Logged for the second time. Care to help me decipher?

[Grinder34]

https://docs.google.com/spreadsheet/ccc?key=0Aics16LC4qhLdFJnM0NCaFRNQnhfbWpnd09fdnNOdFE

Vermont showed me the ropes yesterday. I tried to absorb as much of his knowledge as possible, but theres just so much to learn!

Well anyways, i logged for the second half of my trip home today.

Right now I know that the two knock columns are pretty important, and the couple of -2.45 values for feedback knock concern me. Should i be worried? What else is in my log that I should be concerned with, notice, etc...?

also here's the 3rd gear pull we did together:

https://docs.google.com/spreadsheet/ccc?key=0Aics16LC4qhLdFNkckVPYUx1NFM5c2VNN2lvZFhTOHc

Thanks guys. PS, im not sure where this thread is most appropriate, so feel free to move/delete it.

 

[Vermont]

Moved over to the ask a tuner section man.

As for the log. On the first log there is one point where I do see real knock. Where it goes up to 2.4 then proceeds to move back down. Some might say that the amount of knock is not worth even noticing but in my book any thing under -1.4 is bad. Personally I would remove a degree of timing from that load and rpm set, and then smooth out the surrounding areas to ease into it. Also you seem to have quite a bit of boost error going on at the same time. Once again here I say anything over 1psi error needs to be fix. Right now it is not too bad and is under shooting but still it is something that should be tweaked. The biggest worry I have is the prevalence of FLKC on that whole log. Way to much of it for my liking.

My first suggestion for all of this to fix it would be to remove the timing from that cell like I said before and then go in and clear out those fucking timing added per-cylinder shits.... I know it is there to hopfully catch knock before it happens in the other cylinders, but this just means you have a higher chance to blow a ring land. After that, go out and drive around for little until the DAM (Dynamic Advance Multiplier) reaches one, should take a few minutes of regular driving with a few OL pulls. After that proceed to log for a good 5-10 min and try to get a much different driving as you possibly can. Vary throttle, loads, boost and gears so you can get a mix match of everything. What we are looking for is to see if the reducing of timing in that trouble cell, along with the clearing of the fucking stupid timing per cylinder tables makes a difference.

I would say wait to do this until either George or Td_d chime in with their opinion on this log.

 

[Td_d]

I've actually got to get to bed now, will have a thorough look tomorrow - but at a quick glance, the -1.4 timing at 6000 rpms plus and 2.45 load not a good idea - drop timing by 2* in that cell and smooth around it. In the log where you get knock to -2.45* - before you do anything with timing, my feeling is that you need to massage the wastegate tables a little, that's quite a significant boost error you pick up as you throttle up (up to 7% TD error).

One big piece of data missing - you must log AFRs - for all you know, the knock on throttle up is due to AFRs being too lean.

 

[Td_d]

Oh yes - and i agree - zero the per cylinder timing compensation tables - they are predominantly used to allow the engine to pick up knock quicker by pushing the knock threshold out to make the engine more knock capable in those cylinders.

On the DAM issue - personally I've always had it at 1 on start up, so that I don't have to wait, without any ill effects.

You still riding stock, btw?

 

[Vermont]

Oh yes - and i agree - zero the per cylinder timing compensation tables - they are predominantly used to allow the engine to pick up knock quicker by pushing the knock threshold out to make the engine more knock capable in those cylinders.

On the DAM issue - personally I've always had it at 1 on start up, so that I don't have to wait, without any ill effects.

You still riding stock, btw?

Currently he is Cobb Stage one. His AFR corrections are all within 5%, so we know that at the low end at least the MAF is scaled. He doesn't have a wide band yet, so we can't really log AFR. I was going to say leave the waste gate alone for now, at least until he is knock free. Right now he is just under boosting so that's not a huge deal. If it was over yeah, but not under. Also why drop the timing up top? Is -1.4 deg really not much under that kinda load? He is not knocking and not really putting that much pressure on the engine due to the low boost he is pushing right now. I figure so long as it doesn't knock it should be good. After all he is going to need to make power with timing primarily as he still has the stock DP on.

