tip in knock with turbo

[jamesm]

Originally Posted by arghx7

How do you know these are all cases of real knock?

i don't. after reading more, looking back through old logs, and talking to some folks that have been around subarus for a while, it seems this is just 'the way it is'. i guess there's some argument over whether it's problematic or a concern, but it's an extremely common issue, even in stock cars.

so, i'm not going to lose sleep over it. until i see more than a degree being pulled, or anything being pulled under WOT, i'm just going to keep going as if it isn't there.

what do you think? assuming this is real (light) knock, is it an issue at the loads and frequency i (and others) are seeing it?

[RickyB]

Originally Posted by mike the snake

Would a standalone system like the Hydra be more capable of dealing with this?

Next week I'll be installing my turbo kit tuned with the Hydra EMS.

Not only will you have the ability to datalog all knock parameters and figure out exactly which cylinder(s) are knocking at exactly what load/RPM points, but with the Hydra all the tables responsible for fuel, timing and cam position are completely open to you and their effect fully documented. No need to wonder what transient tables remain undiscovered to rain on your NA to FI parade.

Good luck with the installation and tuning process.

[jamesm]

Originally Posted by RickyB

Not only will you have the ability to datalog all knock parameters and figure out exactly which cylinder(s) are knocking at exactly what load/RPM points, but with the Hydra all the tables responsible for fuel, timing and cam position are completely open to you and their effect fully documented. No need to wonder what transient tables remain undiscovered to rain on your NA to FI parade.

Good luck with the installation and tuning process.

this car has 4 knock sensors? i was completely unaware if that's the case. the more i hear about the hydra the more i like it. then my rational brain kicks in and reminds me i'm not building a race car lol.

[mike the snake]

Originally Posted by jamesm

this car has 4 knock sensors? i was completely unaware if that's the case. the more i hear about the hydra the more i like it. then my rational brain kicks in and reminds me i'm not building a race car lol.

I think the car has one knock sensor, but the ECU can tell by the timing which hole is knocking.

RickyB, Thanks for the info on Hydra! I'm very excited, turbo and related parts going in hopefully at the end of the week, or early next week.

[brn12345]

Originally Posted by arghx7

How do you know these are all cases of real knock?

The sound of ball bearings on a marble floor is hard to miss. For me i have ZERO full throttle knock, but upon shifting during normal driving as soon as i let go of the clutch and put my foot back on the accelerator it knocks (heard/logged). I am not too concerned about it but what bugs me is that it drives the IAM down. By the end of an afternoon driving the IAM is at 0.25 today. If i go out and find a nice stretch of road and floor it for a while it goes back up quickly to 0.6-0.8.

So regardless if its real knock or not, the impact on IAM is real (unless something else is making the changes to IAM). To confuse a bit more check this post of mine, I am still baffled by why IAM dropped like it dropped there. Nice to see you in this post

http://www.ft86club.com/forums/showp...2&postcount=56

[jamesm]

Originally Posted by brn12345

The sound of ball bearings on a marble floor is hard to miss. For me i have ZERO full throttle knock, but upon shifting during normal driving as soon as i let go of the clutch and put my foot back on the accelerator it knocks (heard/logged). I am not too concerned about it but what bugs me is that it drives the IAM down. By the end of an afternoon driving the IAM is at 0.25 today. If i go out and find a nice stretch of road and floor it for a while it goes back up quickly to 0.6-0.8.

So regardless if its real knock or not, the impact on IAM is real (unless something else is making the changes to IAM). To confuse a bit more check this post of mine, I am still baffled by why IAM dropped like it dropped there. Nice to see you in this post

http://www.ft86club.com/forums/showp...2&postcount=56

you can negate the effect on IAM by reconfiguring the base and advance tables to not allow correction at very low loads and rpms. this is what i'm doing now, and it has solved the issue of the iam dropping because of the knock, though of course the knock itself (real or otherwise) is still there. basically this involves ensuring that the areas where you see the knock have all of their timing in the base table, with zeroed out advance (again, this is only is very low load, low-mid rpm range). with no advance to remove in these cells, it won't cause the iam to drop, effecting the entire rest of the map.

[brn12345]

I would think that would solve the problem of removing timing in the lower rpm area yet we would still have fake/real knock which would cause the IAM to trend down. I don't think this would solve the IAM dropping issue. On another tangent, I have it on good faith that temperatures, especially oil have an impact on the mapping, yet in all available definitions I cant find a oil to timing map.

Originally Posted by jamesm

you can negate the effect on IAM by reconfiguring the base and advance tables to not allow correction at very low loads and rpms. this is what i'm doing now, and it has solved the issue of the iam dropping because of the knock, though of course the knock itself (real or otherwise) is still there. basically this involves ensuring that the areas where you see the knock have all of their timing in the base table, with zeroed out advance (again, this is only is very low load, low-mid rpm range). with no advance to remove in these cells, it won't cause the iam to drop, effecting the entire rest of the map.

[jamesm]

Originally Posted by brn12345

I would think that would solve the problem of removing timing in the lower rpm area yet we would still have fake/real knock which would cause the IAM to trend down. I don't think this would solve the IAM dropping issue. On another tangent, I have it on good faith that temperatures, especially oil have an impact on the mapping, yet in all available definitions I cant find a oil to timing map.

It does solve the aim dropping issue. The timing itself is the same. Final timing is base + (iam * advance) + compensations. The tables were reconfigured such that total timing remains the same.

