timk

Spark plug #64 on an 8 bit little endian system? Kudos!

Braineack

iTrader: (62)

But your motor doesnt have a turbo attached to it:

MX-5 Miata Forum - View Single Post - Question about Turbos

Joe Perez

iTrader: (8)

Haha.

I honestly debated with that. At work, all of our SSI stuff is big endian, however it winds up displaying as though it were little endian when viewed on a logic analyzer, so the end result is the same (read from left to right.)

On the other hand, most humans are accustomed to reading binary in big endian left-to-right when presented in print, and the clock on my shelf displays in this format as well.

jnshk

I don't know if this is still relevant to the current direction of the thread, or if it is precise enough for your intentions, but here is a spark map that I put together to emulate the OBD2 computer's indications on my 1997 1.8L California car:

More background on it can be found at the original thread: https://www.miataturbo.net/megasquir...=tuning+octane

18psi

iTrader: (76)

/\ That's very interesting and still very relevant to this thread, thank you for posting.

Leafy

iTrader: (1)

What the heck is up with the -10 degrees of timing in the bottom left corner. The only time I've ever seen negative timing like that is when I was trying to get the formula car to idle below 4 grand with the shitty intake manifold that leaked like a sieve.

18psi

iTrader: (76)

I don't think the car will even hit those areas at idle so probably doesn't matter?

Braineack

iTrader: (62)

it's just 10° After TDC. No biggie.

If we look at 750RPM (typical idle speed) and 10°BTDC (oem timing at idle), it takes 25° or 5.434ms for the flame front to reach 15°ATDC, or about 2mm into the downward stroke.

15°ATDC is an assumption of where the flame front is hitting for this comparison

Since that time never changes we can figure that within the latency of 5.434ms for the spark to ignite the flame and hit the piston since at 500RPM the crank rotates one degree in 0.33333333 milliseconds. At 500RPM 5.434ms is about 16° of crank rotation.

So if we spark at 10°ATDC, then the flame front has moved to ~26°ATDC. This should put the piston top maybe ~4mm into the downward stroke, so less power should be outputted at this time.I know I did this flame front math wrong, but you get the idea.

This is possibly an effort to stabilize the motor if the RPMs drop that low. If you advanced the timing when the RPMs dropped, it would increase the power and increase ultimately bump up the idle speed. I dont believe the stock ECU uses timing control like my MS3Pro to help control idle speed, so I'm assuming it's a way to allow the idle valve ot have greater effect at low rpms.

Joe Perez

iTrader: (8)

Very interesting. It's really becoming clear to me why everyone was able to get away with bumping the static timing by 5-10° (by rotating the CAS) and not have any problems, back in the day before enlightened ECUs.

JasonC SBB

Huge jump in timing from 1000-1200 rpm, 25-42 kPa. That is probably not the static table in the ECU, but observations made, with AC and idle compensation included. Such a static table will cause severe idle instability.

jnshk

Jason is correct. The table was compiled/interpolated based on observations of OBD2 polling, and around idle there was some compensation done to stabilize things between A/C off (low load) and A/C on (heavy load), which is also why it is scaled somewhat non-uniformly so that I could isolate some key areas around idle and cruise.

The stock computer calculates spark based on rpm and MAF load (I believe) and possibly a few other factors. Since the MS is configured differently, this was my best interpretation at emulating it in effectiveness. When I was testing it out, the readings synced up quite well between MS and OBD2.

Note on the negative timing cells -- these cells were basically areas where the car was struggling to not stall. Braking to nearly 0mph with the transmission in gear and crawling along at low speed, high gear, throttle mashed open.

miatauser884

iTrader: (11)

Joe,

What are you seeing as far as knock with your knock sensor? The following is my table that I have been tuning out knock peaks by lowering timing. Mine is much more conservative now that I have a knock module. I need to check my hardware latency as it is set to 21 microseconds.

Joe Perez

iTrader: (8)

Knock sensor? Bah. A Jedi craves not such things.

richyvrlimited

iTrader: (1)

Luddite

18psi

iTrader: (76)

he has a 1.6
doesn't even have a knock sensor.
and he probably cares not about it enough to rig one up lol

y8s

iTrader: (8)

luddite is right. drag your ms paint white pre-paste image to a tiny rectangle and THEN paste your graph. we dont' care about 3/4 of the content of that bs.

emilio700

iTrader: (3)

This is not the flame front you are looking for.

miatauser884

iTrader: (11)

There you go bitches

miatauser884

iTrader: (11)

Although the knock peaks are responding to reducing the timing. I'm wondering if I'm tuning out peaks that represent negligible knock. Unfortunately there are no clearly defined knock thresholds that we have to follow.

What hardware latency are others running on the ms3x?

Joe Perez

iTrader: (8)

You must learn the ways of the tune if you are to come with me to Thunderhill.