 

[Grinder34]

yeah cobb stage 1, no wideband (trying to get that figured out)!

So zero those cylinder 1-4 "Ignition Timing Compensation (Per Cylinder)" tables?

Right now they're all 0 except cylinder 4.

 

[HolyCrapItsFast]

I have nothing to add hear really... you are in capable hands... The only thing I would add is you need to log AFR's some how... who's to say really that removing timing is the correct thing to do here. You need to monitor fuel also. It could be lean for the condition being seen and though removing timing may solve the problem you could run your exhaust temps to high. Just a suggestion but this is what I mean about looking at timing and fuel together. :tup:

 

[Grinder34]

I cant monitor AFR, but i can monitor EGT (currently no logging, but i can look into that. how do you log a gauge?). Let me know what i should look for and I can do my best to get more info.

 

[HolyCrapItsFast]

I cant monitor AFR, but i can monitor EGT (currently no logging, but i can look into that. how do you log a gauge?). Let me know what i should look for and I can do my best to get more info.

Though it is not the most accurate thing you can log the A/F Sensor Ratio parameter. This is very accurate during closed loop and in open loop it is semi accurate. But it can give you/us an idea of where you are AFR wise. Ideally a stand alone AFR gauges is preferred.

I usually just recommend the AEM wide band. It is one of the simplest and accurate gauges out and it simply takes two wires and a DB9 connector. That runs to the laptop serial port COM1. If you you don't have a serial port on your PC then you can buy a USB adapter. The one that RadioShack sells is the one that works consistently.

 

[Vermont]

Though it is not the most accurate thing you can log the A/F Sensor Ratio parameter. This is very accurate during closed loop and in open loop it is semi accurate. But it can give you/us an idea of where you are AFR wise. Ideally a stand alone AFR gauges is preferred.

I usually just recommend the AEM wide band. It is one of the simplest and accurate gauges out and it simply takes two wires and a DB9 connector. That runs to the laptop serial port COM1. If you you don't have a serial port on your PC then you can buy a USB adapter. The one that RadioShack sells is the one that works consistently.

^ This! :tup:

If you need any help just let me know bro!

 

[Td_d]

Yup - not much to add - to reiterate Holy's point - it's key to log AFRs - like I said earlier, without out knowing what AFRs are sitting at, it's more difficult to diagnose the causality behind knock. There's always 3 ingredients that works together - fueling, timing and boost - changing one affects the other. There is a way to recalibrate the stock wideband settings to read relatively accurately down to 11.8 AFR - but frankly, just get a wideband, it's the best value for money you'll ever spend on the car.

Personally, my preference is no knock above 2 load, especially at high RPMs - but as Holy said, it's not necessarily clear what the cause is. I can live with a certain amount of low level knock low down - the engine load is generally not high enough to cause any real damage. But there was one instance of knock above 6000 rpms and 2.45 load in one of the logs if you look at sample times 187.54 to to 189.25 which was likely real knock as it also results in FLKC. Could have been lean AFRs though, as there is close to -2% learned fueling being carried all the way from range D onwards.

 

[Grinder34]

So theres lots of learning going on there, like the fueling throughout. Would it be a good or bad idea to reset the ECU and see what happens?

And barring a wideband, add the A/F parameter to the list and take another log before doing anything to the tune?

I hope to learn all these things myself, but for now I want to ensure I've got a safe tune while i start pouring over all the materials.

 

[HolyCrapItsFast]

So theres lots of learning going on there, like the fueling throughout. Would it be a good or bad idea to reset the ECU and see what happens?

And barring a wideband, add the A/F parameter to the list and take another log before doing anything to the tune?

I hope to learn all these things myself, but for now I want to ensure I've got a safe tune while i start pouring over all the materials.

