Stage 2, knock at low load, need an advise.

[lepich]

Got an update:

After reducing timing again and again in the same area (1-1.3 g/rev - 2-2.5K RPM) without any luck, I've reflashed back to stock and... No knock! Some knock while idling, as usual, but nothing on low RPM, as it uses to be.

So, it was not timng, as I've had about 5 to 10 degrees less then stock in that area.

Then I've copied AVCS and timing low load area (0-1.3 g/revs) to my stage 2 map, reflashed, and knock is gone. I think the problem was AVCS table, low load area.

Having low load part AVCS from a stock map is fine, as it should not affect WOT performance.

[Infamous1]

What did your AVCS looks like before? It was likely fuel related then.

[lepich]

It was like this:

 

Need to drive more.. today, without laptop I had a feeling that timing was pulled again in the same area.. Not sure though.If it is again, this is a very strange knock.. appearing and dissapearing like a who knows what:-)

[thebootyhouse]

It was likely fuel related then.

 

 

Could be a dirty or clogged injector?

________

LIVE SEX WEBSHOWS

[Infamous1]

It was like this:

 

Need to drive more.. today, without laptop I had a feeling that timing was pulled again in the same area.. Not sure though.If it is again, this is a very strange knock.. appearing and dissapearing like a who knows what:-)

Interesting table, high values for the stock turbo.

[Infamous1]

Could be a dirty or clogged injector?

No, I assumed fueling because of the non stock AVCS values (higher) in the cruise regions.

[lepich]

Damn! is still there, see LV.

What about fuel? It is mostly CL on that areas.. SAme as on a stock map

[Infamous1]

Damn! is still there, see LV.

What about fuel? It is mostly CL on that areas.. SAme as on a stock map

Most likely noise then, I only said fueling (from AVCS change) because you thought it had went away.

[lepich]

I guess I'll take IC out to see what is going on there.. knock sensor is under IC, right?

It can't be tune, as I drove for 2 days, had 4-6 knock sum max for each 30 mins run, including WOT, all at idle. Next day I drive without laptop, and such LV as I posted above.

This is so must be a noise!

 

Heh.. I remember these worry-free days when I drive without datalog.. just used OTS St2 map and was happy :-) Hope my engine will forgive me for that.

[Spec B]

Is the knock/percieved knock knocking down your IAM?

I never tracked down the cause of my knock at the low load/low RPM range but since it is not affecting my IAM, I decided to live with it.

[lepich]

No, it is not affecting IAM. sometime it shows FBKC -8 , but never dropped IAM. All is happeing between 0.8 and 1.4 of load.

Are you saying I can ignore it? Even if it is real knock?

[Spec B]

Other than what you have done so far, I don't know if there is anyway to determine if it is real or not. I would ignore it.

 

Also, keep in mind that knock in that range is not something new. A lot of folks are experiencing it with no ill effects.

 

I think if you have:

 

- Pulled timing in that range

- Flashed back to stock

- Confirmed there is nothing loose

 

and the knock is still present, there is very little you can do other than ignore it.

[lepich]

- Confirmed there is nothing loose

This is what I have not done yet. Needs to be something!

 

One more strange things I've noticed today. Looks like RamRaider can provide meaningless values.. Can it be true? See attached screenshot, two parts of my log, same run.

On upper portion you'll see load 7.99 g/rev, which is impossible, on lower part Total Ignition Timing showed 2.5 without any knock or that huge corrections applied (-1.4 only, which should result in about 15 degrees of Total timing).. I do not have such numbers in a map as well..

Is this a bug?

Has anyone seen such crazy numbers in RR logs?

 

Edit:This is 0.5.1 version beta

[LittleBlueGT]

Is what I have not done yet. Needs to be something!

 

One more strange things I've noticed today. Looks like RamRaider can provide meaningless values.. Can it be true? See attached screenshot, two parts of my log, same run.

On upper portion you'll see load 7.99 g/rev, which is impossible, on lower part Total Ignition Timing showed 2.5 without any knock or that huge corrections applied (-1.4 only, which should result in about 15 degrees of Total timing).. I do not have such numbers in a map as well..

Is this a bug?

Has anyone seen such crazy numbers in RR logs?

 

Edit:This is 0.5.1 version beta

 

I get the odd random bit of info in RR. Like my rpm will go 3345, 7, 3389.

 

I wouldn't worry about it.

[NSFW]

It is an SSM parameter (not that I know what that means) and needs to be connected to the car to be available to log.

 

As I understand it:

 

SSM = Subaru Select Monitor. I think this might be the term for the device that Subaru designed (available through 3rd parties, and expensive) for dealers / mechanics to talk with the car. The term is also (or has also become) the name for the protocol those devices use, which is also the protocol that we use for datalogging.

