knock detection, stock ecu VS utec
knock detection, stock ecu VS utec
Whats up all,
Well, I started over on my timing map last night and found a nice little trick as well as a question from today's data logs...
First the trick.... If you let the ecu control 0 and 10% load column, your will have very few transition problems.... the stock ecu ramps down timing fast in load cell 10 under load, in most cases down to between 12 to 16 degrees or so, more in some cases, so that by the time you transition into 20 load you are very close to some good numbers for a smooth transition.
Now the question, which is more sensitive to detonation, the stock ecu, or the utec (with default settings in the setup menu from rev 1.3) reading the factory knock sensor.
Today I saw one spot where the utec picked up knock bu the stock ecu did not:
4728,+23.9, 490, 27, 00, 00, +38.2, 10.0,ECU. , +0.0,ECU. , 606
4246,+23.7, 720,100, 10, 00, +17.9, 27.0,ECU. , +0.0,100.00, 625
4462,+23.7, 970,100, 10, 00, +13.9, 38.0,ECU. , +0.0,100.00, 961
4662,+23.9,1154,100, 10, 00, +8.8, 49.0,ECU. , +0.0,100.00,1152
4916,+23.9,1453,100, 40, 00, +7.5, 62.0,+12.0, -4.3,100.00,1184
5120,+24.1,1377,100, 50, 16, +7.9, 60.0, +8.0, -5.4,100.00,1396
5333,+23.9,1412,100, 70, 00, +9.3, 60.0, +8.0, -8.2,100.00,1428
5546,+24.1,1295,100, 70, 00, +10.8, 60.0, +8.0, -9.6,100.00,1207
5787,+23.9,1412,100, 80, 00, +13.4, 62.0,+12.2,-10.9,100.00,1295
6016,+23.9,1457,100, 80, 00, +15.6, 65.0,+13.1,-11.7,100.00,1381
6218,+23.9,1533,100, 80, 00, +17.1, 66.0,+13.8,-12.1,100.00,1461
6422,+23.7,1623,100, 90, 00, +18.6, 69.0,+14.7,-12.2,100.00,1461
6613,+23.7,1552,100, 90, 00, +20.7, 69.0,+15.5,-12.8,100.00,1322
Notice that the stock ecu did not pull any timing like it normally would if it agreed that there was a knock condition.
The map i'm running is not as "strong" up top as the last map but it is giving about zero knock at tip in and the transition is hard to notice... but i was hopeing that I had some room to advance the right lower third of the map to pick up the top end............
Any thoughts from TurboXS?
Well, I started over on my timing map last night and found a nice little trick as well as a question from today's data logs...
First the trick.... If you let the ecu control 0 and 10% load column, your will have very few transition problems.... the stock ecu ramps down timing fast in load cell 10 under load, in most cases down to between 12 to 16 degrees or so, more in some cases, so that by the time you transition into 20 load you are very close to some good numbers for a smooth transition.
Now the question, which is more sensitive to detonation, the stock ecu, or the utec (with default settings in the setup menu from rev 1.3) reading the factory knock sensor.
Today I saw one spot where the utec picked up knock bu the stock ecu did not:
4728,+23.9, 490, 27, 00, 00, +38.2, 10.0,ECU. , +0.0,ECU. , 606
4246,+23.7, 720,100, 10, 00, +17.9, 27.0,ECU. , +0.0,100.00, 625
4462,+23.7, 970,100, 10, 00, +13.9, 38.0,ECU. , +0.0,100.00, 961
4662,+23.9,1154,100, 10, 00, +8.8, 49.0,ECU. , +0.0,100.00,1152
4916,+23.9,1453,100, 40, 00, +7.5, 62.0,+12.0, -4.3,100.00,1184
5120,+24.1,1377,100, 50, 16, +7.9, 60.0, +8.0, -5.4,100.00,1396
5333,+23.9,1412,100, 70, 00, +9.3, 60.0, +8.0, -8.2,100.00,1428
5546,+24.1,1295,100, 70, 00, +10.8, 60.0, +8.0, -9.6,100.00,1207
5787,+23.9,1412,100, 80, 00, +13.4, 62.0,+12.2,-10.9,100.00,1295
6016,+23.9,1457,100, 80, 00, +15.6, 65.0,+13.1,-11.7,100.00,1381
6218,+23.9,1533,100, 80, 00, +17.1, 66.0,+13.8,-12.1,100.00,1461
6422,+23.7,1623,100, 90, 00, +18.6, 69.0,+14.7,-12.2,100.00,1461
6613,+23.7,1552,100, 90, 00, +20.7, 69.0,+15.5,-12.8,100.00,1322
Notice that the stock ecu did not pull any timing like it normally would if it agreed that there was a knock condition.
