Stock ECU stim - help me out!
 

Let's see if the mt.net collective can help me with a little something...

I have built several ECU stims, much like this one...



...to test all the ECUs I send out of the door. It dawned to me at some point that I can use these with the stock ECUs, and with the use of a scope, I can extract very useful information (ignition timing, ignition dwell, cranking pulsewidth, IAT/CLT corrections, IAC valve %, etc).

So I modified my 90-93 1.6 stim to use on a 90-93 ECU.

I got excited when it actually synced and got spark and fuel on the stim, only to lose fuel a little later. In other words:

1) I power it up
2) I get spark
3) I get fuel
4) After 16 injector pulses, it stops injecting (the LEDs go off).

So I'm trying to figure out why.

Here is a list of what is actually connected to the ECU on this stim:

1) Constant +12V
2) Switched +12V
3) Spark output for cylinders 1&4
4) Spark output for cylinders 2&3
5) A/C relay LED indication (ie if the A/C is actually active)
6) P/S input (through a jumper)
7) A/C thermo switch (ie A/C button)
8) Cooling fan LED indication
9) Clutch/neutral switch input
10) All grounds
11) CAS input (crank+cam)
12) TPS as per the stock TPS sensor: 1N is +12V, 2L is +5V, 0% TPS grounds the 12V, 100% TPS grounds the 5V, 1-99% leaves both floating.
13) Coolant sensor (potentiometer, only two legs connected)
14) Air temp sensor (potentiometer, only two legs connected)
15) AFM (5V potentiometer, all 3 legs connected)
16) Idle valve (LED)
17) Injectors 1/3, 2/4
18) ST SIGN input

I haven't connected:

a) Check engine light (1E)
b) Diagnostic connector wires that go to 1D, 1F and 1K
c) Stop light switch (1O)
d) Heater control unit (1S)
e) Tail fuse (1U)
f) Short connector (2H, present on Cali ECUs only?)
g) Input from igniter module (2I)
h) O2 sensor
i) Purge solenoid valve

Any ideas?

Not a clue, but this is freaking cool!

Could it be that the stim doesn't perfectly replicate the CAS signals and the ECU just loses sync?

Might be worth hooking up an actual CAS to rule that out?

i wouldn't think that. I can drive a MS by just power ground and CAS.

My bets on g and h.
ECU not knowing what the ignitor and AFR's are up to.
Maybe :noob:

An ECU will run without the O2 sensor. I've never tried it without the IGf pin connected, but that would be my guess.

Parallel the spark outputs (you might need diode isolation) and send them back into that input.

I just tried that, no go. It appears IGf is 12V high, and needs to be pulled down, so I need to get a transistor in there to pull IGf to ground on each ignition pulse.

Yup, the transistor did the trick!

So two immediate notes so far:

1) The stock ECU on the batch injection 1.6 will fire ALL injectors simultaneously while cranking - ie while it receives the cranking signal on pin 1C.
2) If the stock ECU does not receive ignition pulses back from the igniter module through IGf for 16 NE events, it stops all injection events to protect the engine from flooding and/or destroying the cat. In other words, if your igniter is fried, you will also not get fuel after cranking for a while.

Again, well done mt.net!

Well done Demitris too.

<Wonders if the MegaSquirt coders would be willing to add simultaneous injection during cranking events to the MS code>

At least that explains why the OEM ECU cranks and fires so quickly compared to the MS.

Oh, many more surprises to come:

Cranking dwell at 0degC. 50ms. Yes. Fifty milliseconds! Does not change with voltage, and is only so during cranking (ie ST SIGN is on).

Yowzer!

Can we replicate that on the MS?

Might fix my complete refusal to start when below freezing.

Looking forward to a complete ignition timing table!

Are you going to stim all the year ECU's? This is fascinating...

So far I've got a 90-93, 94-95, waiting for a 96-97, I have a complete harness for a 99-00 (waiting for the 99-00 ECU, immobilizer + keys to arrive) and I also have my own 01-05 ECU, immo and keys. So yes I am going to try and stim them all. I do have all the stims for these years, I just need to mod them so that will work properly with the stock ECUs (much like this 90-93 stim).

