Why does the IAM value fluctuate?

[BMCracer]

I just got my Nexus 7 / BtSsm running to the point I can use it in the car. I've read lots of places that the IAM value should be 1.0 when everything is happy. Mine is consistently around the 0.15 - 0.4 range, and has been as low as 0.012, and tonight for a very short period was the first time I have ever seen it at 1.0

 

Car is a 3.0 5EAT 2007 ADM facelift Spec B wagon with 150,000km ( 90,000 miles ). The car runs fine, I maintain it and it's never given me a problem, but wondering why the value isn't 1.0, why it keeps fluctuating, and what it actually means / does.

[DaveWaters]

Looks like you have knock issue. Monitor feedback knock and fine learned knock to perhaps give you an idea when it’s occurring. The IAM is global timing pull because the ecu is detecting a lot of knock. In your case it could be as simple as bad gas/ too low of octane or it could be vacuum leak etc. Cobb has good info to help you understand the subaru knock strategy on their website . If I get a chance later on I’ll see if I can find you a link.

[dgoodhue]

What octane gas do you use?

[BMCracer]

I have run 95 petrol in it most of the time I have owned it, but recently changed to 98. It does feel like it goes slightly better on 98, but it has not been tuned for anything, as far as I am aware it is the standard factory tune.

 

Here are some LV tables from some knock events - this is when it was running on 95 ( so should be worse than 98? ). I would have expected to see more negative vales in the table?

 

BtSsm_LV_20190227_2212 by Andrew Metford, on Flickr

 

BtSsm_LV_20190227_1641 by Andrew Metford, on Flickr

 

BtSsm_LV_20190227_1632 by Andrew Metford, on Flickr

 

BtSsm_LV_20190227_1519 by Andrew Metford, on Flickr

 

In this next pic, I'm doing 100kph ( 60 mph ) down the motorway with light accelerator yet it's picked up knock?

 

20190226_173053 by Andrew Metford, on Flickr

[moral hazard]

What gear were you in during those knock events, could it be possibly that you're in a 5th at a high load? Possibly could be avoided by downshifting more often and not lugging the engine. Just a thought.

[JmP6889928]

Too low of RPM at load will definitely cause knock.

[BMCracer]

5eat automatic gearbox. So I don't know what gear it was in. The bottom one was 5th cos it was doing 60 mph. I always run the car in "I" mode, and I never hammer the gas - I drive it like a nana, no hard acceleration.

[Andy_Mac]

Do any of you know a sequence/method to get it out of rough knock correction faster/straight away?

My speculation was that it is getting some moderate knock straight after a reset causing it to straight away pull the IAM rather than adding fine corrections that can be easily tuned out.

[DaveWaters]

Maybe something is causing a false knock. Check under the hood for loose things. Loose heat shield on exhaust, something loose near knock sensor.

[whitexc]

5eat automatic gearbox. So I don't know what gear it was in. The bottom one was 5th cos it was doing 60 mph. I always run the car in "I" mode, and I never hammer the gas - I drive it like a nana, no hard acceleration.

 

I did some experimenting with my '09 LGT 5EAT when I got it and it seems to get better mileage in Sport mode than I mode, unless you are the the highway with the CC set. Just wondering if you switched Sport mode and remained light footed if the readings would change. All that does is introduce boost earlier from my understanding so you won't be lugging the engine as much. I mode feels as though it cripples these cars. Just my observations and I am new to these cars.

[BMCracer]

AndyMac above told me how to reset the ecu too which I did through the BtSsm, and that hasn't made any difference to anything.

 

I can't hear anything that sound like it's loose or rattling etc, but the cars are pretty well insulated so maybe I need to drive with the window down a bit to see if I can hear anything.

 

I might try it in S mode for a bit and see if anything changes. Won't make a difference to getting on boost earlier cos I have the 3L 6 cylinder NA motor.

 

But even if that takes care of the knock issues, why is the IAM all over the place, and so low??

[DaveWaters]

IAM is timing pulled across the board to protect the motor from damage due to higher than normal knock conditions as registered by flkc and fbkc. Once knock issues are resolved IAM will work its way back up to 1.

 

You may not hear any sort of rattles. You’ll need to get under the hood and inspect. There could be a bolt that went missing sitting on top of the block near the knock sensor or even a wire loom or vac hose laying across it. Get your hands in there and feel around for loose stuff and fasten. Zip ties are your friend!

 

From what I’ve been told the 5eat itself tends to register some knock during shifts. But not enough to drop IAM.

[Andy_Mac]

The hard part in this case is that it it does it before entering fine correction mode so it isn't actually pulling the IAM due to high knock counts visible in FKC or FLKC.

I've been through his logs and there isn't a whole lot of FKC events at all, most are the typical 0.7 and 1.41 that you get from general driving and aren't enough to ever pull the IAM this low consistently.

 

http://www.romraider.com/forum/topic1840.html

[BMCracer]

From this afternoon, now running on 98 petrol ( all the previous ones posted were on 95 )

 

BtSsm_LV_20190328_1514 by Andrew Metford, on Flickr

 

Hopefully I'll get some time on the weekend to have a look around the engine bay for anything loose.

[DaveWaters]

Is it just me or should there be values on the maf g/s line?

[Infosecdad]

That is a lot of knock detection at low RPMs and with the IAM already down to .656

 

Really curious what's up with your engine.

[solidxsnake]

The hard part in this case is that it it does it before entering fine correction mode so it isn't actually pulling the IAM due to high knock counts visible in FKC or FLKC.

I've been through his logs and there isn't a whole lot of FKC events at all, most are the typical 0.7 and 1.41 that you get from general driving and aren't enough to ever pull the IAM this low consistently.

