DIY open source engine control unit

[russian]

I want to share a DIY engine control unit project I am involved in, maybe someone would find it interesting

It is still work a progress, this platform has started a few engines so far but not much driving yet. Most of the crowd involved with the project are there for the potential performance mods but the same platform could be used for eco modding.

It this point our ECU board is still a snap-on module for stm32f4discovery

but we are working on a complete PCB which would contain an automotive-grade power supply and the MCU.

There is some wiki @ https://rusefi.com

 

[stevewaclo]

Are you going in via OBDII port?

Other than the Eco-mod, are performance members of your group looking to develop an electronic "Nitrous" button?

Watched the brief video at your site. And the point was?

Good luck and best wishes.

Beware the warranty cops!

 

[russian]

Are you going in via OBDII port?
... And the point was?

Nope, I am not going in via anything - this thing completely replaces the original engine control unit, so the point of the video is to show that my thing is controlling the engine instead of the oem ECU. By bypassing the oem unit I gain full control over the algorithm and parameters.

 

[stevewaclo]

Russian,

What is your country of residence, or am I overlooking the obvious?

I am frankly in awe of the effort you have put into Frankenstein 0.1, as evidenced by the circuit board picture, but other than an intellectual exercise and challenge, I fail to detect a useful purpose. As you are no doubt aware, there are many OBDII plug-in devices available for engine performance modifications, as well as hundreds of engineers in the automotive industry laboring daily to balance performance and emissions in modern vehicles.

"I gain full control over the algorithm and parameters."

Tell me again where you are going with this?

Best wishes.

 

[russian]

Location info added

While you are right that some (and 'some' means 'not all') oem ECUs are to some extent adjustable one way or another, there is only that much you can adjust. Think of adding more output channels while turbo-charging an engine. If an OEM channel does not have the circuitry to control a wastegate, no amount of OBD-II adjustment would add it. Google for "engine management system" and you would see a world of aftermarket EMS/ECU.

 

[picbits]

Have you had a look at the long established and quite widely used Megajolt and MegaSquirt units ?

 

[russian]

Have you had a look at the long established and quite widely used Megajolt and MegaSquirt units ?

https://rusefi.com/forum/viewtopic.php?f=2&t=239&p=1009#p1009

 

[russian]

I have an electrical issue I cannot figure out - I am just a software engineer

I have my ECU plugged to stock car harness, I also have stock transmission control module. Ignition key is OK, the engine is not running nor cranking nor spinning - my camshaft position sensors should report nothing. I have VR sensors and I am using MAX9926 VR signal input chip

In reality my VR sensor reports some 40Hz noise.

I know that one of transmission solenoids is controlled with a 40Hz signal, and the noise is gone if I disconnect the TCM (transmission control module)

The stock harness schematics is https://rusefi.com/wiki/index.php?title=File:Honda_1.png - I am connected to B13/B14/B15/B16. There is even some GND shielding in the stock harness.

I have a DSO Quad and I do not see the noise on the input - I only see the noise on the output, but that could be me.

Ideas? Anyone close to JC?

 

[KeepItSimpleStupid]

It may be the watchdog timer. I'm having a hard time reading your plots. But the watchdog interval is 45 to 140 mS and would occur with the absence of a signal. I didn't stare at the data sheet for too long.

If the signal is regular and periodic, then that's what I would expect.

Low level signals should basically be twisted pair shielded in this case. The shields should only be connected at ONE end. The end chosen is usually the source end or in this case the sensor end.

I'm probably around 2.5 hours from JC.

 

[russian]

We have resolved this by grounding the negative side of the VR sensor - that's the way it's done in OEM ECU anyway. Once the sensor negative got grounded, the noise has disappeared.

 

[KeepItSimpleStupid]

You always need a bias return path, Grounding turns it into a single ended sensor. A resistor can provide the bias return path with the advantages of the differential input. Without the twisted pair shielded cable, some of those advantages go away.

 

[audioguru]

As you are no doubt aware, there are many OBDII plug-in devices available for engine performance modifications, as well as hundreds of engineers in the automotive industry laboring daily to balance performance and emissions in modern vehicles.

But some people CHEAT.
Every day I hear a racket coming from a kid's lowered car with big pipes replacing its mufflers and it stinks like cars from the 60's because its emission controls were removed.

 

[russian]

 

[russian]

Looks like stable voltage is on my critical path for this project, need to figure out better PID parameters to improve things (video is boring in the beginning but it gets better)

 

[russian]

Circuit Cellar #307 has an article about this project!