I'm afraid it will get significantly more difficult with the 99-05 ECUs. I have no idea how they will react to all the things missing (EGR valves, purge valves, etc) and whether or not the data collected will be skewed or not.

Getting the ignition timing on the stim manually might be a little difficult - I may have to rig something to measure it automagically and report it on a small LCD.

If your cranking settings are based on the DIYAutoTune basemaps, then that would explain it. It is actually quite off compared to the stock ECU, below 30*C.

My settings *are* the DUY basemaps :facepalm:

Do you have anymore data on it then I can plumb it in?

Thanks!!

I need to get the data out first - there is MASSIVE information to be had from this ECU alone.

i want to know the timing differences with and without a/c

0.

Interesting notes about the cranking dwell

interesting. thanks.

Hope you have a 94+ euro 1.6 ECU to work on as well :bowrofl:

On a different note, the stock 2002 is manipulating the idle using the ignition advance - with a timing light you can see the white dots drifting left and right, while jumping GND to TEN makes the timing solid.

What it looks like, adjusting RPM up and down, activating A/C.

I have an spare 1998(99) 1,6 euro ecu here, including 1 key, the lock and the immobilizer unit. I could send it to you if you would be interested.

Greets Sven

Absolutely! I'll cover the shipping, I don't need the lock, just the ECU, immobilizer and key. I already have the coil (antenna) that reads the key. I'll PM you my shipping address.

It already does.

At least, it does it you have your cranking RPM configured properly.

?

My MS1-equipped car fires up at least as fast as one with a stock ECU. You gotta wait about a second after switching the key to "ON" before continuing to "START", as it takes some time for the ECU to boot up, but that's about it.

The MS2 fires all injectors simultaneously during cranking...if you have untimed injection and not sequential injection. Booooh.

Btw the MS1 did start the car as fast as the stock ECU, as it didn't wait for a full sync on the wheel pattern AND it fired all injectors simultaneously as well.

MS1 does not = MS2 :P

Seriously?

(I mean, I know that MS1 :ne: MS2, I just assumed that this specific behavior was so fundamental in nature that its inclusion in MS2/MS3 was axiomatic regardless of injector timing or configuration.)

It would be interesting to see what the ecu does to fuel and spark by atifically altering the atmosphereic pressure it sees.

Btw did anyone one that the 90-93 ECU does barometric correction? It has a baro map sensor inside it...

Some region cars do, not all. That baro map sensor isn't always populated.

Read about it on the 'slocketing ECU' thread on m.net

Interesting. This ECU came out from a euro-spec 91 if that matters.

The Adaptronic can do batch fire when cranking too, which seems to help.

I have a JDM 99 ECU that has no EGR and runs fine without any emissions stuff, so if you get stuck hit me up! I'd love to know how the OEM ECU is dialled in!

Cheers

The $1,000,000 question is whether it needs the immobilizer to work...which would it make significantly easier for me if it doesn't need the immobilizer module.

I don't think it even has an immobilizer, my key is just a chunk of metal??

Ah yes, that would help A LOT!!!

Message me your email address.

The ECU is currently running my race car but I may be able to get my hands on a spare tomorrow. I'm flying to London on Saturday so if it works out I will post it from there otherwise you will have to wait a couple of weeks.

Cheers

What does it do for fueling with high intake temps, especially after startup?

I haven't been there yet, I've extracted all the data on cranking PWs at various system voltages, and a single running PW at 250rpm and almost zero load, again at various voltages, coolant temps and time stamps. This is a really long and painful process, as I have to change multiple parameters by hand (ie potentiometers) and then measure with the oscope, switching between the stock ECU and a MS to measure the parameters I am sending (CLT, IAT, RPM).

get on it bro!

Patience is a virtue bro!

I've been driving my car for the last 6 months with an obvious electrical problem...I know about patience.

Lol

I know that the US-spec '90-'93 cars had a baro sensor on the ECU board. I'd assumed that any Miata (or any car) with a VAF sensor would need one, as it's the only way to be able to correlate volumetric airflow to actual air mass. Without knowing barometric pressure, the ECU would not be able to distinguish between 100 CFM in San Diego (which is at sea level) and 100 CFM in Alma, Colorado (3.156 meters / 10,578 feet above sea level) despite the fact that there is 30% less air per cubic foot in Alma than in San Diego.

