Tip in knock

[nikitopo]

Originally Posted by steve99

From my experience its very difficult to hear knock on these cars unless its rearly bad ...

The only time ive heard knock is usually on stock tuned cars

Try adding some octane booster or try 5 litres of e85 in full tank and see if it makes any difference to noise or logs

Steve, I don't think the issue is always the fuel quality. In Germany we have very high quality. When I had a similar issue in my Forester, I couldn't fix it even with V-power racing which is rated at 100 RON. The knock was for sure there, the engine was gutless and I could see it even in the logs. I believe that some engines are not so tolerant to changes and maybe the table definitions were not a 100% accurate. I don't blame my tuner (throttlehappy) which in fact was very good and I could also verify myself that the parameter changes were minimal and careful made. Nevertheless, I am a programmer and I can understand the whole process of tuning and its pitfalls.

I think that Tor is on his own and depending the cause of the problem he might solve the issue or not. Personally, I wouldn't take the risk and I would flash immediately the tune back to stock. I guess everyone is different.

[Tor]

The AC is off most of the time I drive the car. I'm using 100 octane fuel and have no other knock worth mentioning. The stock tune has tip in knock too. I'm not going back to stock because of (potential) knock in low load transients and only in one specific circumstance. This is not something that really worries me, more it annoys me.

Also I'm not too sure how much I trust the logs in this specific scenario, as MotoEast wrote that it doesn't always show up in the logs. I could imagine that the rapid change could make it difficult for the ECU to react and the logged values (if any) may not correspond with the actual severity. Of course that personal imagination/theory.

Originally Posted by steve99

Also have a look at the transient ignition retard table from the stock K00C 2017 rom, much more agressive than the older roms, note the table size is different

Thanks for the tip, Steve. Left is A01G, middle MY17 and right is what I might try out. Was it ever settled for sure if the offset should be -30 or -50? The definitions I have, say -50 for both the old and new ROMs.

[steve99]

Originally Posted by nikitopo

Steve, I don't think the issue is always the fuel quality. In Germany we have very high quality. When I had a similar issue in my Forester, I couldn't fix it even with V-power racing which is rated at 100 RON. The knock was for sure there, the engine was gutless and I could see it even in the logs. I believe that some engines are not so tolerant to changes and maybe the table definitions were not a 100% accurate. I don't blame my tuner (throttlehappy) which in fact was very good and I could also verify myself that the parameter changes were minimal and careful made. Nevertheless, I am a programmer and I can understand the whole process of tuning and its pitfalls.

I think that Tor is on his own and depending the cause of the problem he might solve the issue or not. Personally, I wouldn't take the risk and I would flash immediately the tune back to stock. I guess everyone is different.

haha,


my reasoning here is to eliminate fuel quality, if you add octane booster or better fuel then theoretically knock should reduce. If it doesn't reduce at all then the cause may not be knock but a mechanical problem like exhaust hitting on chassis fooling knock sensor faulty or loose knock sensor or other mechanical noise being interpreted as knock :-)


or you have some other major tune issue or hardware issue causing for example a very lean condition.


for example incorrect scaling of 02 sensor. can cause you to run incorrect AFR but logging will show correct afr, always best to check with another independent sensor if weird stuff happening.


Flashing to stock is always a good check to do to eliminate hardware/tune issue, however not always possible if you have changed components that require "tune" or running say E85.

[steve99]

Originally Posted by Tor

The AC is off most of the time I drive the car. I'm using 100 octane fuel and have no other knock worth mentioning. The stock tune has tip in knock too. I'm not going back to stock because of (potential) knock in low load transients and only in one specific circumstance. This is not something that really worries me, more it annoys me.

Also I'm not too sure how much I trust the logs in this specific scenario, as MotoEast wrote that it doesn't always show up in the logs. I could imagine that the rapid change could make it difficult for the ECU to react and the logged values (if any) may not correspond with the actual severity. Of course that personal imagination/theory.



Thanks for the tip, Steve. Left is A01G, middle MY17 and right is what I might try out. Was it ever settled for sure if the offset should be -30 or -50? The definitions I have, say -50 for both the old and new ROMs.

yes as long as scaling/offset is same in both defs should be ok


the more aggressive transient retard may dull on/off throttle response

[nikitopo]

I agree that in weird cases the best approach is to check with independent sensors. In my case it wasn't possible because of e-tuning. I feel lucky that the engine survived and the last 4 years runs strong.


Wishing good luck to Tor.

[Tor]

Originally Posted by steve99

yes as long as scaling/offset is same in both defs should be ok


the more aggressive transient retard may dull on/off throttle response

Since it's only happening between 3 and 4000 rpm, I suppose I could try and edit the A01G table in that area and less aggressively first and see if it has an effect.

