Who has demonstrably damaged/blown their engine due ONLY to tuning mistakes?

[mwiener2]

Depends on your definition of crappy. To alot of people, not maxed out = crappy tune.

 

The COBB OTS maps are rich and run moderate timing. They also come with a description of the mods allowed for each specific map. Alot of people assume some changes are minor enough that they won't hurt. Depending on the mod (intakes specifically) it may be more than enough to hurt the car.

 

What is the first mod most people do? Mess with the intake!!

 

 

Another thing to consider is that every car is different. Even car number 2 off the assembly line is going to be different than car 1 and 3. I see it on the dyno every day. I have 2 2005 LGT's in my shop right now. They both came in with basic stage 2 setups. Both had completely different dyno plots (and they're both gonna leave with much larger much different plots )

 

Because there are so many differences, an OTS map has to be as basic as possible. Again, this = crappy to some people.

 

An excellent point about self tuners and e-tunes.. they have a higher failure rate than OTS and pro-tunes. Most self-tuners don't know enough to do it right themselves no matter what they've read on the internets. You're also not a good tuner unless you've blown up an engine or two... otherwise you wouldn't be learning. If you're a self-tuner, that means your own engine is gonna go sooner or later, it's part of the process.

 

E-tunes are a joke, I don't care what anyone says. One of the reasons most self tunes fail is because the proper monitoring equipment is not in place. The E-tune is worse because not only are those monitors not in place, but it's not even happening in real time.

[Boostin1657615274]

The cobb 93 octane map advances total timing (primary + max Dynamic advance) an average of perhaps 3 degrees in high load areas up to about 5000 rpm. The fuel maps are basically stock. It leans out a few areas slightly; there are hardly any meaningful differences.

 

There's no danger of a lean condition with a Cobb map if you use it with the parts it was intended for. None. You can smell fuel when you boost it! I have actually leaned out the Cobb stage 2. As for the timing, I've seen minor FLKC (2 degrees max). Big deal. You can't live by knock sensors alone, just like you can't live by widebands. They are only instruments. The Cobb map is perfectly safe if you use it the way it was intended to be used. You may or may not be happy with the power output, but I wouldn't worry about engine damage. I have actually leaned out the Cobb map some.

 

I have attached screenshots of the stage 1 and 2 93 octane maps, as well as the stock fuel and timing maps. I have also attached a spreadsheet (it is actually in .xls format) with three tabs that compare the timing of the stock, stage 1, and stage 2 map. Finally, I am including a screenshot of the leaned out Cobb stage 2 map I have used successfully. That's used with Agency power catback, STi uppipe, stock downpipe with front precat gutted.

Cobb_timing_comparison_2005_LGT.csv

[LittleBlueGT]

I agree with Boostin, but will add a bit:

 

I buddies 09 WRX has boost creep (it is worse with a new intake, that we scaled properly) but in anything less then 80F temps it creeps, and that causes knock at high rpm. In the 40F temps it creeps soo much that even at 0% WGDC and 10.8:1 AFR I had to pull 4 degrees of timing off of COBBs map for it to work.

 

FWIW he does have a HFC in the DP.

 

That is not so much an issue with COBB's map, but an issue with it not working as a package as advertised safely. (If he didn't have a TBE, and only a DP, maybe it wouldn't creep)

 

The issues I have had with COBBs map (and Christian from COBB has replied to my e-mails, but so far TTBOMK has done nothing) is they turn off FLKC at too low of a load:

 

2009 WRX they turn it off at 3.1 load, and his car easily hits 3.3 or 3.4 load. Now any knock that is seen above 3.1 load will only be reacted to by FBKC, and never be learned or affect DAM (IAM)

 

2005 LGT vs#1 AP tops FLKC at 2.2 load IIRC (not sure what it is on any updated vs#2 maps). Same problems as above, as a stage2 LGT will easily hit over 2.2 load.

[NSFW]

I think I've seen a roughly equal number of dead engines tuned by pros vs. DIYers. However there are probably a lot more cars out there tuned by pros vs. DIYers, so you could be right about the ratio.

