Good evening!
I’ve started a little research project. I want to make my SLK faster, just like how I make every car I own a bit faster
One important part of this, is ECU remapping. Tossing on a pulley to spin the supercharger faster and installing a more beefy intercooler is one thing, but you’ll be lacking fueling especially at the top of the rev range.
Unfortunately for me, there are no tuners in my area that want to work on a car this old. Also, any future modifications I make will mean going back to the tuner and paying a nice bill.
This is geared specifically towards the facelift SLK230. On the pre-facelift it was possible to replace the fuel pressure regulator on the fuel rail to help out with fueling. The post-facelift has the fuel regulator built into the fuel pump assembly. It’s still possible to modify this, but it’s a bit harder.
Having experience with ECU programming myself, I decided to look at things I can do myself. Now this will be a very slowly updating topic, but I thought I’d let you ride with me and see where things go. I’d love to mess with this car a bit more and see how far I can get.
I have previously built an Arduino based throttle body injection ECU for my old-timer Skoda. It worked using a DIY crankshaft position sensor and trigger wheel and a throttle body from a 1998 Opel Astra 1.6.
I used 5v sensors so I could easily read them using my Arduino, used a H-bridge to control the idle air motor in the throttle body and used a MOSFET to trigger the fuel injector.
It then ran using a pre-programmed map, no o2 sensor. Basically an electronic carburettor. I implemented some very basic idle learning, but the rest of the map was trail and error from modifying an existing map for a similar engine.
Now back to the Mercedes. Unfortunately getting the required stuff to remap the stock ECU myself costs a real pretty penny. I’m looking at piggyback solutions. Either through a piggyback ECU, or an Arduino based system I can develop myself.
These are the goals I’m looking to achieve:
There are different ways I can achieve this, but there are also hurdles.
The Mercedes ECU is pretty picky and pretty smart. If it detects anything is off, it will immediately throw a code and possibly throw the car into limp mode.
Let’s first look at fueling.
A common way to try to get a car to inject more fuel, is by modifying the MAF signal to the ECU. This won’t work on the Mercedes. When installing a pulley, the car is already aware of the extra air that is flowing, but due to it reverting back to a simpler base map under WOT and full boost it’s not really taking these values into account that much.
A second issue is boost cut. Above a certain voltage from the MAF sensor, the ECU will go into limp mode as it considers it an overboost situation. Therefor modifying MAF signal going into the ECU is not really viable.
Another method, and the method I want to try is modifying the injector duty cycle. When the car is going WOT and above a certain RPM, I’d like to capture the injector pulse signals coming from the ECU and post-process them. Let’s say the ECU is saying the injectors should open for 5ms every x amount of milliseconds. I want to read this signal, and apply a correction factor to it. So let’s say, I’ll add 0.5ms to it.
The actual signal going to the injectors will be 5.5ms. The ECU will be none the wiser any of this is happening. The only important bit here is only apply the correction factor during WOT and full boost, but keep the Mercedes ECU in full control during any other part of the drive.
This should hopefully keep it from registering it and adjusting the fuel trims if my theory that the car isn’t looking much at the O2 sensors during WOT is correct.
How am I going to do all of this? I’m not sure. I was thinking of using an Arduino to sniff the signal and either have it pass it through unmodified (during cruising) or pass it through with the correction factor during WOT.
Another option is to use an off-the-shelf piggyback ECU that is capable of driving the injectors. Either solution will enable me to build a new map that gets applied on top of the Mercedes stock ECU map.
One thing I don’t know yet, is whether the Mercedes is using batch fire or sequential injection when at WOT. (whether it fires all injectors at once, or controls them individually).
Many cars revert to batch fire under high load. Does anyone maybe know how the Mercedes ECU does that? If not, I’ll hook up a scope to the injectors some day.
As for timing, I have not looked into how I might achieve that.
That’s it for now, there are no doubt people smarter than me on this forum so drop your thoughts if you have any!
If in the end I can make this work, I’d like to share all the information about it so anyone can replicate it themselves.
I’ve started a little research project. I want to make my SLK faster, just like how I make every car I own a bit faster
One important part of this, is ECU remapping. Tossing on a pulley to spin the supercharger faster and installing a more beefy intercooler is one thing, but you’ll be lacking fueling especially at the top of the rev range.