 

SSM (the protocol) has standard parameters that are queried by index (they're listed at the top of RomRaider's logger.xml file) and extended parameters that are queried by their memory address in the ECU's RAM. I don't think the set of standard parameters has changed much, if at all, but the set of extended parameters gets bigger every time Merchgod releases an update.

 

The extended parameters don't show up in RomRaider until you connect to the car.

[NSFW]

If you want to get a really good picture of how much Knock Sum increments in your car, and what RPM/Load points, try the Sawmill program in this package:

http://www.codeplex.com/ssm

It's got a screen that records knock sum increments in a table form, across multiple drives. I've had it running on my CarPC for a couple months now.

 

Knock Sum increments a lot in the low-rpm, low-load region, even on the stock tune. Since going stage 2, I haven't changed my timing in that area at all, and it's the closed-loop area, so the fueling is the same. And it increments no more or less than it did when I was running the stock exhaust and stock tune. So I just ignore the 'knock' increments in that area.

 

How many people here have logged that during startup, idle/warmup, and on long boring (or short boring) drives? How much knock sum activity do you see in the area below 1.0 g/s and 2500 rpm?

 

I see quite a bit - yet I see very little activity in higher loads and RPM ranges. My guess is that this is either noise, or it's actual knock that Subaru's engineers decided wasn't worth mitigating.

[lepich]

Nice link, Thank you!

It is interesting to know when/how ECU decide when IAM should be dropped? I use to see -8 or even -12 FBKC, without any IAM drop. And, having less FBKC may cause IAM drop.. I guerss load is in consideration?

 

Edit: One day I'll have carPC. I can't imagine how convenient and nice it will be for a tuning! NSFW, do you remember how much total you've paid for your carPC? Considering I'll do all instaliation, only for parts..?

[LittleBlueGT]

How many people here have logged that during startup, idle/warmup, and on long boring (or short boring) drives? How much knock sum activity do you see in the area below 1.0 g/s and 2500 rpm?

 

If it is warm (and the car is warmed up from before) I can sometimes get no low end knock activity.

 

If it is cold I can sometimes get the counter to go through a whole cycle (from 0-32, or 35 or something and back up to 10 or 12) just driving out of the driveway.

[merchgod]

Nice link, Thank you!

It is interesting to know when/how ECU decide when IAM should be dropped? I use to see -8 or even -12 FBKC, without any IAM drop. And, having less FBKC may cause IAM drop.. I guerss load is in consideration?

FBKC will NEVER cause the IAM to drop - doesn't matter what it is. FBKC is a response to a knock event just like a drop in the IAM and a change to FLKC are responses, but only one of these are active at any given time. Conditions dictate which one is active (if any) at any given time (that is, which is appropriate given the conditions). Only extremes of applied FLKC have potentially a impact on the IAM as one of the conditions for changing the IAM (other conditions must also be met).

[NSFW]

Edit: One day I'll have carPC. I can't imagine how convenient and nice it will be for a tuning! NSFW, do you remember how much total you've paid for your carPC? Considering I'll do all instaliation, only for parts..?

 

$500 laptop (probably $350 today)

$400 screen (also lower now

$250 bezel fab (but you're going to DIY...)

$130 iGuidance software + GPS antenna

$40 120v-12v inverter

$35 power module

$35 materials for the shelf the laptop is mounted on

 

And some odds an ends... a pair of $20 USB hubs, a USB bluetooth antenna, a long VGA cable, some wiring, etc. Going to replace the inverter with an $80 12v laptop power supply at some point, it's sitting here ready to go in.

 

Some text and pics: http://www.legacygt.com/forums/showpost.php?p=1992851&postcount=2

 

LittleBlueGT: I'm glad to hear it's not just me.

[lepich]

^^ Hm.. Not bad, considering I already have an old laptop...

merchgod, Thank you for the explanation.

BTW, can non-secure engine cover cause a knock sensor to read a knock? I' think it may rattle, never sits securely. have lost all my pins which suppose to keep it in place.

[Spec B]

I don't think it will cause a false knock reading since it is "sealed" between the TMIC and the hood....but who knows. Try removing it and logging.

[LittleBlueGT]

FBKC will NEVER cause the IAM to drop - doesn't matter what it is. FBKC is a response to a knock event just like a drop in the IAM and a change to FLKC are responses, but only one of these are active at any given time. Conditions dictate which one is active (if any) at any given time (that is, which is appropriate given the conditions). Only extremes of applied FLKC have potentially a impact on the IAM as one of the conditions for changing the IAM (other conditions must also be met).

 

I am going to try and explain this in layman's terms. If I am wrong, and with tea cups I usually am, feel free to let me know.

 

FBKC will not cause your IAM to fall. That is a true statement, but it is also kinda true that it will cause it to fall from what I have seen.

