ECU autopsy anyone?

[russian]

It's winter and racing season is over, so I am kinda bored and want to learn something new. Wet dreaming about turboing and hating the idea of p&p meqasquirt I've decided to try to hack stock ECU. I am a software developer and I know nothing about analog electronics so my idea is to take a stock ECU, remove the 20 years old chip and snap my own chip into it, one I would be able to program. So far that's just an idea, because I really know next to nothing about ECUs I have a couple of these 079721-2351, I have a multimeter and I know that I am inventing a bicycle here.

Here are some notes on what I have learned so far

https://sites.google.com/site/mazdaecutechnicaldata/data-flow

and the ECU schematic is attached. So, if anyone wants to help or join me in this senseless waste of time, I would appreciate any help or company

[FB71]

I'm a little bit of a hardware guy, so I'll input some, if that's ok.

ok, first comment on your data flow notes;

yes, the distributor is responsible for base ignition timing. However, the PCM is responsible for ignition advance. From what I can tell, advance lookup is based on RPM, modified by VAF, ECT and TP.

Fuel delivery is based primarily on VAF, modified by RPM, IAT, ECT, O2 and TP. Remember, unlike more advanced PCMs, this one only looks for one one of three states from the TP; idle, WOT and inbetween. Idle and WOT are fairly fixed, primarily because they operate in open loop. Everything that's not idle or WOT will be calculated on VAF, RPM, and ECT. IAT and O2 shift the cursor on the table less.

[russian]

Welcome Jim!
ECU=engine control unit, same as PCM=Powertrain Control Module
RPM=revolutions per minute
VAF=vane air flow
TP=throttle position
IAT=intake air temperature
but what is ECT?

Right, on manual TP is only idle or WOT (Wide Open Throttle) or just not_idle and not_WOT, which is funny. On automatic they do have a potentiometer in the TP, but auto is none of our business.

Can you please explain the whole ignition advance thing? We have the mechanical distributor cap and ignition coil. So, what part of this is variable? What is the PCM controlling here? I mean withing a revolution we have continuity in the distributor cap for maybe like a degree or two for each cylinder, do not we?

[KingFish]

ECT= Engine coolant temp sensor.

On both the manual and auto trans the Festiva sends a simple digital (on-off) signal. No potentiometer, as the auto Festys are mechanically controlled. The TPS is closed for the first 7 degrees or so of throttle opening, then opens untill 70 or so percent (I don;t remember the exact numbers).

As for the ignition timing being adjusted, only on the carb cars is the advance done in the dizzy. The FI cars use a ignition module (igniter) located close to the coil on the pass side strut tower. The ECM send a signal to fire the coil that is based off of the signal form the pick -up in the dizzy (basically just a cam position sensor), and adjusted based on the signals received from the sensors FB71 mentioned above. I posted a chart at some point that broke down what inputs affected what outputs, but I am just too lazy too look it up right now. Try searching for ECM inputs.

FB has much more exp than me. So if I miss represented something I hope he can catch it.

[russian]

So PCM has this Green/White output wire pin 1G/A4 which goes to Ignition Module which control Ignition Coil ground. I guess this controls when the high voltage for the ignition spark happens on the central distributor wire. Now under distributor cap we have the rotor which connects the center wire with one of four spark wires? But do not we have two cylinders igniting at the same time? That's where I am loosing it I guess.

[bravekozak]

Reprogramming an ECU is pretty complicated. I saw how much work it was to program a hydrogen fuel cell computer to work in conjuction with the signal pulses from various stock ECUs on different big rigs. It ain't easy. Good luck.

[russian]

Reprogramming an ECU is pretty complicated. I saw how much work it was to program a hydrogen fuel cell computer to work in conjuction with the signal pulses from various stock ECUs on different big rigs. It ain't easy. Good luck.