I suppose a lighter flywheel would counter any dulling effect? My throwout bearing is making a lot of squeaking noises so a new clutch might be in a near future anyway.

[Tor]

Originally Posted by nikitopo

I agree that in weird cases the best approach is to check with independent sensors. In my case it wasn't possible because of e-tuning. I feel lucky that the engine survived and the last 4 years runs strong.


Wishing good luck to Tor.

In statistics, there is the saying "correlation does not imply causation".

Originally Posted by nikitopo

I decided to flash the ROM back to stock and left the tune to balance the parameters. Afterwards, I used a carbon build-up chemical cleaner. After 4 years the engine still goes fine, remains strong and the knock issues disappeared.

Maybe the knock issue would have disappeared if you had used the carbon build-up chemical cleaner before getting rid of the tune. Or maybe it was a 3rd undiscovered reason you had knock.

Just saying that it is not really possible to know if you tune was actually to blame without having investigated more thoroughly.

[nikitopo]

Well the car was almost new and the problem appeared after using the tune and disappeared after I flashed back the stock definitions. I tried sometimes back and forth, but I had always problems. I am not saying that the tune was completely rubbish. There was an amazing improvement in fuel consumption. In fact the car had a 30-40% improvement, by working more time in closed loop! However, I wouldn't risk to lose my engine just for this reason.


In general I am a fan of what tuning can achieve, but only if it can be done by correct means. Check here an older post of what tuning improvements the GT86s had in the SP3 class category at the Nürburgring 24h race:


http://www.ft86club.com/forums/showp...3&postcount=13

[steve99]

That seems to be an unrealistic increase in fuel economy on the forrester. If your engine smelt hot id suspect it was running way lean for some reason, were you able to log afr and commanded afr and exhaust gass temp ( cat temp) overly high exhaust temps can jndicate lean condition.

If you still have tactrix you can log your 86. Is interesting excersise. Which rom calibration are you ? All you need to do is used sd mam card in tactrix and load the logcfg.txt and tactrix will intertogate ecu. ( No alterations just log )

Or just use toyota techstream software, free, with your tactrix to monitor 86

[nikitopo]

These were the fuel maps of the forester. I don't think it was something extreme:

Stock


Tuned


The burning smell was in the middle range (1800 - 3200), where the tuned map adds a bit more fuel for knock resistance.


Anyway, I don't have the Tactrix cable anymore. So no more logging or changes for me

[steve99]

Originally Posted by nikitopo

These were the fuel maps of the forester. I don't think it was something extreme:

Stock


Tuned


The burning smell was in the middle range (1800 - 3200), where the tuned map adds a bit more fuel for knock resistance.


Anyway, I don't have the Tactrix cable anymore. So no more logging or changes for me

at low rpm cruising it probably wont be using the Open loop maps, it will likey be in closed loop. The closed loop tables are an offset from 14.7 and negative so if the closed loop tables contains -0.3 target is 14.4. If the def is bad or you put 0 or positive values in their then rearly weird stuff happens and it can run way lean and ltft is disabled and/or do other strange things.


mr throttle happy may have "hacked the 02 sensor scaling to fool ecu so he could run leaner cruise AFR than 14.7


see here


http://www.romraider.com/forum/viewt...p?f=15&t=11266


.

[oldgunfan]

Was there anything you changed just before you started hearing this noise? or did it start right after a track day or something like that. Also I would like to know what oil you use.

[Tor]

Update:
I tried the MY17 transient table. I can't say anything for sure yet as I also removed 2 deg timing in the same area between 3 and 4k rpm due to some AVCS adjustments in the same range. The noise went away due to one or the other of these changes. Also, the throttle felt less responsive as Steve predicted, but not too bad.

After the AVCS adjustments didn't give the results I wanted (tried to adjust AFR with Intake cam - didn't work, or at least the few deg I adjusted didn't have any visible effect) I went back to my previous tables and restored the 2 deg timing as well. Since I didn't hear any noises I halved the values in the transient table and flashed again. The noise came back. Whether it was due to restoring the 2 deg timing or halving the transient values I don't know yet.

I wanted to flash the full MY17 transient again but started working on scaling my MAF instead (it was a productive day ). Will let you know how it works out with the full MY17 table and my normal timing.

[freerunner]

Originally Posted by Tor

In my case, it happens when blipping the throttle during downshifts. E.g. cruising in 4th with 3500 rpm and downshifting to overtake. If I blip the throttle aggressively, then I get an audible knock like the video examples in the link above. I.e. it only happens in neutral.
[/url]

Do you get throttle lag when doing this? Do rpms drop before it's revving up?
I once had this issue and solved it with increasing the corresponding values in my exhaust-avcs tables.
The cause was my own fault, I advanced the exhaust cams before for the while load range. Now they're back to stock (0-2k rpm, up to 0.7 load).