Printable View
I am looking for a description and operation of ignition timing control on an S550 5.0L. I am interested in how the knock control in particular is meant to operate. I am tuning a car that is continuously pulling timing regardless of ignition timing requested at load above 80%. Car is currently running 20-23* ignition advance at WOT. I would like to adjust knock sensitivity but I cannot find a proper description.
Thanks.
It's probably backwards than what you think and adding timing if you're not used to tuning Fords. Negative spark retard is advance. Ford uses knock sensors to add and remove timing. S550 knock sensor calibration seems good to be and they can run good timing.
Look at the knock Advance Limit Table. It will have 10s stock indicating that the computer can advance timing up to 10 degrees if no knock is detected. If you zero that your knock will go away but I wouldn't do it. Just bump up your Borderline and let the sensors keep adding a few degrees. You should find it likes it in the high 20s or low 30s depending on temperature, fuel, and mods
In another thread on the forum we have tested the sensitivity and found that the knock sensitivity increases with the values increased. Eric indicated that they would probably modify the description to more clearly indicate how this parameter affects the knock sensitivity. As ridenrunwv indicated, the KR values are reversed. Negative is adding timing and vice versa.
S550 knock sensor calibration is different and more complex than the S197 calibration we were discussing in the other thread.
Thank you Ridenrunwv. That is exactly the case. I appreciate the quick response.
No problem. Cylinder Pressure Limit will be your next confusion if you don't look at it also by the way. I max it out
Sorry - I neglected to look closely at your year. ridenrun wv is absolutely correct; the S550s are significantly more sophisticated (complicated) than the S197. I have searched through the Ford patents today after realizing my mistake, and so far, nothing. I did notice that of the few modified S550s in the repository, they left the KS values at OEM. If I find some additional data I will post it. Or maybe someone on the forum with some hands-on experience with these models might provide some explanation, that would be beneficial. BTW - is it still removing timing? or adding it?
It's the same as far as adding and removing timing. Negative knock retard is adding timing and positive removing.
If your tuning an engine, why would you ever want it adding spark your going to find what made the most power safest? Id think it best to zero out its adders (, only let it pull when it needs to, and do testing to find its max happiest timing?
In this file are you saying the descriptions are wrong knock Retard Limit is all -10s and Knock Advance Limit is 6 to 4 in most areas. I was thinking I had to 0 out advance limit to keep pcm from adding timing over whats in the tables?
Zeroing the adder past partial throttle load values is how i've always done it. On the s550 if you zero the adders the timing ends up much lower than optimal and you'll have to go back over your timing tables to get it where you want it on a dyno.Quote:
Originally Posted by GrannySShifting
For tuning you want your spark at one of two values, Borderline or MBT(which ever is lower). These will result in the highest cylinder pressures, most torque, and least hammering effect from rapidly changing pressures.
MBT (minimum advance for best torque) ignition value has a few different names people call it, a common one is Maximum brake torque. This is the value of ignition timing that will allow the engine to produce the greatest torque. It does this by matching the rise in cylinder pressure, caused by the flame front, to match the cylinder volume/ compression pressure change as the piston is moving. No excess pressure before TDC and no drop off in pressure after TDC. It is the "perfect ignition value". So why would you ever want it to use borderline and not just want the engine to use this value and only this value? Well for one the air fuel ratio has an effect on flame speed, RPM has an effect on piston speed, and devices such as IMRC tumbling the air creating a better in cylinder distribution(centered around the ignition source) can also have an effect. The air fuel ratio in the cylinder can have an effect of the speed of the flame and therefore a lambda modifier table is needed. In older vehicles they had IMRC open/closed modifiers as well. New vehicles with the dynamic relationship of VCT and IMRC optimization we have twice as many MBT tables(why we can't just use the OP tables). The second and main reason for not using only an MBT value, the fuel quality just can't achieve MBT before its detonation point(borderline).
Borderline ignition is the highest ignition value that a fuel can still resist detonation. Above this value the engine will ping, knock, or detonate whatever you like to call it. This value is highly dependent on many other factors... fuel quality, temperatures, lambda being commanded, engine load, RPM just to name a few... It is a constantly moving target based on all these factors being so dynamic.