By zeroing advance in that area, flkc has no timing to pull in those areas, so iam never drops as a result.

i haven't found any maps referencing oil temp at all, but then i havent really been looking for it. i'm using brzedit and romraider, so not sure if ecutek has them or not.

[arghx7]

Originally Posted by brn12345

The sound of ball bearings on a marble floor is hard to miss. For me i have ZERO full throttle knock, but upon shifting during normal driving as soon as i let go of the clutch and put my foot back on the accelerator it knocks (heard/logged). I am not too concerned about it but what bugs me is that it drives the IAM down. By the end of an afternoon driving the IAM is at 0.25 today. If i go out and find a nice stretch of road and floor it for a while it goes back up quickly to 0.6-0.8.

So regardless if its real knock or not, the impact on IAM is real (unless something else is making the changes to IAM). To confuse a bit more check this post of mine, I am still baffled by why IAM dropped like it dropped there. Nice to see you in this post

http://www.ft86club.com/forums/showp...2&postcount=56

I know you're not the OP but I'm going to address your posts because it is still relevant. I downloaded your logs and made some plots.




Top chart shows engine rpm in blue, knock retard events in red. Bottom chart shows vehicle speed kph in black and IAM scaled up where 100=1.0

I'm going to throw something out there. I think in some cases you're getting LSPI. Considering the nature of LSPI it's not going to happen on every detected knock event, but the fact that you are hearing "ball bearings on a marble floor" sometimes is a big sign that this isn't always false events.

What the heck is LSPI? It stands for Low Speed Pre Ignition also known as Super Knock or Mega Knock. It's not very well known in the aftermarket. It sounds scary, and it kind of is. It's what can happen when a turbo engine spools up at low rpm, especially turbo DI engines. It usually happens at 2500rpm or less, especially under 2000rpm. You don't commonly hear about it because DI engines with factory turbos are tuned to stay out of the "no man's land."

Usually an LSPI event will hit peak cylinder pressures of 140-190bar, similar to a diesel. Your pressure rate of gets very very high. That's where the noise comes from. It's different from regular old knock or preignition from hot spark plugs and such. There's basically no simple solution for it, and it's somewhat random. Pulling timing often doesn't solve it, although a richer mixture can help sometimes. Theories for its causes abound and I won't dive into them here.

So basically, if you're actually hearing a loud "marbles in a can" sound at low speed tip-in, and some simple spark retard is not working, you have limited options to solve it... there are guys with PhD's who have been trying to figure it out for years. I'm not trying to sensationalize anything--you probably make it go away with enough fuel and spark retard. It's important to recognize though that this problem exists and if you really do have it there isn't an easy solution.

[ft_sjo]

@arghx7 - The tip-in knock exists throughout the rev range, even when the transient spark table is heavily modified.. it can be reduced, but not eradicated.. all the tables are not defined, e.g. there are no speed compensation tables around transients.. which is pretty crucial!

[jamesm]

Originally Posted by ft_sjo

@arghx7 - The tip-in knock exists throughout the rev range, even when the transient spark table is heavily modified.. it can be reduced, but not eradicated.. all the tables are not defined, e.g. there are no speed compensation tables around transients.. which is pretty crucial!

what did you find to be effective in reducing the knock (in the transient table)?

[jamesm]

Originally Posted by arghx7

I know you're not the OP but I'm going to address your posts because it is still relevant. I downloaded your logs and made some plots.

Top chart shows engine rpm in blue, knock retard events in red. Bottom chart shows vehicle speed kph in black and IAM scaled up where 100=1.0

I'm going to throw something out there. I think in some cases you're getting LSPI. Considering the nature of LSPI it's not going to happen on every detected knock event, but the fact that you are hearing "ball bearings on a marble floor" sometimes is a big sign that this isn't always false events.

What the heck is LSPI? It stands for Low Speed Pre Ignition also known as Super Knock or Mega Knock. It's not very well known in the aftermarket. It sounds scary, and it kind of is. It's what can happen when a turbo engine spools up at low rpm, especially turbo DI engines. It usually happens at 2500rpm or less, especially under 2000rpm. You don't commonly hear about it because DI engines with factory turbos are tuned to stay out of the "no man's land."

Usually an LSPI event will hit peak cylinder pressures of 140-190bar, similar to a diesel. Your pressure rate of gets very very high. That's where the noise comes from. It's different from regular old knock or preignition from hot spark plugs and such. There's basically no simple solution for it, and it's somewhat random. Pulling timing often doesn't solve it, although a richer mixture can help sometimes. Theories for its causes abound and I won't dive into them here.

So basically, if you're actually hearing a loud "marbles in a can" sound at low speed tip-in, and some simple spark retard is not working, you have limited options to solve it... there are guys with PhD's who have been trying to figure it out for years. I'm not trying to sensationalize anything--you probably make it go away with enough fuel and spark retard. It's important to recognize though that this problem exists and if you really do have it there isn't an easy solution.

well that's no fun. does it help that my knock usually occurs when the car is still in vacuum?

[ft_sjo]

Originally Posted by jamesm

what did you find to be effective in reducing the knock (in the transient table)?

I have a completely different setup to you, so the values will be meaningless.. but here you go..





FWIW, i've only ever seen det on vacuum once (not on an FA20), and that was really odd.

[jamesm]

Originally Posted by ft_sjo

I have a completely different setup to you, so the values will be meaningless.. but here you go..




FWIW, i've only ever seen det on vacuum once (not on an FA20), and that was really odd.

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.