Tuning crank enrichment

I'm having trouble getting consistently good starts. Running an Emtron KV8 ecu, Ford Duratec 2.3 engine.

I can tune it one day and it starts up great. But then 10 minutes later without making any changes it cranks and then backfires. And after several tries, then it will finally run.

It seems to like cold air, it cranks and starts on the first try when the air temp is cold, less than 10 degrees.

But on hot summer days, or when the engine bay is heat soaked, that's when the problem is the worst. It will just crank then backfire. And eventually start up.

Any changes I make either make it worse or make no difference. I'm not sure what I should be doing with it.

Cranking compensation should change with engine temperature, and how long your are cranking. Usually lots of fuel for the first few cycles, then reduce it quickly to prevent flooding. Are you sure the engine crank & cam sensors are getting good consistent signals (no errors) when warm?

Yup, that is how it's working. I started with the default cranking table, which varies by engine temp and cranking time. And just been tweaking it attempting to get better results.

Both sensors go green right away. If they didn't it wouldn't be injecting/firing at all I believe. But I'll be sure to look at those in the log next time.

How effective/accurate is the lambda sensor during cranking?

you cant have accurate reading while cranking.

my recommendation is to remove fuel 10% at a time until it stops turning on and then add 10% until it feels just right.

maybe start by removing 50% and see if it still starts it means you have way too much fuel to start with

you might have to also do pre-cranking and afterstart if you have it.

before you tune your cranking make sure your fuel take is tuned right.

also, check your timing during cranking and make sure it's synced.

what is your injector pulsewidth?

From some of the better startups, these are the observed pw:

Cold cranking pw is around 25.5ms, hot cranking is 6.2 to 6.3ms. FIC 445 injectors, running 58psi returnless, so they're 530cc injectors.

Checked the sensor syncing in some logs I took. It gets 360 sync almost instantly when cranking starts. 720 sync happens about 1 second later every time. When it works well, it fires up almost immediately after getting 720 sync.

Whoa, I think I may have found my problem!

Take a look at the scope trace I took while cranking today. Blue trace is the single-tooth cam signal.

While cranking it's working perfectly normal. It's pulled high by the pull-up, and triggers to ground the way it should.

But then as soon as the engine fires and starts to run, the sync signal gets inverted! It's being pulled low, and triggers high. I don't know if this happens all the time or just sometimes, but it could definitely explain why I sometimes have startup issues!

The trigger tooth is small enough that the change in sync angle from the inversion is "close enough" that the ecu can just go with it (The 720 sync status is still green). But depending on exactly when it happens, it might not work every time.

When the engine does start it always runs great. It's just hard to start sometimes.

I have spare sensors, I'll give another one a try. I'll double check the wiring, but I'm confident it's correct.

Besides a bad sensor, what else might cause something like that?

first thing i am not a professional tuner, just a hobbyist. so i am looking at this from a simplistic/basic point of view.

the BADSYNC picture to me looks like a Hall effect sync sensor and a reluctor style trigger sensor out put, as its amplitude increases with speed. with this in mind:

do you have your threshold voltages for the sensor set to vary with speed and also temp? old sensors in some cars are notorious for dropping output voltage when hot and at low cranking speeds.

Yup, it's a VR crank, and Hall cam sensor. No signals are being dropped, they all come through clean. Problem I see is the sync signal inverts polarity when the engine speed goes up...

I think it's a bad hall sensor. Since you've identified that it's worse when hot, I would look for one that can live in whatever environment (valve cover?) is required. I know there are some high-temp OEM ones if you can fit them.

Why don't you try heating it up with a heat gun with the engine cold and see if you can reproduce the problem. Or reproduce your bad sync capture.

Just tried a new sensor, and it's doing the exact same thing.

Did some googling, and it seems like some of these OEM sensors do this by design. Though I have no idea how or why...

I should point out that the stock cam sensor for this engine is VR. But I've replaced it with the hall sensor from a newer version of the same engine. The VR sensor is designed to trigger on a single-tooth cam. And the hall sensor version is designed to trigger on a multi-tooth cam for the VCT system. But the hall sensor is much easier to work with, which is why we use that one.

I've taken a closer look at the scope logs, and I don't think this is a problem.

I counted the number of teeth from the falling edge of the sync to the missing tooth of the crank, and it's always exactly 20 teeth. No matter if it's the normal or the inverted, and even at the time when it inverts it's still 20 teeth. So in theory the ecu should be fine, and shouldn't even notice the difference. Since the ecu only triggers on the falling edge, and that happens at the exact same point all the time. So I think this is fine and normal for this sensor.

Screenshot attached of the new sensor, this scope was taken during cranking and startup. You can see as soon as the engine fires (yellow trace) that's when the sync signal (blue trace) gets inverted.