If you just compare a MBT table and a borderline table you will notice in low loads MBT is lower than borderline. This is because pump gas which the tune was calibrated for is capable or reaching the MBT value before it will detonate. Raising your MBT value to the borderline value would result in less torque as ignition would be too advance, you would get a large pressure spike before TDC fighting the pistons movement and creating more than necessary pumping losses. Ford calibrated the borderline tables at 1 stoich, 100 IAT, and 200 ECT. The speed density values are calibrated at these values as well(keep that in mind if playing with the new calculator).
In the middle loads you will notice a region where MBT and borderline are very close to the same value. At some point the fuel will not be able to achieve the MBT value and instead will detonate. If you take in to account how dynamic the actual borderline value is you can see that a borderline table is not enough to represent reality. That's when the spark advance correction tables come into play to modify the table to make it reflect more closely the actual borderline value at those conditions.
The knock sensor systems main job is to adjust for fuel quality and find the transition point from running MBT value and running a borderline value. You can zero out all the spark correction tables(environmental tables as I call them), but all this does is put more burden on the knock sensor to find the borderline of the fuel you have in the tank. If you zero out spark correction tables, advance limit, all other modifiers, and raise the limit tables, you will see it run the border line values. On a dyno you could then find the optimal ignition timing to get the best numbers or dial them back to add a bit of "safety" in the tune. If you do this on a cold day after getting some really good pump gas you will see good results. Then 6 months later you get a tank of junk gas and its a day 60+* warmer, the car has just been idling in traffic/staging lanes so IATS are 130*+, all that safety you put into the tune by taking out 3* from the borderline is totally worth less. The same goes for the opposite situation, you tuned it on a warm day with bad gas. The on the cold days with good gas there is power left on the table.
You want to tune the ECU, not to make the most/ safest power right now, but to give it the tools to find the optimal settings on its own with the least amount of effort so that on any day in any conditions you have a safe tune thats also giving the optimal amount of power. Most tuners don't have the time or resources to properly simulate and calibrate for all these conditions. Its why Whipple brags so much about their tunes. In that case don't undo what ford did to make it easier on your self as it will not lead to a better/ safer tune. Ford leaves a lot of room in the borderline values to compensate for a variety of octanes and fuel qualities. Decreasing this and requiring the use of quality fuel is basically what most email tunes are, besides setting it up for what ever modifications the customer is running. That's why " you have bad gas" and not "its not the tune" are the common responses to "why is there knock in my logs?" Some people will only have access to the worst fuel the the stock calibration is setup for and a "tune" will not help at all. Unless we are talking about automatic transmission settings.
If you are tuning where power levels are possibly approaching the physical limits, then yes limit the amount of ignition it can reach by calibrating the advance limit and spark correction modifier tables. Take note of the lambda you are targeting and temperatures you are logging and modify the tables as you see appropriate, use the factory tables as reference or make up your own more or less aggressive scheme for colder and warmer conditions. The goal isn't to be perfect, but reduce the burden of the knock sensors adding ignition to find the actual borderline or your maximum ignition value. you also want it to modify the borderline value so that the knock sensors never need to remove timing below your set borderline table value. Once below the borderline table value it treats this as a fuel quality issue and only makes coarse adjustments. Get enough and it will learn and apply the octane adjustment table. Don't forget the multipliers as they are useful to have it apply as much, as little, or not at certain RPMs and loads.
In eco boost and newer cars these coarse adjustments are learned as octane adjustment ratio OAR, kinda like long term fuel trims are learned, but to adjust knock sensors for small variances in fuel quality. If you tune by zeroing out spark correction you will not have an easy time figuring out optimizing OAR as it will be all over based on temperature variations trying to learn the best compensation.
TIP:
I read a white paper about stoich 1.1 giving the fastest flame front. In the MBT modifier you can see it raises the MBT value the most at that stoich(3.25). Running stoich 1.1 in the lower loads(MBT limited not knock limited) instead of 1 can result in higher low end torque and give a better partial throttle feel. Better fuel economy.
