1st Gen Haltech ECU Discussion

I'm prepping for a trip at the moment so not looking to mess with my tune, but I will definitely be giving it a go when I get back.

Hmmm, wanna help me fund an R3 and buy my 750?

 

Just based on the name, I would imagine it works better. The whole point of the transient tables are to predict & offset where your cursors are going to be vs where they are because the sensors are too slow

 

It's not even because the sensors are too slow, it's because the actual reaction of the injection event takes too long compared to how fast the air can get to the same destination, that's why we inject more than is necessary, to overcome the delay in delivery and the waste in delivery from the film on the runners. Part of why the new system involves a prediction and a film model.

 

It should go throttle, air, injection in terms of speed. So ya reaction time is probably the better perspective. Either way you're 'predicting' next-cycle MAP and subsequently fuel requirements

 

This new method reminds me a lot of how OEM ECUs do things. I like that they give both options, I'd imagine like any other feature it will work great for some and shit for others. Like this prediction method could be garbage for independent throttle body cars or maybe even something supercharged. But could be rock solid for turbo and na (totally just guesses, I've tested none of those but I used them as examples because they address manifold vacuum changes wildly different from each other)

 

What makes you say it wouldn't work well for a SC application?

 

Like I said I don't know that it wouldn't but you have a much larger transient event when stabbing the throttle versus a turbo or na, you go directly into full boost pretty much. They might have logic to predict this, I'm not sure. Perhaps this is where that induction drop down factors in somehow

 

Do we have any idea if the factory ECU will throw a code based on factory knock sensors being disconnected from the block, but still plugged in electrically?

Like is it looking for a certain response like it does for the O2 voltage, or is it just a resistance, or it is only when a voltage output threshold is exceeded that the code is thrown? (I would lean the latter if I had to guess)

Right now I have P0325 and P0330s (both banks), but that's due to legitimately exceeding the knock thresholds. I'd like to have a contingency plan for removing them for when I have to deal with emissions next year.

 

They don't seem to throw codes when they're deadened... plenty of accounts of people throwing the less sensitive GM ones on and not tossing a code... I'd imagine putting them elsewhere is the same as deadening them. They generate voltage like a microphone, the ECU is just interpreting it... but it looks for resistance to know if you've unplugged them or the circuit is broken somewhere. You'd still have to ground them though with the threads somewhere.

 

I wonder... what mechanism does it use to detect a misfire though? Would hate to try and alleviate the knock codes, just for 4 new ones to pop up

 

I can't answer that fully, it can't detect every type of misfire like new cars can though so my guess is it can only deal with ignition related misfires since it can detect a successful collapse of the secondary field in the coil. My BMW used its knock sensor for misfire detection but I don't think this ECU can. Lemme check my old TCCS manual

 

Auto trans and drive by wire here, don't think the 750 would cut it for me.

I wonder what the differences will be on the new S3 vs the R3

 

Makes no mention of misfire detection



That's all i can find in regards to the Knock sensor duties, there could be more but I didn't find any specific misfire logic section...

You're welcome to check the document yourself, it's very handy

 

Nah it would work for that, you just buy the piggyback harness so the stock ECU can control the trans and the e-throttle.

 

Wow that's great info, and a great document though, definitely saving it.

I think you're definitely correct on the ignition basis - the stored misfire codes were from around the time I had to narrow my spark gap, so like you said, the field wasn't able to collapse during some ignition cycles and it detected that.

Unrelated, I also had a random P1300 (igniter circuit) a little over a month ago show up, and I have no idea why, 10 days after a P0335 (crank position sensor). These codes haven't been active since.

 

Bout 90% sure the S3 is going to be an R3 without WiFi or PDM capabilities. Should have all other R3 capabilities though based on the default map they loaded in NSP.

 

Hope it helps, I missed this earlier... but this is basically what I outline with the collapsing field.

So I'd say they only have ignition based misfire detection.

 

I have the day off, so I went poking around the new settings....

So let's say I want to keep using Transient Throttle until the MAP Prediction has some time to learn cells. You guys think we should set this value as high as it goes so it's not accidentally correcting to bad data during that process?



I'm also kinda surprised you can't disable Fuel Film separately from MAP prediction. I have no idea how to estimate pooling percentages or time constants.

 

What is this ‘fuel film’ y’all are talking about?

How does one tune with it/around it?

 

Also found this odd


So it won't adjust the prediction table to boost if you set to Turbocharged (which I have), but it lets you manually enter boost numbers in the table anyways...