So I can pretty much guarantee that as barometric pressure increases, you are going to see an increase in fuel PW and a decrease in spark timing at all operating conditions, because an increase in barometric pressure will cause an increase in computed load.



Do you have a two-channel scope? I ask because I would think that RPM would be easy to measure without using the MS- just attach CMP to one of the scope channels, set a counter for frequency, and divide.

Or you could always do a parallel install on the bench. :D

Yes, it is a 2Ch 60Mhz scope. I am using both channels to correlate the injection timing (one probe on NE, one on the injector) and the ignition advance.

Thanks reverant for finding out this information.
Hopefully this info leads to better understanding of what the stock ecu does for startsup and other conditions.

Well, there you go. If you've already got Ne (which we often call CKP, I mis-spoke earlier when I said CMP) connected to the scope, then all you need to do is dial out the horizontal far enough to do a frequency count on that signal (I'm assuming you have a digital scope with built-in measurement functions) and you can then derive RPM from that without needing the MS. Just divide the frequency of Ne by 120 and you have RPM.

If your multimeter does frequency counting, you can use that instead, and not even have to bother adjusting the scope's timescale.

Any new nuggets of wisdom extracted from the stock ECU? :)

Yeah, the stock ECU fires the injectors half the times the Megasquirt does, and that fucked up everything that I've measured so far, concerning the injectors. I need to figure out how to get the MS to fire the same number of injections as the stock ECU.

this still the early 89-93 ECU i.e. batch?

How very odd, almost like simultaneous injection on the MS, with 1 squirt per cycle?

When you said the ECU adjusted idle by altering advance, what ranges are we talking here? Thanks ;)

Exactly. It does half the injection events, without about twice the pulsewidth.

Still on the 89-93 ECU. I also have a 94-95 ECU to play with, but no stim for that year yet.

That was with the stock 01-05 ECU. I did not measure any numbers as that was with the ECU in the car, when I was still running a parallel setup. When I switched the ignition control to the MS, the idle was all over the place (the idle valve was still under the command of the stock ECU). Switching back to fuel only gave an instant rock-solid idle. My spark map was flat at 10* during idle, so the only reason that would explain the oscillations (+/- 200rpm) would be that the stock ECU uses the ignition as well to control the idle.

Btw that was the main reason why I decided to go for a full standalone setup.

altered the baro sensor pressure and seen any results?

I can't WAIT for the 94-95 ECU info. Which ECU do you have; there were 2 in 95 in the U.S. IIRC.

You know, I never noticed this before.

Rooting around on my hard drive, I found the following scope trace, which I must have taken on a running car, given how ugly the INJ pulse is:

https://www.miataturbo.net/attachmen...ine=1315332142

My notes say that this is the INJ 1/3 wire, but that's about it. What bugs me is the filedate- this picture claims to have been taken in 2008, and I'd have already been running a Megasquirt by that time. Maybe the date got corrupted in a file copy operation...

But yeah, there's clearly only one injector shot per engine cycle here. At moderate to high load, I would think that the MS's two-shots-per-cycle would, in theory at least, produce slightly more even fuel distribution, as each amount of fuel will spend the same length of time "lingering" in the port regardless of which cylinder it is on. At idle, however, it may well be that the stock ECU's method is superior, as one shot of fuel can be more precisely metered than two half-shots.

Alas, this capture was taken at idle, so perhaps the stock ECU switches to two-shots-per-cycle at some higher load condition? Can you test for this?

I can't with the onboard baro sensor. At least not someway I can think of, I couldn't see a port on it somewhere.

BPL9 ecu. Got it from ChicksDigMiatas.

Up to 2000rpm and full load on the AFM (pot), it was still 1 injection event per cycle. I haven't tested further.

there is a hole in the small window of it, I was figuring a small tube glued onto that connected to syrunqe (spelling) and pressure gauge

2002 ECU. Look at the funny injection sequence - semi batch or something?!?!? It is fully batch below 500rpm - but only during the first cranking event. Above 500rpm this is what you get:



It goes like this:

1-2
1-3
4-3
4-2

Oh and y8s...alternator no chargy below 300rpm.