ecoboost engine tends to misfires at idle. This is because sometimes the air and fuel do not effectively mix and light off. This is a common issue with direct injection engines where the fuel and air are injected directly in the cylinder with poorer mixing characteristics. Engine designers and engineers get around this in multiple ways – sometimes with tumble flaps that increase air velocity (which the Ford does not have) – and through extensive internal combustion chamber and injection designs. These random idle misfires are undesirable for smooth running and emissions; however they do still happen even in healthy engines. The Misfire Count will pick these up and increment but a Check Engine Light (CEL) will not necessarily be triggered.

Another common reason for misfire is spark plug gap. This happens at high load under wide open throttle operation. A gap that is too large may prevent the plug from lighting when there is a dense mixture in the chamber under boost. We recommend a 0.026 inch spark plug gap to avoid these. You can feel these as small intermittent losses of power at wide open throttle but they are a simple fix.Spark Plug Gap

Other failures of the ignition, injection, or air induction system can cause misfires (including carbon buildup) but the above two causes are the most common in healthy engines.

If there is internal damage to the engine such as a damaged ringland and associated loss of compression this can trigger a continuous misfire for that cylinder. The misfire count will increment much faster and eventually it will trigger a check engine light. This constant misfire is something you will FEEL and HEAR with the car at idle and it will be outside of the norm. A quick compression test will verify the internal health of the engine so there is no guesswork in this diagnosis.

Ecoboost Compression Test

When are the Misfires Codes (P0300, P030X for each cylinder) Triggered

The OEM ECU has a complex mechanism for determining when a misfire is part of normal operation, or when it means there is something that needs to be addressed. When it determines that something needs to be addressed; it will illuminate the Check Engine Light (CEL) and usually indicate which cylinder the issue occurred in (P0301 indicates cylinder 1, P0302 indicates cylinder 2, and so on).

There are two programs built into the ECU for determining two different types of failures. We will call these Type A and Type B.

Type A detects a failure that has recently happened or is sudden in nature (failed coil pack for example). This detects failures over a period of 200 revolutions. If the engine misfires in excess of 23% of the time (46 misfires per 200 engine revolutions) at low RPM or in excess of 5% at high RPM it will trigger the fault code under this condition.

Type B detects a condition that is progressing over time. This has a detection period of 1000 engine revolutions. For example in the Focus RS it triggers if 1.2% of revolutions  result in misfires (12 misfires per 1000 engine revolutions). This is designed to catch progressing conditions such as worn spark plugs.

For the sake of simplicity, assume that your car is idling at 1000RPM. This means that for you to have enough misfires to indicate an out of tolerance condition, you will need to have at least 12 misfires during 1 minute of running at idle.

When should you worry and what should you do?

check-engine-light-ecoboost

Because the ECU is keeping track of the misfires, you don’t need to! When a fault code is triggered for a misfire, this is when you want to investigate and look for an issue. Otherwise, keep spark plugs fresh and gapped correctly and keep in mind that misfires can and do happen on perfectly healthy engines!

The Stratified Team