 

If you log the knock sum parameter you will see anytime the ECU perceives what might be knock. You will also notice that it may go up one count in a WOT run, usually this will be accompanied by a FBKC event and the subsequent 1.4 or 2.0 (or whatever you have it set at) degrees of pulled timing. If you do a ton of logging, you will start to see the odd increase in the knock sum parameter (all WOT stuff here) accompanied by no FBKC, but by FLKC. This is ussually because there has been a patern of knock in that load/rpm cell. So instead of just pulling timing as a reaction, the ECU has decided to learn some pulled timing.

 

The same reasoning here also applies to the IAM (or in COBB speak DAM). I have yet to drop my IAM but I would imagine that if it were to drop during a WOT run it would be indirect response to a knock event (as seen by an increase in the knock sum paramter). This would obviously not be a one time thing, it would be in response to multiple knock counts that the ECU had previously either used FBKC or FLKC to react to, but this time (due to some logic that I don't fully understand, nor do I want to) decided to lower the DAM. I am assuming this would be due to some persistent knock at not just one or two areas, but in multiple areas there-by triggering the ECU logic to respond by dropping the whole DA table to a lower multiple.

 

So if you see FBKC a fair bit in your logs, you can likely expect your DAM to fall soon. FBKC is not directly responsible for the DAM falling, but it can be a fairly sure sign that will happen.

 

How is that for explaining something with more words then original explanation?

 

I actually think that explanation will help the relatively new tuners that already have a little bit of an idea of whats going on.

[merchgod]

FBKC will not cause your IAM to fall. That is a true statement, but it is also kinda true that it will cause it to fall from what I have seen.

 

If you log the knock sum parameter you will see anytime the ECU perceives what might be knock. You will also notice that it may go up one count in a WOT run, usually this will be accompanied by a FBKC event and the subsequent 1.4 or 2.0 (or whatever you have it set at) degrees of pulled timing. If you do a ton of logging, you will start to see the odd increase in the knock sum parameter (all WOT stuff here) accompanied by no FBKC, but by FLKC. This is ussually because there has been a patern of knock in that load/rpm cell. So instead of just pulling timing as a reaction, the ECU has decided to learn some pulled timing.

 

The same reasoning here also applies to the IAM (or in COBB speak DAM). I have yet to drop my IAM but I would imagine that if it were to drop during a WOT run it would be indirect response to a knock event (as seen by an increase in the knock sum paramter). This would obviously not be a one time thing, it would be in response to multiple knock counts that the ECU had previously either used FBKC or FLKC to react to, but this time (due to some logic that I don't fully understand, nor do I want to) decided to lower the DAM. I am assuming this would be due to some persistent knock at not just one or two areas, but in multiple areas there-by triggering the ECU logic to respond by dropping the whole DA table to a lower multiple.

 

So if you see FBKC a fair bit in your logs, you can likely expect your DAM to fall soon. FBKC is not directly responsible for the DAM falling, but it can be a fairly sure sign that will happen.

 

How is that for explaining something with more words then original explanation?

 

I actually think that explanation will help the relatively new tuners that already have a little bit of an idea of whats going on.

The knock sum is incremented during any non-consecutive knock event as perceived by the ECU. When there is a knock event, the ECU either responds to it or not.

 

So, if the ECU is going to respond to a knock event, it will, based on the current conditions, choose only one of the following: decrease in FBKC, decrease FLKC (current cell), OR decrease the IAM. Extremely simplified, FBKC would be chosen when the change in load is higher. When the change is load is relatively lower, then either the IAM or the FLKC (current cell) will be decreased. A decrease to the IAM will occur when there are extremes of negative FLKC applied in the current cell and when you are in hitting cells in the advance map that meet a minimum threshold. The current cell of the FLKC table will be decreased when the above is not true. There are other conditions, so this is simplified for all 3 of these. Also should be noted when the IAM changes, the entire FLKC table is cleared.

 

Bottom line, changes to FBKC, FLKC, and the IAM never occur at the same time. The ECU does nothing with negative FBKC - it does not store the value, it does not trigger the other knock responses. So, to say that it will impact the IAM or eventually impact the IAM is false. However, if you have knock events when FBKC is active, there's also a chance you will also have knock when changes to the IAM or changes to the FLKC table are active. Also, remember that a lot can happen in-between lines of your log and even in-between the sampling of each parameter on the same line, because logging is relatively slow.

[lepich]

Great explanation, thank you, guys! I do see some FLCK here and there, but usually it gets cleared after each 15-20 min drive.

BTW, if IAM was lowered, how fast it can be restored? Usually I did not log IAM, just cheking it in LeraningView after each logging session. It was always 1. Could it be lowered during a run, and then restored to 1 again, before I check it in LV?