Attachment 76055
The scanner was setup to show negatives as a good thing and positives as a bad thing, Eric just made a post about why and a new PID you can use to have it display in the opposite way. In the editor they represent the values that can be applied to the borderline tables to find the actual borderline.Quote:
Originally Posted by GrannySShifting
I know the difference between MBT and the actual timing you may be able to run when octane limited. What I wasnt sure of is how its actually determining what the timing is going to be - GM is very obvious you have a high octane table-THAT is what the spark will be period unless its pulling due to KR, or IAT retard table, or torque management event etc etc
Thats exactly how Ive tuned LS cars for 15 years, and have hurt 3 stock pistons out of the thousands Ive done making 500-1200 hp on them. I tuned for the coldest densist air, when its 60 degrees warmer and the IATs are hotter it is setup to pull timing and even in speed density its not a problem, as hotter less dense air just makes it richer still. I dont care if you leave a tenth on the table in 115 degree track conditions. If its a hot ass day Im tuning on, I dont push it near the edge in A/FQuote:
The knock sensor systems main job is to adjust for fuel quality and find the transition point from running MBT value and running a borderline value. You can zero out all the spark correction tables(environmental tables as I call them), but all this does is put more burden on the knock sensor to find the borderline of the fuel you have in the tank. If you zero out spark correction tables, advance limit, all other modifiers, and raise the limit tables, you will see it run the border line values. On a dyno you could then find the optimal ignition timing to get the best numbers or dial them back to add a bit of "safety" in the tune. If you do this on a cold day after getting some really good pump gas you will see good results. Then 6 months later you get a tank of junk gas and its a day 60+* warmer, the car has just been idling in traffic/staging lanes so IATS are 130*+, all that safety you put into the tune by taking out 3* from the borderline is totally worth less. The same goes for the opposite situation, you tuned it on a warm day with bad gas. The on the cold days with good gas there is power left on the table.
So it targets borderline table as starting point if its 100% weighted to MP14 uses that table? , Spark Preign Limit says -1.3 at .8 lambda its pulling 1.3 degree, it has to be under the preignition limit of 28.whatever degrees there, not sure what MBT Spark Lambda Correction does (is it taking out that amount of timing in degrees when commanding .80 lambda?) has to be under Spark Cylinder Pressure Limit, not sure Spark Borderline Lambda Correction factors in? and if I have the KR adder zeroed out above .80 load, no IAT/ECT modifiers that should be the final amount?
Not really sure that seems to correlate scanner shows 17.7 degrees spark on chart at one point, in channels shows 29 degrees, with MBT at 27.1 and Borderline at 16.3??? If borderline is supposed to be the max (and I have KR adder 0d out) how is it higher than Borderline?
What is Spark Retard vs Piston Rattle doing?
I would work with support as your MP weights are not reading correctly. Cams are moving but its staying in MP14 is not correct.
It uses MBT + MBT spark correction, Borderline+ BL spark correction, or a limit/ maximum/ minimum what ever is the lowest value.
Spark retard vs piston rattle is just another spark correction applied to the value before the spark advance final total. It along with combustion effects table only retard the total. They are mainly to prevent abnormal combustion and unwanted engine noise.
Yeah I asked you abut the MP14 thing in a PM a few days ago thought it was oddQuote:
Originally Posted by murfie
Why have a table that retards by load only when you could build that into the regular spark tables? Its entirely redundant and retarded (pun intended). It has 18 at .5 load and 3000 rpm. Is that supposed to mean its retarding 18 degrees there always (0 this out itll have TONS more timing in the midrange?), or that a high limit or? The redundancy in these ECUs is silly in some places and the way they scale things - IAT retard is 0 at 100 and -25 at 150 degrees, but mult by .14 at .4 airload (for -3.5 total) and .1 at .8 airload (for -2.5) why not just use reasonable numbers in the first place?
Can you look at this log, and tell me at 9:24:15 how the timing is added up, or what even the final result is? Chart shows 17.9, channels shows 13.5, Borderline is 14.5, MBT is 24.3 (which it should never even try to go to with Borderline being 14.5, but KR correction is adding 3 - well the chart shows it at 3.5??? Nothing jives as far as I can tell
Not really sure I follow building it into the regular spark tables. There is no base values its working up from. The MBT and borderline tables are the main tables and the represent those exact spark advance values or as close as they could measure for the conditions the represent. The ECU is working toward minimum spark advance value(base source), with the exception of the knock advance logic which occurs off the borderline values. Spark source PID is very helpful in figuring out where the spark is coming from. If you lower the MBT values you mess up the TQ model and all the other IPC's. MBT is usually as close to a constant as it gets with internal combustion engine and is a great thing to base other outputs off of. If you lower the borderline values it will most likely be over ridden by the knock advance logic which is inherently very dynamic, for better or for worst. You need these modifier tables to accomplish things like retard a specific amount for combustion stability or noise and vibration dampening.