 

It worries me to see posts on NASIOC where people ask where to find an open source tune for their mods. I think there's a growing expectation that open source tuning means going to a web site, finding an open source tune, flashing it, and being happy. That's going to end badly as often as not. Pretty much every tune I've seen at RomRaider has been something like "here's a tune that works well on my car, it may be a good starting point for your car if you have similar mods." Probably they are good starting points, too. But I wonder how many people will just flash them and hope for the best.

[mwiener2]

There's reasons they(COBB) do that. At those higher loads, there's alot of noise going on. The knock sensor hears all this "power" and interprets it as knock, even though it may not be. This doesn't happen all the time, but enough that they worry about it. They are, after all, selling a product. If everyone's car started running like crap cause the IAM dropped due to false knock, there'd be alot of cranky people. If 2-3 motors a year pop cause on their particular cars something wasn't quite right, that's better odds.

 

 

Your buddies WRX sounds like it's got something else going on. Is it mostly stock or he's got a huge turbo?

[LittleBlueGT]

There's reasons they do that. At those higher loads, there's alot of noise going on. The knock sensor hears all this "power" and interprets it as knock, even though it may not be. This doesn't happen all the time, but enough that they worry about it. They are, after all, selling a product. If everyone's car started running like crap cause the IAM dropped due to false knock, there'd be alot of cranky people. If 2-3 motors a year pop cause on their particular cars something wasn't quite right, that's better odds.

 

 

Your buddies WRX sounds like it's got something else going on. Is it mostly stock or he's got a huge turbo?

 

I completely disagree with your assessment as to why COBB turns off FLKC at those loads. COBB hasn't read the stickies on RR, and they don't understand completely how it works.

 

FWIW there are many cars (yours likely too) that allow the knock sensor to affect learned timing at higher loads then the ones mentioned, especially the LGT at 2.2 load, good grief!

 

Here is a link to a thread about my buddies WRX:

 

VF52 major boost dreep!!!

 

There is nothing wrong with his car, the turbo creeps, plain and simple. Actuator is good, no leaks in the hoses going to actuator, we checked everything.

[mwiener2]

my ECU sees loads of 4+. not kidding

 

Rumor has it COBB has actually hired the Open Source Gurus. I think COBB has a pretty good understanding of the Subaru ECU.

[LittleBlueGT]

my ECU sees loads of 4+. not kidding

 

Rumor has it COBB has actually hired the Open Source Gurus. I think COBB has a pretty good understanding of the Subaru ECU.

 

I believe your car sees that load, now tell us what your tuner set FLKC load range at? Still on stock sensor? I thought so.

 

The rumor is true, Merchgod from RR is working for them on a trail basis right now part-time. We will see what happens in the future.

 

The reason COBB left those tables in tact is because they did not bother to change them from the OEM settings.

 

Stop posting stuff you don't understand Weiner!

 

If the FLKC turns off at a certain load, and the car knocks at loads above that (which mine did on stage1, we have super crappy gas) the ECU is handicapped with only having FBKC to use to control the knock. It will never learn anything there, and if the car does knock at that load, it will do it everytime, and the whole knock learning feature that the ECU has is rendered useless.

 

I am not saying that COBB's tunes are bad. AFRs were good, boost control was adequate, timing conservative on every COBB tune I have logged. But it is not perfect, and I just pointed out a couple of flaws.

[mwiener2]

That's why they turn it off... so it doesn't learn false knock.

If it hears something and pulls a degree to be safe, great. If it was real knock, it's trying to protect itself. If it was false knock, then better safe than sorry.

 

The reason it doesn't learn is that everything is assuming a good tune. It shouldn't be knocking in the first place. If it's a bad tank of gas, don't store long term changes for short term problem.

[LittleBlueGT]

That's why they turn it off... so it doesn't learn false knock.

If it hears something and pulls a degree to be safe, great. If it was real knock, it's trying to protect itself. If it was false knock, then better safe than sorry.

 

The reason it doesn't learn is that everything is assuming a good tune. It shouldn't be knocking in the first place. If it's a bad tank of gas, don't store long term changes for short term problem.

