rusefi.com

We do have some cool hardware, but hacking a stock ECU would always be the cheapest options. My way of hacking an ECU would be to remove the old, closed & slow chip and hack my fast ARM or something in place with a small daughterboard.

So, OBD-I OEM Honda. Looks like the chips I want to take away use 1.778mm/0.07" pitch. For a prototype I would probably simply solder some wires in, but what would do for a more permanent solution? A "solder tail"? Something along the way of PRECI-DIP 350-90-1xx-00-012?

The only problem this one is not in stock. Looking for a 32+ pin 0.07" solder tail.

Old Honda?



Pre-OBD for you.

Actually, good OBD- told you if an injector was shorted or open, any one of the 3 MAP sensors or other sensors were faulty....
Didn't need a special tool to do it either- a few LED's on the side of the board gave fault codes.

OEM ECU from Honda was hacked i think many times... Search about HonData

acab wrote:OEM ECU from Honda was hacked i think many times

Thank you Captain Obvious. This does not really help me with my 0.07" connector?

This sounds really interesting.
However I imagine even if a Honda solution were to be worked out it would still be pretty tough to port it to a different ecu type. Way above my skill level
An alternative (less $) approach would be to use a parallel processor and modify the crank and injector signals externally to the stock ecu. By delaying the crank signal, timing can be retarded (and advanced) relative to the stock settings. For injection, the start pulse timing is fine just the end of the pulse needs to be adjusted. (According to the injector size/ manifold pressure and fine tuned by lambda.)

Is there room on rusEFI for a parallel processor ECU? AKA the dreaded piggyback

acab wrote:OEM ECU from Honda was hacked i think many times... Search about HonData

The ECU I posted hasn't been hacked, until recently, and even now the 'hacks' are experimental; it didn't come on any Civic, Accord, Prelude or Integra- couldn't get one of those EFI in 1982, could you?
There are only a few thousand of those ECU's out in the wild.

If you're capable, I'll send you a complete EEPROM dump.

sturovo wrote:This sounds really interesting.
However I imagine even if a Honda solution were to be worked out it would still be pretty tough to port it to a different ecu type. Way above my skill level
An alternative (less $) approach would be to use a parallel processor and modify the crank and injector signals externally to the stock ecu. By delaying the crank signal, timing can be retarded (and advanced) relative to the stock settings. For injection, the start pulse timing is fine just the end of the pulse needs to be adjusted. (According to the injector size/ manifold pressure and fine tuned by lambda.)

Is there room on rusEFI for a parallel processor ECU? AKA the dreaded piggyback

Most easy would be intercept table reads (Honda did 'invent' the PGM thing- where values were programmed in and read out and interpolated... starting with that ECU I posted earlier- Honda's first EFI system) and substitute data out to the ECU. Hard part is knowing what data the MCU wants, then calculating what it needs and handing it back before the MCU faults.

Let's get back on-topic: this thread is about soldering rusEfi body into some OEM ECU, Honda in this case

Just to save the link - QuarterHorse is another example of HW hacking an OEM ecu http://www.moates.net/quarterhorse-for-fords-p-199.html

I am working on this with my Vitara ECU, my idea was to get the software running on a proven platform before building a new circuit board.
From what I recall the Hondas use a non-Motorola chip which would add complexity. On a Denso with a Motorola type processor the tables aren't too difficult to locate, most Densos that I've looked at use similar code on ECUs of the same age. I hacked the early Miata ECU using a lot of knowledge gained from hacking the early Vitara 16V ECU.

I tried to find a way to hack OEM ECU BOSCH MOTRONIC by MOTRONIC ONLINE

But this one is not far so cheapest to be so useless. Besides it not open source, then needs to pay for firmware and other software

If we know enough about each pin, it should be easy enough to spin a small board with the STM32 on it. The primary questions are, what voltages will be at the pins, how much current does each pin need to drive or sink, are they analog or digital signals, and what bandwidth do they need to be. With those hand full of questions, I could see a carrier board being made that would allow rusEFI to happen as a possible modification to an OEM ECU.

I think the idea is not entirely capable of carefree life.
Because for each unit will needs in a specific processor PCB. You can make a module that can be connected with the free pigtails, but the end user in this case would not be free from errors.

By the way, I dug up ECU from SUBARU EA82T, that worked in a normal mode with VAF flow-plate, and a mechanical ignition distributor on the camshaft gear.
But opening it I found the MAP sensor.

Information on the network is not found on any pinout or even datasheet.

But I really need to launch this motor .
Inside LED blinks error code like a Morse, breakage/short circuit some nozzles, when was the last time the ECU patched to engine harness plugs.
I couldnt start it during normal produced spark, fuel delivery and proper compression in the cylinders.
Any idea how to resolve it?