The map i'm running is not as "strong" up top as the last map but it is giving about zero knock at tip in and the transition is hard to notice... but i was hopeing that I had some room to advance the right lower third of the map to pick up the top end............
Any thoughts from TurboXS?
Newbie
Joined: Feb 2004
Posts: 27
Likes: 0
The problem here is that what you changed may have actually caused the knock event.
Notice that you went from 8 degrees under ECU control to 12 degrees under UTEC control. That's were you got the knock. The logs always report the knock events after the fact and when the UTEC is retarding the ignition.
Changing timing by more that 2-3 degrees under boost puts you at a higher risk of knock. Try transitioning timing by no more that 2 degrees at a time when under boost.
The ECU already knew it was close to knock as it was already retarding the timing. It went from 13.9 to 8.......
As for the knock thresholds, I know that they are on the concervative side. In your case I think this was real knock. The knock volume, or count, was 16 which means the noise spike was pretty big. Unless something was banging against your knock sensor, this was knock. Counts of 1 or 2 can be put down to phantom knock or false knock, but 16 is going to show up all the time.
Typically the transition "jerk" you can feel is fueling based, not timing based. Try using a complete ECU timing map and just use the UTEC for fueling. See if you still get the "jerk" If you do you know it's all down to fueling. (I only start to feel the effect of timing with a 10 degree difference. Below that, nothing) Please feel free to prove me wrong in respect to this issue. I know this is the case for the Subaru's, my assumption is the same for the EVO's.
Have fun tuning.
Notice that you went from 8 degrees under ECU control to 12 degrees under UTEC control. That's were you got the knock. The logs always report the knock events after the fact and when the UTEC is retarding the ignition.
Changing timing by more that 2-3 degrees under boost puts you at a higher risk of knock. Try transitioning timing by no more that 2 degrees at a time when under boost.
The ECU already knew it was close to knock as it was already retarding the timing. It went from 13.9 to 8.......
As for the knock thresholds, I know that they are on the concervative side. In your case I think this was real knock. The knock volume, or count, was 16 which means the noise spike was pretty big. Unless something was banging against your knock sensor, this was knock. Counts of 1 or 2 can be put down to phantom knock or false knock, but 16 is going to show up all the time.
Typically the transition "jerk" you can feel is fueling based, not timing based. Try using a complete ECU timing map and just use the UTEC for fueling. See if you still get the "jerk" If you do you know it's all down to fueling. (I only start to feel the effect of timing with a 10 degree difference. Below that, nothing) Please feel free to prove me wrong in respect to this issue. I know this is the case for the Subaru's, my assumption is the same for the EVO's.
Have fun tuning.
I hear what your saying with respect to the jump in timing causing the knock event and I have spent the last hour revising the timing map based off of my data logs from today....
My observation has been that the stock ecu (from the datalogs i have anyway) always retards the timing on heavy tip in then ramps it up. Again, I do agree that I needed to pull back a few more degrees for the transition to prevent the knock.
Also, I was not feeling any transition with this map at all, which jives with what your saying. The base map provided by turbo xs was the one that had harsh transitions. That map was leading to big drops in timing during transition. By allowing the ecu to control load cell 10%, you give it the chance to do it's thing and pull down the timing on tip in to a value that makes the handoff timing change smaller, or so it seems.
I ran the turboxs stage 1 map with timing map set to ecu (-100) for several days. This map still had the txs fuel values, and in fact i am still using there fuel numbers as they were reported to be conservative.. With the ecu controlled timing i was getting knock and rough part throttle transition in regions were the fuel was not touched from the factory setting (in load cells where the fuel trim was set to 0). Not saying that proves anything one way or the other.. just someting I observed.
My observation has been that the stock ecu (from the datalogs i have anyway) always retards the timing on heavy tip in then ramps it up. Again, I do agree that I needed to pull back a few more degrees for the transition to prevent the knock.
Also, I was not feeling any transition with this map at all, which jives with what your saying. The base map provided by turbo xs was the one that had harsh transitions. That map was leading to big drops in timing during transition. By allowing the ecu to control load cell 10%, you give it the chance to do it's thing and pull down the timing on tip in to a value that makes the handoff timing change smaller, or so it seems.
I ran the turboxs stage 1 map with timing map set to ecu (-100) for several days. This map still had the txs fuel values, and in fact i am still using there fuel numbers as they were reported to be conservative.. With the ecu controlled timing i was getting knock and rough part throttle transition in regions were the fuel was not touched from the factory setting (in load cells where the fuel trim was set to 0). Not saying that proves anything one way or the other.. just someting I observed.