 

 

I give up, you are an idiot:

[mwiener2]

If I'm the idiot what's YOUR explanation for turning it off?

[Guest]

Stop posting stuff you don't understand Weiner!

 

That would a be a good start.

[NSFW]

That's why they turn it off... so it doesn't learn false knock.

 

I don't think that's it. My load gets up to 3.6 or so, and I have FLKC active up to 4.0, with no false knock issues so far.

 

If it hears something and pulls a degree to be safe, great. If it was real knock, it's trying to protect itself. If it was false knock, then better safe than sorry.

Funny that you put it that way, it seems to me that "better safe than sorry" would mean pulling timing and keeping it pulled until the knock sensor quiets down. Which is basically how FLKC works.

 

The reason it doesn't learn is that everything is assuming a good tune. It shouldn't be knocking in the first place. If it's a bad tank of gas, don't store long term changes for short term problem.

Seems to me that the "tank of bad gas" is exactly what learned knock correction is for - keeping knock down until the octane goes back up. It will un-learn the knock corrections too you know.

[mwiener2]

But it would take another 3-4 tanks to unlearn 1 bad tank. Have you fine tuned your map?

[Boostin1657615274]

When I was out tuning with AccessTuner, a few WOT 2nd and 3rd gear pulls were done and no "knock events" (I'm unsure how that translates to open source lingo) came up during the WOT. But I saw a couple from free revving? And I think I had FLKC scaled to a little over 2.0 load, not up to 2.5 which is what I was getting as my max.

 

I find no fault in wanting an aggressive knock control strategy as a personal preference. Is it necessary to prevent engine damage if the stage 2 map is used as intended/marketed? I doubt it. I think Cobb made a conscious decision to keep the FLKC where it is, just like they made a conscious decision to keep the dynamic advance table stock. You may have very good reasons not to agree with them, but that doesn't mean Cobb doesn't understand the ECU or that people with advertised "stage 2" parts and a "stage 2" map are going to pop their motor.

 

Remember that such heavy use of "stage" terminology is mostly a Subaru deal. I'm not sure if Cobb's marketing made this so prevalent. I do agree that it promotes complacency with a tune: if I'm "stage 2" and this map is "stage 2" then I'll flash it in there and be fine. You do find that mentality among other turbo car owners but not to the same degree as Subaru owners.

[JoeFromPA]

I think the things said here are worth noting. Cobb Stage 2 map leaves power on the table, and is by no means an ideal map, but it is an incredibly easy and time-efficient manner to gain a solid 20% HP over stock with a gutted or aftermarket DP (and gutted or aftermarket uppipe for 05-06 LGTs) with a very low risk of problems.

 

I have absolutely no problems with open source, e-tuning, or pro-tuners. My point is simply that there is something to be said for the "crappy OTS Cobb tunes".

 

Joe

[LittleBlueGT]

When I was out tuning with AccessTuner, a few WOT 2nd and 3rd gear pulls were done and no "knock events" (I'm unsure how that translates to open source lingo) came up during the WOT. But I saw a couple from free revving? And I think I had FLKC scaled to a little over 2.0 load, not up to 2.5 which is what I was getting as my max.

 

I find no fault in wanting an aggressive knock control strategy as a personal preference. Is it necessary to prevent engine damage if the stage 2 map is used as intended/marketed? I doubt it. I think Cobb made a conscious decision to keep the FLKC where it is, just like they made a conscious decision to keep the dynamic advance table stock. You may have very good reasons not to agree with them, but that doesn't mean Cobb doesn't understand the ECU or that people with advertised "stage 2" parts and a "stage 2" map are going to pop their motor.

 

Remember that such heavy use of "stage" terminology is mostly a Subaru deal. I'm not sure if Cobb's marketing made this so prevalent. I do agree that it promotes complacency with a tune: if I'm "stage 2" and this map is "stage 2" then I'll flash it in there and be fine. You do find that mentality among other turbo car owners but not to the same degree as Subaru owners.

 

There are a number of reasons why COBB doesn't understand it properly IMO.