 

http://www.romraider.com/forum/topic1840.html

 

IAM is never pulled because of "high knock counts visible in FKC or FLKC." Every time the IAM changes at all (either increases or decreases), the entire FLKC table is zeroed out. FKC only registers values in real-time, and only when the ECU is not in fine or coarse learning mode. For example, if the ECU is in fine-learning mode and registers a knock event, instead of pull 2° in feedback response (i.e. -2° FKC), it will just subtract another 2° from the current FLKC value. As such, you can't just look at FKC or FBKC and correlate it to IAM.

 

If that's the case, I'd suspect the timing correction (i.e. the IAM drop) is actually helping the problem, and that it's not false knock. Since that's the case, I'd suspect a mechanical issue, or a failing MAF or O2 sensor, since the car is stock.

 

To answer your question, as far as I know there isn't any way to get the ECU out of coarse-correction mode after a reset besides some varied driving (low load, high load, etc.).

[BMCracer]

So apart from replacing them, is there a way to test the MAF and O2 sensors? Can they be tested in situ?

 

I don't think ( hope not ) it's a mechanical issue as the car runs very well, and if not for having the Nexus 7 / BtSsm I would have no indication it is knocking all the time.

[solidxsnake]

On second thought, I forgot that the NZ spec.B is a 3.0 NA H6 motor, and not the turbo'd 2.5 H4. Which brings me to remember some of covertrussian's findings about the stock NA tunes being very aggressive and knocking quite a bit from the factory. Since the car feels the same/fine to you, it's likely the knock you're seeing is "normal" and not much to worry about if the car feels the same as it always has. However, it's possible you could get more power and better fuel efficiency with a proper tune.

 

Remember that everything you see about how "IAM NEEDS TO BE AT 1.0 OR YOUR MOTOR IS UNHAPPY AND WILL BLOW UP" is referring to the turbo'd motors, where some minor detonation at WOT will shatter ringlands. Not saying knock is good on an NA motor, but it's definitely more tolerable than on the turbo'd motors.

[xt2005bonbon]

solidxsnake, can you clarify this statement: "IAM is never pulled because of "high knock counts visible in FKC or FLKC." ?

I agree with the first part, but the second part not really. For example, let's suppose your latest LV shows a tons of FLKC registered in various cells (with especially large values), while your IAM is at 1.00. You continue driving and the ECU registers more FKC on top of the FLKC, then at one point it will say 'enough is enough' . And it would eventually zero out the table (as you say) and decrease the IAM (implying pulling a certain amount of timing across the board). So in essence, the amount of registered FLKC appear to play a role in whether or not the IAM will decrease. Am I missing something?

[solidxsnake]

solidxsnake, can you clarify this statement: "IAM is never pulled because of "high knock counts visible in FKC or FLKC." ?

I agree with the first part, but the second part not really. For example, let's suppose your latest LV shows a tons of FLKC registered in various cells (with especially large values), while your IAM is at 1.00. You continue driving and the ECU registers more FKC on top of the FLKC, then at one point it will say 'enough is enough' . And it would eventually zero out the table (as you say) and decrease the IAM (implying pulling a certain amount of timing across the board). So in essence, the amount of registered FLKC appear to play a role in whether or not the IAM will decrease. Am I missing something?

 

 

Yes, I suppose I was a bit too broad in my description there. You're correct that in order to switch away from fine-learning mode, the applied FLKC value must be greater than a certain amount. However, that threshold is a fixed number (by default I believe it is 4deg). Meaning that if your car has an FLKC value of -4 in a cell that occurs in the area rough correction can be activated, the car may switch over to rough correction. That doesn't necessarily mean it will, however. You could have a FLKC value of -8 indefinitely and the IAM still won't drop.

 

Once the IAM changes, the FLKC table will be set to zero, and will stay that way until the ECU has determined the IAM has settled.

 

All of this from here: http://www.romraider.com/forum/viewtopic.php?f=40&t=1840

 

 

The important take away from that post:

 

 

Q: I've logged a large negative value in FBKC, but the IAM and FLKC did not change? Why didn't the ECU change the IAM based on this severe knock?

 

It is better to think of following as three distinct means of knock control:

 

1. Negative FBKC

 

2. Changes to the current IAM.

 

3. Changes to the current FLKC cell (note this is different than applied FLKC).

 

That is, only one is active at any given time based on what would be appropriate for the given conditions. So, in order for the IAM to change, the "severe" knock would have to occur when the ECU is in rough correction mode, FBKC is disabled, and the conditions to make a change to the IAM are met. So, basically, the corrective action that is taken is dependent on which knock control method is active at the time and whether certain conditions were met (although some values, such as FLKC, may partially impact mode selection as described earlier).

[Volta55]

I did some experimenting with my '09 LGT 5EAT when I got it and it seems to get better mileage in Sport mode than I mode, unless you are the the highway with the CC set. Just wondering if you switched Sport mode and remained light footed if the readings would change. All that does is introduce boost earlier from my understanding so you won't be lugging the engine as much. I mode feels as though it cripples these cars. Just my observations and I am new to these cars.

 

Same.. in my 05, 80% of the time I drive with the +/- or Sport mode... nothing fully auto. Full auto loves to downshift at 2.5K when your trying to friggen merge onto i95.. slow boi

 

Is it just me or should there be values on the maf g/s line?

 

I was thinking the same thing

 

 

For the OP.. I would try disconnecting the battery for 15 minutes, pump the brakes, then connect the battery. ( hard reset )

And check for knicked wires going to MAF and o2