OSHW Turbo Diesel Tuning Box

[Wilksey]

Am I the only one wondering for a self confessed OSHW project, where the project files are? 

[JPortici]

If you do something to your car in Germany with stuff that has no universal operating permit (Allgemeine Betriebserlaubnis), you have to let the TÜV (an agency that checks if a car meets the requirements to be road-legal) check your modification.
If the engineers there say everything is okay, it gets entered into papers of your car and is therefore road-legal.
At least that's how several people in school had to do to make their tuned up scooters legal

same here.
for those who wonder about an OSHW project.. most if not all the hardware is trivial, all the magic lies in the software

[brainwash]

I've already started working on my own version - at a slow pace - documenting the steps on my blog.
The hardware is trivial if only you've done things like this before. Even if it's not approved it should still have the possibility to be so by ensuring adequate protection:

Here's what could happen on the power supply: spike of 30, 150, 20kV, with various durations, reverse polarity, 24V for a minute. All modules should accept shorting data lines to VCC and GND without failure. Are the parts being used documented for this? I haven't seen anything related to this in the ESP8266 datasheet for example, what the ESD protection diodes are able to withstand.

The original sensor is specced for 12-20mA current draw, 2-5mA current output, is the module able to work within those limits? How much load does the ECU expect on the VCC (5V) line? What's its input impedance?
You could get by using some existing 12V from somewhere else but then everything changes.

What connectors are in use? There's Bosch Kompact, there are some from TE (available only in qty >4k or 'sourced' from ali*), but there might be quite a few others. You could splice the wires, but what then?

The sensor has 2 voltage options: 5V and 12V - with different requirements. The 5V one has error signaling at 0-0.5V and 4.8-5V. Is the module able to forward these voltages back to the ECU? Rather, is the opamp you are using for buffering able to swing that much?

What's the response time of the original sensor? Is the ECU using that information?
What's the voltage offset? How much is the ECU able to compensate for? (that's why older Bosch ECUs work with a resistor and new ones don't)

Failure modes?

Temperature range, water ingress, vibration tolerance...

Doable? Yes. Trivial? Up to you to decide.



Anyway, my plan is to feed the data back via wireless (something low-power) to get telemetry data: idle, snap acceleration, city and highway rpm/pressure range. Since it's a two-way communication you should be able to send a response curve back to the device that applies an offset based on the input. A curve with 5-10 data points should be enough. It would output the unmodified signal if no commands are given or the MCU locks up.

[f4eru]

I find it strange that some people insist in detuning their engine...
Too much money I suppose.

Now a really interesting project would be recreate an ECU from scratch.
But knowing the complexity involved, that won't happen soon.

[janoc]

I find it strange that some people insist in detuning their engine...
Too much money I suppose.

Now a really interesting project would be recreate an ECU from scratch.
But knowing the complexity involved, that won't happen soon.

That did happen already:
http://rusefi.com/wiki/index.php?title=Main_Page
or this one:
https://en.wikipedia.org/wiki/MegaSquirt

[brainwash]

Here's my progress so far: http://hackcorrelation.blogspot.de/2016/05/designing-better-diesel-tuning-box-part_23.html

TL;DR version: ">30% fuel savings while idle." "25% highway mileage increase".

This was a version done as a proof of concept, I still have to wait and see if the ECU will adapt itself to the new values and revert to the old consumption, but so far it was on the car for more than 600km.

[JPortici]

Doable? Yes. Trivial? Up to you to decide.

there is one line of high class tuning boxes that is two PIC18 and two can transceivers. period. nothing else on the board (well, a bit of supply filtering and protection but that's it)
That one can be used with a limited range of cars (also because you need to decode/encode the messages from/to the ECU and for that you have to know the right people) but that's an example on how simple it can become on the hardware side
others that i've opened never had impressive hardware or protection anyway

[russian2]

That did happen already:
http://rusefi.com/wiki/index.php?title=Main_Page

And this is OSHW by the way. See also https://www.eevblog.com/forum/projects/rusefi-open-source-standalone-engine-management-system/

I would love to add diesel logic into the firmware once we find somebody to try this on a diesel.

[harlesk]

Did anyone get any further with this?

I have had Diesel boxes on several of my cars but have always been interested if there was something more open source available.

[brainwash]

I did link above to a blog post which was eventually continued and a project was published some years ago. But after extensive testing I've figured out that it doesn't do much. The consumption is almost exactly the same, there is some additional low-end torque (depending on settings). Basically the computer things it's pumping less fuel so it's compensating. The estimated consumption looks lower than the real one since the car computer doesn't have the correct data to calculate it.

The new TL;DR is: don't bother.

[nctnico]

What I did with my turbo diesel was make the ECU think there was more pressure in the inlet then there actually was causing it to inject more fuel at low RPM. That was a 5 cent diode in series with the pressure sensor signal to give it a bit of offset. Worked like a charm to get more power at low RPM.