 

1) I have spoken with Justin, and Christian at length about the issue, and they don't seem to get it. It is easy to see by there descriptions of the FLKC in their ATR manual, they give the correct tip, but only for engine speed (which they change to suit their rev limits), but then don't even mention the load range!!!!!!!!

 

From COBB's manual:

 

Knock Learning

Knock Detection Range (High, Low)

Table description – two vales for engine speed that represent the low or high range for active knock detection.

Tuning Tips – The range for knock detection should span the entire operating range of engine speeds.

2) If you are driving your car hard, say on a track, and your car was knocking at say 2.5 load/4000 rpm. The ECU would only respond with FBKC. That means in a few laps your motor will see hundred of hard knock events that the ECU will never learn from.

 

3) The guys at COBB do not seem to get the WHOLE interaction of knock control. They don't get the FLKC (once ECU is out of rough learning mode) is what will trigger the IAM to fall in time.

 

4) OS tuning had the whole per/cyl timing comp thing figured out well in advance of COBB. They do not undersatdn everything about the ECU

 

This does not mean their tunes knock badly, they don't. On my car it did, and now that I know how to read logs, I can see that my car it was knocking at high load (even on stage1 map) over and over and over, but it never learned anything. You can tune that way, you see FBKC, and then mod the tune accordingly. But in the off chance that your car is knocking in that range, and you are not logging to monitor, it can and will cause engine damage.

 

The bad thing is many tuners use their maps as a base, then don't change that table (I have seen it many times).

 

I am not guessing here, or speculating.

[LittleBlueGT]

No more speculating:

 

Just got off the phone with Christian at COBB. He seemed to kind of understand most things about how FLKC and rough correction fit in. He said it was a matter of time and resources. (working on 80+ other Honda maps right now)

 

He is releasing new maps within weeks (more launch control stuff) and he said he will make changes across the board (if he has time) to coarse learning load and Fine Learning load ranges to encompass the entire operating range for the engine (we figured about 3.6 load should be enough for any OTS tune, even in cold temps).

 

Good guys there at COBB (Christian has always been helpful, but is only human, and can only do so much).

 

PS: he also hopes they will define (in newer roms) the target boost IAT comp table (which drastically cuts boost in really cold temps) as only the OS community has this table currently defined.

 

 

 

Weiner: shut-up!

[Boostin1657615274]

have you read their article on Dynamic Advance? http://www.cobbtuning.com/info/?ID=3487

 

you are talking about the load axis on this table (pic taken from their article)? http://www.cobbtuning.com/images_products/DynamicAdvance_4.jpg

[LittleBlueGT]

have you read their article on Dynamic Advance? http://www.cobbtuning.com/info/?ID=3487

 

you are talking about the load axis on this table (pic taken from their article)? http://www.cobbtuning.com/images_products/DynamicAdvance_4.jpg

 

 

We are way beyond the elementary info in that article. It has much good info, but does not paint the whole picture.

 

Read the knock sticky on the RR forum, much more in depth and accurate:

 

Subaru's knock control startegy explained

 

That is the FLKC table.

 

I am referring to this tables:

 

http://i248.photobucket.com/albums/gg194/littlebluegt/FLKCload.jpg

 

This from an 09 WRX. OEM and COBB has the table maxed out at 3.1 load, which does not encompass the operating range of the motor running more boost.

 

I get a little annoyed when I post info that I know about (not asking questions) to try and shed more light on subjects that I had previously not fully understood, and then people try and refute your reasoning with incompetent logic. Hence the phone call to Christian at COBB, and hence why COBB is changing the maps.

 

I am trying to help, and I fell I have to fight it every step of the way.

[Boostin1657615274]

I have read the sticky and I have a pretty good understanding of how the system works, but I am always trying to learn more. It is simply a matter of mapping Cobb's nomenclature to Romraider's, that's all.