I would do that, yes

 

[Grinder34]

Ok i reset the ECU last night and then logged to/from work. I didnt log the boring freeway/cruise-control portion of the commute. The logs are here:

https://docs.google.com/open?id=0Bycs16LC4qhLYzFyYU8tM1pzQk0

What i noticed is that especially on the "afternoon 1" i've got feedback knock of up to -3.85! I have a feeling that's bad.

I'm a bit confused about how long it takes to get learning. For instance, even by the afternoon, there was still no FKLC, but even first thing in the morning i had AF Learning. Also lots of AF Correction for most of the logs.

I'm working on getting a true AFR logger instead of the OBDII AFR values, but i'm not there yet.

Thanks for your help guys.

 

[Vermont]

Ok i reset the ECU last night and then logged to/from work. I didnt log the boring freeway/cruise-control portion of the commute. The logs are here:

https://docs.google.com/open?id=0Bycs16LC4qhLYzFyYU8tM1pzQk0

What i noticed is that especially on the "afternoon 1" i've got feedback knock of up to -3.85! I have a feeling that's bad.

I'm a bit confused about how long it takes to get learning. For instance, even by the afternoon, there was still no FKLC, but even first thing in the morning i had AF Learning. Also lots of AF Correction for most of the logs.

I'm working on getting a true AFR logger instead of the OBDII AFR values, but i'm not there yet.

Thanks for your help guys.

WOW! that is no bueno! Looks like under that one big knock event -3.85 you were over boosting and running lean. Something is not right with this map, at least as far as I am concerned. We need a wide band to be certain but what I can tell you for sure is that MAF scaling is out of whack. Your AF correction is what the ECU is seeing it needs at that minute to hit the right called for AFRs and AF learned is what the ECU has learned that it needs for a certain range on the MAF to meet AFR targets, any value over or under 5/-5 for either of them shows that you have a poorly scaled MAF, which is odd since you have the stock air box still on. Do you have a K&N filter in or something? How the knock works is it has to see knock in a certain load and rpm cell multiple times before it will add it too the FLKC table and remember to pull it. Just in that first log alone I am seeing a lot of things I am not happy with.

I am going to with hold my recommendation at this point until some one with more experience chimes in...

 

[Td_d]

Something's odd with that afternoon log - there's a big gap of time between the last point and the -3.85 flkc? Looks like tip out knock to me, as throttle is declining, but difficult to tell without the previous data points. You want to also log final fueling base (commanded fueling I think in the AP), so that you can check the gap between what is being requested and what the AFRs are (to rule out lean knock).

This is why I prefer the tactrix for logging, I can log 25 parameters at a much higher rate, without loss of resolution. Ideally, you would want both the open loop fueling target, and the final fueling base, to see if fueling is being richened up for load or tip in, and also total timing and open loop / close loop state. Often, if you mash throttle, the rapid change from closed loop to open loop does not allow the timing to drop (through interpolation between the cruise and non-cruise maps) quickly enough, resulting in timing that is too high for that resulting load in open loop. That's why you'll see in the stock maps, the cruise timing is much lower in certain low RPM areas of the base timing. But I digress :lol:

 

[Grinder34]

td_d. I'll add that to my logging parameter list. Are there any parameters i'm logging that I (really you), DONT need that I'm currently logging? I'll try to keep the list to a minimum to get a higher resolution.

ps commanded fuel map or commanded fuel final?

if you give me a list of what you'd specifically like logged i'll be happy to do that.

 

[Td_d]

Final fueling - that gives the actual requested fueling after all compensations have been added. Map gives you the corresponding value in the open loop fueling table. I'll check on your logs tomorrow, 1am here now and on my phone, could be 1 or two parameters that could be knocked off to increase resolution.

I'm curious what happened with the timing of that log though - there are time stamps which are non sequential all over the place.

 

[Grinder34]

oh haha they're all sorted by AF correction.

you should be able to sort by any column you choose. Maybe only I can, as the owner.

I'll leave them sorted by time as default

 

[Vermont]

Td_d is it possible for me to log his car with my taktrix fable even though he has a cobb map on? If so me and grinder could meet up again and do some logging and some small tweaks to the map to get him safe.