tip in knock with turbo

[ft_sjo]

Originally Posted by jamesm

mine happens the instant you tip into the throttle. if i open the throttle ever so slightly and ease into it before stabbing the throttle to the floor it doesn't knock, and never at any point other than after that initial tip in if it does. it always at least begins while the manifold is still under vacuum.

i'll try to find a log that illustrates it clearly and post it.

It's possible you're being mislead by the pure nature of what causes tip-in det. The MAP sensor may be lying to you. You may not be in vacuum at all!

[arghx7]

All I'm saying is, if it's audible, at 2000rpm ish or less, and adjusting timing doesn't help, it may be LSPI. And in that case, there's not much you can do about it. Other knock retard on tip in may just be noise.

[ft_sjo]

I use my ears to monitor det, not logs from a conditioned knock sensor.

[vgi]

Originally Posted by ft_sjo

I use my ears to monitor det, not logs from a conditioned knock sensor.

maybe you could rent those for a small fee?

[Ross]

vgi there not hard to make. Just look up knock ears on google. there are several different ways to go about building them. You can go amplified, non amplified, ect...

[ft_sjo]

You don't need any electronics to make det cans. Just some ear defenders, fuel hose and fittings.

[mike the snake]

I wonder if placement of the MAP and O2 sensors might have something to do with this. Maybe a blank spot caused by turbulence in the intake/exhaust causes incorrect readings momentarily?

[brn12345]

Ok so if the Arghx7 logic on LSPI is correct then I will go one step further and say that the fuel itself is a cause. To test I have access to VP Motorsport 103 (helps that I race circuit in Honda's) which ill fill up with for 2 tanks and observe by logging and hearing (wish I had unicorn ears) and will report back.

In the mean time, I discovered a nice little trick today to get my IAM back quickly. Basiacally I am logging as much as I can even during normal driving. I tried going to 6th gear around 3000rpm and pushing the throttle pedal to the floor. Interestingly in 3 seconds of that my IAM went from 0.36 to 0.47 (see log photo attached) I confirmed this again on my way home moving the IAM up by 0.4 just by doing the same 6th gear low rpm full throttle 2 times in a row. Now only if we had a way to access the RAM and reset IAM manually (hoping we can do this with RR).

[jamesm]

i misspoke earlier when i said zeroing the table caused the iam to not drop, when in fact this works because one of the conditions of entering rough correction mode is an ignition advance value of >5.something (i think 5.9 iirc). so, it doesn't actually have to be zero in those cells, just less than that value.

i've tested this on my car and it works... still have the tip-in issue but IAM stays steady at 1 regardless.

[Td-d]

There's another part of the puzzle which is not being picked up - transient conditions are not only managed with respect to spark, but also with respect to fueling. There are a set of tables that manage transient fueling - the so-called Tau tables (google it for some context). These tables either enrich or lean out the mixture based on a rapid transitions in engine load.

[arghx7]

Originally Posted by brn12345

Ok so if the Arghx7 logic on LSPI is correct then I will go one step further and say that the fuel itself is a cause. To test I have access to VP Motorsport 103 (helps that I race circuit in Honda's) which ill fill up with for 2 tanks and observe by logging and hearing (wish I had unicorn ears) and will report back.

In my experience, race fuel basically eliminates LSPI. It's mostly a pump fuel phenomenon. Ethanol is a different beast--in some ways it has its own unique tendencies.

Originally Posted by Td-d

There's another part of the puzzle which is not being picked up - transient conditions are not only managed with respect to spark, but also with respect to fueling. There are a set of tables that manage transient fueling - the so-called Tau tables (google it for some context). These tables either enrich or lean out the mixture based on a rapid transitions in engine load.

I'd have to look closely at what the tables supposedly due, but that is mostly a port injection phenomenon. In the days of the original AE86, the fuel injectors just kinda sprayed and as long as there was no major hesitation it was all good. So on modern port injection engines, you have what's called a wall film model.



The wall film model calculates the steady-state size of the fuel "puddle" in the port. It figures out how much is sticking to the port and how much is flowing into the engine as the valve opens. Then it will adjust the main injection time to help maintain that puddle. In the event of rapid opening of the throttle, you might have additional injections. They can happen when the fuel is sprayed through an open intake valve and could potentially induce knock with the right temperature distribution.

On the direct injection side, transients can involve two injections per cycle. Sometimes it's two in the intake stroke, sometimes it's split between intake and compression stroke. It tends to help with knock There's no indication that this engine uses multiple direct injections. Keep in mind that with direct injection you don't have to worry about maintaining the fuel film on the back of the valve. However transients aren't fully understood to the point where we can say how exactly port and direct injection might be blended on this particular implementation of port + DI.

There's no easy way to know what's going on without acquisition of cylinder pressure, injector waveform, spark waveform, and individual crank angle degrees. You can certainly play around with tables and see if it helps though--but understanding the "why" is going to be murky. If changing a couple cells in some map makes it go away perhaps at the expense of some negligible fuel economy hit, that would be great.

[Td-d]

^ There's a good thread on the RR forum by NSFW, who dug these parameters out and went through the various iterations of testing them. In a nutshell they add or remove fuel in either rising or falling load condition, and as you pointed out, are to do with so-called wall wetting.

What's interesting for me is that not only are they present in the BRZ ecu, but they are significantly expanded (from 1 in early mid 2000's subarus, to 3 in later CAN models, to 14 in the BRZ). Furthermore, in previous models the tables are either load OR manifold pressure referenced - in the BRZ both are present.

[arghx7]

Originally Posted by Td-d

^ There's a good thread on the RR forum by NSFW, who dug these parameters out and went through the various iterations of testing them. In a nutshell they add or remove fuel in either rising or falling load condition, and as you pointed out, are to do with so-called wall wetting.

What's interesting for me is that not only are they present in the BRZ ecu, but they are significantly expanded (from 1 in early mid 2000's subarus, to 3 in later CAN models, to 14 in the BRZ). Furthermore, in previous models the tables are either load OR manifold pressure referenced - in the BRZ both are present.

Remember that the software is going to carry over across most engine programs. So the tip-in control tables used for an early USDM WRX or STi were for Subaru engines that had to meet a much less stringent emissions standard.

See http://www.dieselnet.com/standards/us/ld_t2.php

They were what's called California TLEV certification -- transitional Low Emissions Vehicle, roughly Federal Tier 2 Bin 9. That's when the standards were moving from LEVI of the 90s (basically, early OBD II cars) to LEVII ULEV (Roughly, Federal Tier 2 Bin 4) of the mid-late 2000s. Software on a BRZ is most likely shared with some random family hauler that has been certified to Federal Tier 2 Bin 2 California SULEV/PZEV standard. It's a big regulatory clusterfuck.

So Subaru had to make an engine family designed in the late 80s to meet a much much stricter standard. Ever wonder why Subarus seemed to get all sort of changes, like smog pumps and new closed loop delays around 2008 model year? That's when the emissions standards tightened up significantly. I'd have to double check the certification applications for the exact model years, but the STi for example went to a Tier 2 Bin 5 (California ULEV) in 2008, when 2005 it was Tier 2 bin 9. That's when you had exhaust AVCS added, and the closed loop delay, and more torque-based engine controls--a big software update.

[Td-d]

Yeah, I've diassembled a couple of earlier PZEV models (mostly 2005 Legacy and Outback models) - they were a pain in the arse to trawl through, as large chunks of code are substantially different from non-PZEV models (not least of all the 5 O2 sensors!).

But we digress from the OP, apologies James