 

Don't take this the wrong way, but please chill with the whole Us (Rom Raider etc) vs Them (Cobb) deal. It's not really helpful. It reminds me of some Honda D series people (most 90s Civic engines) attitudes toward those with twin cam B series engines (Integra). Cobb's hardware allows realtime tuning and an external display. To some people, that's worth the extra hundreds of dollars, despite having to deal with software that has some limitations and uses slightly different terminology.

[NSFW]

But it would take another 3-4 tanks to unlearn 1 bad tank. Have you fine tuned your map?

 

I'm pretty sure you could unlearn that in less than one tank if you drove hard (and if you're not driving hard, you probably don't care much). I haven't tested that theory myself but I think I've seen enough evidence.

 

When I was running the stock tune I saw small FLKC values in different cells at different times, indicating that it was learning and unlearning the corrections. A while back someone posted a log on RR where you could see it unlearning a single FLKC cell over the course of a single log. If you really wanted to unlearn a bunch of correction it might help to some 1/2 to 3/4 throttle pulls to cover more of the FLKC table but I'm sure it's possible in well under a tank if you tried.

 

Resetting the ECU may or may not be a special case, but IAM goes from 0.5 to 1.0 as soon as I find one freeway onramp.

 

The only part of my timing table that isn't fine-tuned yet is the area where it goes rich during and after spoolup. That'll happen after I go to a blow-through MAF to fix the rich dip. Hopefully by Christmas.

[LittleBlueGT]

I have read the sticky and I have a pretty good understanding of how the system works, but I am always trying to learn more. It is simply a matter of mapping Cobb's nomenclature to Romraider's, that's all.

 

Don't take this the wrong way, but please chill with the whole Us (Rom Raider etc) vs Them (Cobb) deal. It's not really helpful. It reminds me of some Honda D series people (most 90s Civic engines) attitudes toward those with twin cam B series engines (Integra). Cobb's hardware allows realtime tuning and an external display. To some people, that's worth the extra hundreds of dollars, despite having to deal with software that has some limitations and uses slightly different terminology.

 

I use COBB ST AD, and I have a few ATRs on my laptop as we speak.

 

I have spoken to Christian many times about tuning issues, and I don't at all feel that it is us vs them. If I come across that way, it is not intentional.

 

I use both OS and real-time tuning on my car, and others. Thus I know first hand the advantages and disadvantages.

 

I posted valid info about what COBB did not do with their OTS tunes. Some here disputed that. I knew there was no "good reason" why COBB didn't do it, other then they haven't thought of everything (and nobody has). I felt it was necessary to get the truth out. I already knew COBBs stance from other e-mails, but some didn't believe me. Hence the phone call today to Christian.

 

Is this important info? I feel it is. I lost a motor a few years ago that was running on COBB OTS tune, then on a vendor's tune. I was logging, but didn't know how Subaru's knock control strategy worked. When I looked back at logs (just showing DA) I can now see that my car was over the FLKC load range, and was knocking. It was just using FBKC to deal with it, and it knocked on every single WOT run I did. Of course no learning took place, since the knock I had was occurring at higher loads then 2.2.

 

Would this have saved my motor? I don't know, but it sure would have helped.

[LittleBlueGT]

I'm pretty sure you could unlearn that in less than one tank if you drove hard (and if you're not driving hard, you probably don't care much). I haven't tested that theory myself but I think I've seen enough evidence.

 

I am sure you know this, but how fast the car learns to add timing back in, is something that we can alter in the tune, if desired. I believe the newer models seem to add it back faster.

 

At any rate, if you have learned timing in a particular load/rpm cel, all that is needed to learn it back is a min time spent in that cel again. For each min time spent w/o knock the ECU will add back a set amount (common amounts subtracted are -1.4 degrees, and +.35 degrees added back at a time). I find that it takes very little time to learn back mid-rpm learned knock, but high rpm learned knock takes a bit longer, due to not spending much time there, unless you are going hard in a high gear at high rpm. (not much time is spent at 6000 rpm in a 1st or 2nd gear pull to meet the min time)

[JoeFromPA]

LBGT - Did I read you correctly that Cobb is going to be adjusting their LGT maps in, hopefully, the next few months? Or perhaps all Subaru maps?

 

Joe