Unfortunately for me, there are no tuners in my area that want to work on a car this old. Also, any future modifications I make will mean going back to the tuner and paying a nice bill.
This is geared specifically towards the facelift SLK230. On the pre-facelift it was possible to replace the fuel pressure regulator on the fuel rail to help out with fueling. The post-facelift has the fuel regulator built into the fuel pump assembly. It’s still possible to modify this, but it’s a bit harder.
Having experience with ECU programming myself, I decided to look at things I can do myself. Now this will be a very slowly updating topic, but I thought I’d let you ride with me and see where things go. I’d love to mess with this car a bit more and see how far I can get.
I have previously built an Arduino based throttle body injection ECU for my old-timer Skoda. It worked using a DIY crankshaft position sensor and trigger wheel and a throttle body from a 1998 Opel Astra 1.6.
I used 5v sensors so I could easily read them using my Arduino, used a H-bridge to control the idle air motor in the throttle body and used a MOSFET to trigger the fuel injector.
It then ran using a pre-programmed map, no o2 sensor. Basically an electronic carburettor. I implemented some very basic idle learning, but the rest of the map was trail and error from modifying an existing map for a similar engine.
Now back to the Mercedes. Unfortunately getting the required stuff to remap the stock ECU myself costs a real pretty penny. I’m looking at piggyback solutions. Either through a piggyback ECU, or an Arduino based system I can develop myself.
These are the goals I’m looking to achieve:
- Alter injector duty cycle when under boost to fix the leaning out that occurs when installing a bigger pulley
- Keep the Mercedes ECU in full control during cruising, as the engine is running mostly atmospheric or under very light boost during these scenarios.
- Alter ignition timing for a 98 octane tune.
- Possibly, obtain control of the recirculation flap to stop the ECU from pulling boost.
There are different ways I can achieve this, but there are also hurdles.
The Mercedes ECU is pretty picky and pretty smart. If it detects anything is off, it will immediately throw a code and possibly throw the car into limp mode.
Let’s first look at fueling.
A common way to try to get a car to inject more fuel, is by modifying the MAF signal to the ECU. This won’t work on the Mercedes. When installing a pulley, the car is already aware of the extra air that is flowing, but due to it reverting back to a simpler base map under WOT and full boost it’s not really taking these values into account that much.
A second issue is boost cut. Above a certain voltage from the MAF sensor, the ECU will go into limp mode as it considers it an overboost situation. Therefor modifying MAF signal going into the ECU is not really viable.
Another method, and the method I want to try is modifying the injector duty cycle. When the car is going WOT and above a certain RPM, I’d like to capture the injector pulse signals coming from the ECU and post-process them. Let’s say the ECU is saying the injectors should open for 5ms every x amount of milliseconds. I want to read this signal, and apply a correction factor to it. So let’s say, I’ll add 0.5ms to it.
The actual signal going to the injectors will be 5.5ms. The ECU will be none the wiser any of this is happening. The only important bit here is only apply the correction factor during WOT and full boost, but keep the Mercedes ECU in full control during any other part of the drive.
This should hopefully keep it from registering it and adjusting the fuel trims if my theory that the car isn’t looking much at the O2 sensors during WOT is correct.
How am I going to do all of this? I’m not sure. I was thinking of using an Arduino to sniff the signal and either have it pass it through unmodified (during cruising) or pass it through with the correction factor during WOT.
Another option is to use an off-the-shelf piggyback ECU that is capable of driving the injectors. Either solution will enable me to build a new map that gets applied on top of the Mercedes stock ECU map.
One thing I don’t know yet, is whether the Mercedes is using batch fire or sequential injection when at WOT. (whether it fires all injectors at once, or controls them individually).
Many cars revert to batch fire under high load. Does anyone maybe know how the Mercedes ECU does that? If not, I’ll hook up a scope to the injectors some day.
As for timing, I have not looked into how I might achieve that.
That’s it for now, there are no doubt people smarter than me on this forum so drop your thoughts if you have any!
If in the end I can make this work, I’d like to share all the information about it so anyone can replicate it themselves.
