Tuning the Mazda Skyactiv-D 2.2l engine to GTD level of performance

Hello everybody, I want to share with you my humble personal project, tuning a Mazda 3 (3rd gen, 2015) Skyactiv-D 150 hp to Golf GTD performance level (200-210 hp).

Let's start from some data about the engine stock performance:

- 2.2l inline 4 diesel engine producing 150 hp at 4500 rpms and 380 Nm of torque at 1800 rpms;

- low compression ratio diesel (14:1)

- two-stages turbo gradually switching between 2600 and 3000 rpms (top boost 1.65 bar);

- fuel rail operating at maximum pressure of 197 MPa with piezo injectors;

- DENSO ECU reflashed using VersaTuner.

We are talking about a C-segment car weighing 1395 kilos (curb weight), with great driving dynamics. The engine is relatively large for such car and it's producing only a decent level of power. It's pretty obvious Mazda left some margin here. Also, engine and transmission are part per part identical to the Skyactiv-D 175 (175 hp - 420 Nm) sold on the Mazda 6 and CX-5. A 175 hp Mazda 3 has been sold in Japan and Australia and it's exactly the same car as mine, just flashed with a slightly different tune.

VersaTuner gives the possibility to unlock the 175 hp basically flashing the original Mazda tune, which has only a bunch of modified tables compared to the 150 hp (unsurprisingly, fuel injection limit and boost limit tables). I did it for some time and the results were very nice. But I wanted to push the boundaries a bit further, given that the engine is basically producing stock power. So I booked the wonderful diesel tuning course here and learnt a lot thanks to Nick (man, you are an amazing explainer, I loved the course so much) and upgraded my VersaTuner so I could create my own tune.

Why going on my own instead of a professional tuner? Mostly to be able to control what was done. And there are much less tuning options for the Skyactiv-D platform compared to, for example, TDI/GTD. Also, the engine is known to have some critical weak spots, potentially leading to catastrophic failure, that I didn't want to push.

And it was also a lot of fun :)

My goals were:

- keep maximum torque the same as the 175 hp Mazda tune (420 Nm), to not overstress the internals and the infamous injector seals;

- keep driveability at low RPMs the same as stock, to not impact fuel economy and smoothness of the car;

- reach 200-210 hp range with smooth transition and extracting such extra power from higher ranges.. basically getting the power only when I want power;

- control exhaust gas temperatures (pre-turbo) and avoid turbo oscillations;

- keep the tune as stock as possible, meaning not re-writing entire tables for the sake of doing it.. Mazda already mapped required injection quantities, timings, pulse width, boost and more for way more torque/power than the engine delivers. So I wanted to use as much as possible their tables. Also, I'm an amateur tuner with no experience, so the less I change, the better.

To do so, I basically only changed injection limits above 2000 rpms and accelerator mapping on the higher end, as Mazda already mapped everything already. The increases are gradual, very small at the beginning and proportionally bigger as the rpms go up, to create a gentle slope across all the range.

Here's a comparison between stock fuel scaling vs mine (I'm a data analyst so I created some fancy plots):

I made the fuel rail pressure go up steeper, to reach maximum value of 197 MPa around 700 rpms before the stock tune (around 3200 instead of around 3900). The reason is to speed-up fuel injection which has a positive impact (albeit small in this case) over EGTs. I was also planning to slightly advance the injection timings for the same reason, but after calculating them I noticed they are already quite advanced on the top-end range and any touch up would have been very tiny, so no need to bother.

Also the boost has been adjusted only for the second hi-end turbo, in the 3000-4000 range. The increase in boost pressure is quite small. The reason? In the 2000-3000 range there's not much changing in my tune, so we can work with stock boost. Then, lambda above 4000 rpms is not really a problem (I get 4-5% O2 concentration on hi-end WOT operations) and requesting more boost on that range generates huge oscillations and crazy wastegate behaviour.

I tried one tune version pushing up +5% the boost from 2000 rpms on, I actually got less boost than stock :(

But adjusting only the second stage turbo I was able to actually pull a bit more boost in the 3000-4000 range. You can see below Engine rpms, MAP and Wastegate Solenoid DC of stock tune and my tune:

The feeling for now is very good and the data logging showed no problems (I collected a long autobahn session and a short city one with quick bursts of acceleration). I didn't have the car dynoed yet but I will go in the next months. Just as reference, in the 4500-4600 range the amount of fuel getting into the cylinder is around 19-20% more than stock, so I expect the power to be between 200 and 210 hp (flywheel power).

On the autobahn, EGT limits (790 degrees) are reached after around 30-40 seconds of full throttle operations, which resulted in that specific session in 220 km/h speed on my Mazda 3. The ECU successfully reduces the injected quantities and stabilizes the temperature but it’s better to not spend too much time in that range.

Temperature here ranges from 2 to 7 Celsius, so it's quite fresh. I will check how the tune works in summer, for sure EGT limits will be reached much sooner.

For keeping everything under check I created my own dashboard with OBD Fusion and a BT OBD2 adapter. I can only recommend that if you tune your car. And if you spend some time on it the results are pretty cool:

It has been a nice journey so far!