Then I am also not sure how you could simplify scaling an output modifier based on three variable inputs. Granted they setup the IAT multiplier table in a way the RPM axis is not needed, But the ECT modifier it definitely was. The IAT VS RPM relationship may not be that significant right now, but in the future it could be the difference they need in MPG or emissions. Then as for the values the smaller the multiplier the finer the adjustment can be the more precision you can get out of the ECU. Takes advantage of the speed ECUs operate at. Normal numbers would be too coarse of an adjustment.
Ive looked at the log and am not sure what you mean by 9:24:15 as the log is only 4m 26s long.
Looking at it the only thing I see is after you accelerate and let off you get into spark source "torque control". This is because target RPM become the Idle RPM and spark/fuel/airflow is all determined by the TQ model.
Ok 3:35 in looks like, I was reading the timestamp.
What I meant by he retard by load piston rattle table building into regular tables - it appears to remove 18 degrees timing at 3000 rpm and .50 load.. you could just subtract 18 from every borderline and MBT table at that same rpm and load, they are all scaled the same way. Im pretty sure that table isnt removing 18 degree however at all times as it implies?? Same goes for spark retard for combustion effects. They could have been combined into one table if so, or meshed in with all the MP timing
Most logical modifiers Ive seen would be done as in remove 3 degrees at 150 IAT temp, with a 1.0x correction at some normal load say .8 and 1.4x correction at .14 load to get tthe same final result but looks a hell of alot more sane to me
There may be some kind of knock sensor logic that applies that table in certain conditions that we cant see, or it could just be applying all the time I'm not sure. I just zero it as excessive noise and vibrations don't bother me.
3:35
Attachment 76081
borderline is 10.1* MBT is 22.8*. So It would use the borderline value. the knock correction is at 6* so the final spark advance is 16.1*.
Your MP weights are not working so cant use that. cam angles are ~ 15 IVO and 30 EVC. Load at 1.2 and RPM at 2200. VCT mode is fuel economy drive. Looking it up in your VCT setup that would be some blend of MP 19 and 21. The scaling on the distance tables could be setup a bit better. Looking in your tune and interpolating borderline values they are ~10*. All your corrections would be 0 or darn close to it(lambda is adding a small bit).Seems right to me.
Attachment 76080
Not sure why your getting knock advance when the limit is 0? I would do a write entire of the stock file then another one of the current tune just to make sure your settings are getting applied. I know in the editor the load axis is tied together with the retard limit. If that not the same in the ECU right now a quick write wont work with what you are trying to setup.
So you zero Piston Rattle and Comb Effects like I was thinking to do. MBT Spark Lambda correction is adding 1 degree if we go by what your commanded graph says, little over 2 degrees if we go by what mine says. Borderline Lambda adder doing roughly the same. Its going to start with the borderline value 10.9+6=16.9, which its getting from whatever MP points its looking at (inferred that its MP 19/21 blend looking at cams position) and its not going to be clipped unless MBT was lower than it in which case it would start with MBT value? But MY commanded EQ should be taking @2 degrees out of Borderline and MBT values so really should be 14.9. No IAT/ECT corrections. So what other stuff could make it not there? And what is timing advance SAE on my channels referencing then because it doesnt correlate to any of that? The spark value on the chart vs time sure looks suspect to me, rarely is ignition timing a bunch of smooth slides up and down????
Do the tables when referencing Lambda base off of the averaged commanded EQ - or does it use the commanded for each bank and vary spark bank/bank, or blend the two individual bank commands? My Bank to Bank doesnt match the overall commanded which doesnt make sense most of the time, fuel trims are pulling fuel so youd think they would commanded differing amounts leaner, but the individual banks are commanding RICHER than the overall singular commanded EQ
For some reason at that time point my values dont match yours, Commanded EQ is shown pretty differently, MBT is 22.8 vs 23.0, Border is 10.1 vs 10.9, rpm is same, odd.
Attachment 76083
So what gets written with a quick write (which isnt quick with the Fords takes 3 minutes vs 30 seconds Im used to) and what additional gets written with entire? Still brickable writing entire?Quote:
Originally Posted by murfie
Not sure why Im still getting knock advance when its 0 above .7 load
Another thing I feel like my load values are kind of high, is there a way to rescale that?
There isnt a clear way to log all the multipliers adders subtractors etc and where its picking from and percentages of MP is there?
Yes, here:Quote:
Originally Posted by GrannySShifting
Attachment 76208
Stock is 1,05. You already have this raised to 1,3. Your air-charge at WOT in speed density is also raised.