I know it's not trivial, but I am a software engineer by trade and want to learn more about cars, so should be doable

[FB71]

ok, think of the distributor as two separate things, but in the same can. First, there's the camshaft rotation sensor. It has three wires; a power, ground and signal, all primary voltage (12v). The signal wire simply switches low when an ignition event approches. This happens four times per cam revolution, or twice per crank revolution (crank-to-cam is a 2-1 ratio). The PCM takes the signal from the sensor, and then determines when to trigger the ignition module (small module next to coil). The delay between the sensor signal and the ign module trigger is calculated based on RPM, modified by VAF, ECT and TP. Now, all of this occurs several degrees of crank rotation BEFORE each cylinders top-dead-center. Thats why this called "Spark Advance". We're determining how many degrees, in advance of top-dead-center, to trigger the spark event for each cylinder. Typically, maximum advance occurs 35-40 degrees Before Top Dead Center (BTDC), with minimum advance being around 6-10 BTDC. Minimum advance is also called "base timing".

Now, inside the distributor is also the rotor. The rotor and cap distribute the secondary voltage (10kV-25kV) to each spark plug. The secondary voltage parts are completely separate (electrically) from the primary voltage parts. Since they both rely on cam rotation, they share a housing and shaft.

[russian]

ok, think of the distributor as two separate things, but in the same can. First, there's the camshaft rotation sensor. It has three wires; a power, ground and signal, all primary voltage (12v). The signal wire simply switches low when an ignition event approches. This happens four times per cam revolution, or twice per crank revolution (crank-to-cam is a 2-1 ratio). The PCM takes the signal from the sensor, and then determines when to trigger the ignition module (small module next to coil). The delay between the sensor signal and the ign module trigger is calculated based on RPM, modified by VAF, ECT and TP. Now, all of this occurs several degrees of crank rotation BEFORE each cylinders top-dead-center. Thats why this called "Spark Advance". We're determining how many degrees, in advance of top-dead-center, to trigger the spark event for each cylinder. Typically, maximum advance occurs 35-40 degrees Before Top Dead Center (BTDC), with minimum advance being around 6-10 BTDC. Minimum advance is also called "base timing".

Now, inside the distributor is also the rotor. The rotor and cap distribute the secondary voltage (10kV-25kV) to each spark plug. The secondary voltage parts are completely separate (electrically) from the primary voltage parts. Since they both rely on cam rotation, they share a housing and shaft.

So I guess the key point here is the fact that distributor rotor connects ignition coil with each spark plug in some range of angles, I guess somewhere around 30 degree total?

And with four strokes during two revolutions of the crank each cylinder is ignited at it's own time, right? With injectors working in pairs I now really need a good animation of this sequence

[russian]

Still playing with the whole thing - just finished a datasheet on the D151811-0360 chip... Cannot attach a "huge" 60K file but hopefully the link will work - https://docs.google.com/viewer?a=v&p...MGVjNGY5NDUzMQ

[russian]

First "practical" result - a handy digital tachometer for Aspire http://www.youtube.com/watch?v=q2DJNHqTCvM

[FestYboy]

can you filter the Hz at which the tach looks at the engine RPM? smooth out the display...?

[KingFish]

That is GREAT! Talk about dedication.

[russian]

smooth out the display...?

I can only update it once a second... Or better only update it if value has either changed by 5% or changed a second ago - should be both smooth and not too slow to react... Anyway, the whole tachometer is just a proof of concept - now I want to control the injectors...

[russian]

Anyway, I've got the car running with me controlling the injectors! 96 Aspire, I used BLue CKP wire signal (ECU 3E/3) - it has a stable signal with 8 spikes per crankshaft revolution. Just the simplest batched implementation so far - 7ms squirts to start and 3ms squirts to run. Next step would be controlling the ignition.

Need help understanding something. What is the purpose of the 2nd CKP sensor wire - the LT GRN one? ECU 3H/17. The signal is less stable - usually it's 5 spikes per revolution, but less than that while cranking. But maybe it's just something wrong with my analyzer. Seems like this 5th spike would identify the cylinders?

Also what are the three ignition system wires? BLK/RED is sparkout - the ECU signal to the igniter. What is GR/BK wire? It seems to be a 50% low-frequency something, if that's simply PWM power supply why would the frequency be variable? Or does it control something? The 3rd wire - GN/R, says it's CID. CID sensor? What is the voltage on this one? I am a total idiot when it comes to electronics, my attempt to measure voltage on this one has actually increased the running engine RPM! Total misery for me...