I have a graph setup in 2.25 to log KR(+ and -) against RPM and Load. I thought I had found a fool proof way to tune the spark tables, but alas......No.
I have a cell that has a Minimum KR value of -4 , a Max value of 1 and an average of 0.
I don't want to have knock anywhere, but I seem to be chasing my tail. Most of the knock occurs and low RPM(1k to 3k) and low load(.4 to .6).
Has anyone here dealt with this and come up with a solution?
Last edited by txcharlie; 10-17-2015 at 10:05 AM.
Not sure if this is helpful but I've also been trying to figure out how timing is calculated and I've learned the following:
1. You need to log "Spark Source", "Torque Source" and "Mapped Point Weight" as well to better understand what tables the ECM is using
2. Mine seems to only add timing when running off the "Borderline" tables, otherwise KR is 0
3. Even when running off the "Borderline" tables plus the "Knock Retard" added, it still has never exceeded the "MBT" values, but gets pretty close
4. I have no idea how the "Knock Retard" timing add works, maybe just keeps adding timing until knock occurs
5. When spark source = "Torque Control" I have no clue how the timing values are calculated, and there seems to be nothing added by "Knock Retard".
My ECM seems to use mapped point 20, 21 and 22 when at full throttle, "Borderline" table + "Knock Retard" = pretty close to the timing I'm actually getting, so when I change my timing I'm going to bump those 3 "Borderline" tables and see what happens.
I've got a 2015 base model M6 with the Shuffle turned Off.
I know about the change in KR reading, but I still had both + and - KR in the same cell on the same log.Originally Posted by Higgs Boson
Thanks for you input.
1. Yes, I log all of those
2. I agree, I've only seen KR sensor add timing when looking at Borderline
3. It can't exceed MBT. It's a clip for Borderline
4. You can change how much and how fast the KR sensor will add timing.
5. When you see Torque Control for Spark Source, look at Torque Source and you will see what torque limit it is using. If someone has a list of definitions of the Torque Source readings, please share.
* 0 - Driver Demand
** 1 - Transmission torque truncation
** 2 - Traction Control
** 3 - Vehicle Speed Limiting
** 4 - Engine Speed Limiting
** 5 - Anticlunk Tipin torque limiting or Tipin rate limiting
** 6 - Tipout torque
** 7 - Transmission Shift Modulation
** 8 - Engine Oil Over temperature
** 9 - Target N
** 10 - Powertrain Limiting Protection FMEM Limit
** 11 - ETC FMEM - RPM Guard
** 12 - PATS
** 13 - OSCMOD
** 14 - Speed Control
** 15 - Transfer Case Torque Limit
** 16 - Torque Based Decel
** 17 - VDE
** 18 - Engine Indicated Torque Limit
** 19 - Engine Combustion Stability Torque Limit
** 20 - HEV Torque Reduction at engine start or stop
** 21 - HEV Tipout Torque Reduction
Courtesy of SultanHassanMasTuning
http://www.hptuners.com/forum/showth...-Torque-Source
Thanks.
I'm going to mess with number 4. this weekend. Saw that my tables allow "Knock" to add up to 10 degrees but it never adds more than 2 to 5 degrees, so maybe if I change the add rate it will get closer to "MBT" faster.
Thanks! Some of them are self explanatory, but some are not. Target N,OSCMOD,VDE, HEV, FMEM??? It's like being in the military. LOL!
It's probably transitioning to another MP, load, or RPM before it can add more timing or that's all it could add before it detected knock and stopped adding.
Target N occurs during gear changes, so must be something related to being in neutral?
OSCMOD is the shuffle logic, I've seen it pull timing down during hard pulls in the lower gears. I turned mine off by changing the max rpm for the logic to operate, I think you posted earlier that you disabled yours. Odd thing it shows up in the Electronic Throttle section of the ECM but mine seems to pull timing, sometimes all timing.
Lets keep going......
Yeah, thats a good point, mine seems to be able to add more advance in the upper gears when the rpm is climbing slower, but in 1st and 2nd it never gets more than 2 to 3 degrees of advance. I'll share what I learn this weekend.
My car is a 2014. I believe the shuffle logic is only on 2015 cars.
I believe I'm dealing with some false knock. I have read that real knock readings look like stair steps. Most of the knock on my logs look like blocks.
No other input??
I have seen the same knock around 40-60% load as well with my setup and both values in the graph. What I've done is logged all enabled mapped points, as well as the usual load, spark, KR, etc. I also plot the KR and I can view the knock as a dip in the line chart. I can click this point on the chart and look at the channels to find the load, rpm and mapped points it is using, and adjust those timing tables accordingly. Hope this helps!
Ok, so the 32412 - Knock Fast Advance Rate vs. RPM table is related to the step change amount of advance added. Stock for me was 0.5 and that was the step up is advance added each time, changed it to 0.125 and that was the step change in advance. BUT sometimes it incremented a different step amount that was not a table value AND sometimes it stepped down an amount not a table value. Still a mystery??
I also changed the:
1122 - Knock Fast Advance Rate Shifting vs. RPM:
50236 - Knock Advance Rate vs. RPM:
tables to 0.125 from the 0.25 stock values and I don't think I was adding timing any faster or slower than before. Logs seem to indicate may actually be slower than before.
So in conclusion the only thing I learned is the values in the 32412 - Knock Fast Advance Rate vs. RPM table are somehow used in calculating the incremental step up change that is added to timing. I've changed it to 1.00 from the stock value of 0.5 and I expect timing to be added in 1 degree steps, but won't be able to verify until weekend again.
I had all MPs except OP disabled, that way I'm not dealing with blending of the tables. I thought this would simplify things.
I set up a graph for KR like this:
In fact, these are from the log showing knock and no knock in the same cell. Look at 3k rpm and load 55 and you will see 1(knock) as maximum and -4(no knock) as minimum.
I'm not aware of any parameter that allows you to adjust the increments of spark. I've seen .1 increments on my logs.
Do you guys start with a stock spark table and just modify the high load rows until you get knock? I have been using the KR and letting it add up to 4 degrees of advance and then adding that to the spark tables. Is this an incorrect way of doing it?
One issue I have seen is what I stated in my opening post. It will add and subtract spark in the same areas of the same table.
What I have decided to do, is ignore any areas the knock sensor is adding spark below a certain load/rpm to avoid these areas that may be false knock.
What do you guys think?
Last week I noticed how much the spark changes. The low and highs were seperated by 6 degrees in the same cell(even at 6500rpm and 100% load it varied by 6 degrees from one pull to the next). I can't log knock, I'm limited to 2.24 right now. According to 2.24, the motor was pulling over 100,000 degrees of timing last week. My air fuels were in the millions too lol. Wanna add some timing to this thing. If anyone was able to log KR successfully in 2.24 let me know.
I'm pretty interested in the approach others take to adding timing. I was planning on just letting the knock sensors add a few extra degrees in the high rpms as a start.
KR does not work correctly in 2.24 for copperhead pcm.
Reply With Quote