TDStuart Tuning Adventure

[engineermike]

@tdstuart if you log your desired airflow and MAF, they should match very closely. Based on the log you posted, yours is off as much as 30%. Assuming your MAF, TB, and Injector data are correct, this can only point to the SD model as being the problem. That said, the more I think about it, the more I think you can get it in the ballpark without actually having to measure the manifold pressure. You see, the throttle body flow is modelled in the logic and the blade angle is sent to a pre-calculated position to achieve a desired airflow. Then the MAF reads the actual airflow and a feedback loop correction is made, but the feedback loop intentionally has limitations. Therefore, you can tune the SD model and get it "in the ballpark" by modifying it until the desired airflow matches up to MAF.

If this were my car, I would do the following steps:
1. Disable the throttle feedback loop. The feedback loop corrects errors so it will mask problems that you want to correct. This is a temporary change just until the SD model is calibrated, and would be disabled by setting auF0080 to 0.
2. Lock it in to one Mapped Point in order to focus on just tuning that one. I would start with Mapped Point 5 since that is where it hovers at cruise. This is done by setting auF33355 to all 5's. This will force it to MP5 at all times not at idle or WOT.
3. With the above changes, desired airflow and MAF can be logged and compared to see the "raw" error. Calculated MAP, rpm, MP5 weight should be logged as well.
4. Start tuning the MP5 SD model. This will be iterative and each value only applies to that specific rpm. It's best to try holding engine speed constant and varying the load with the throttle as best you can. Specifically, I would start by modifying the slope values, auF33197, as this will scale the entire curve up and down by roughly the same % change. My bet is the slope needs to be scaled up and I'd start in 10% increments and see how the MAF vs desired airflow error responds. If there is no discernable change, I'd try larger adjustments. I would try to get the error under 5% by modifying just the slope values.
5. Put auF0080 back stock to re-activate the feedback correction.

Once this is done, MP5 SD model is fairly well calibrated and you'd have a decision to make. You can elect to just leave auF33355 as it is and just let it use MP5 at all times not idle and WOT. Lund often takes this approach, except they don't bother using a high efficiency point like MP5. Or, if you are successful and want to continue on, you can proceed to other Mapped Points that you might want to use. At this point, you can begin to see why many, such as myself, try to reduce the number of active mapped points as much as possible. I'm running only 5 points, and others have reported similar strategies.

Sponsored

 

[GregO]

Therefore, you can tune the SD model and get it "in the ballpark" by modifying it until the desired airflow matches up to MAF.

Exactly !
Trust the OEM Airbox/OEM MAF transfers so long as the transfer function is unaltered. Use a new Motorcraft pleated paper air filter not a high flow gauze type for testing purposes.

Good call EngMike, the placement of a MAP sensor is critical especially in respect to response time.

 

[Pistol_91]

I agree with Mike. It's possible to tune SD without a map sensor. I have done it. It works just fine.

 

[tdstuart]

@tdstuart if you log your desired airflow and MAF, they should match very closely. Based on the log you posted, yours is off as much as 30%. Assuming your MAF, TB, and Injector data are correct, this can only point to the SD model as being the problem. That said, the more I think about it, the more I think you can get it in the ballpark without actually having to measure the manifold pressure. You see, the throttle body flow is modelled in the logic and the blade angle is sent to a pre-calculated position to achieve a desired airflow. Then the MAF reads the actual airflow and a feedback loop correction is made, but the feedback loop intentionally has limitations. Therefore, you can tune the SD model and get it "in the ballpark" by modifying it until the desired airflow matches up to MAF.

If this were my car, I would do the following steps:
1. Disable the throttle feedback loop. The feedback loop corrects errors so it will mask problems that you want to correct. This is a temporary change just until the SD model is calibrated, and would be disabled by setting auF0080 to 0.
2. Lock it in to one Mapped Point in order to focus on just tuning that one. I would start with Mapped Point 5 since that is where it hovers at cruise. This is done by setting auF33355 to all 5's. This will force it to MP5 at all times not at idle or WOT.
3. With the above changes, desired airflow and MAF can be logged and compared to see the "raw" error. Calculated MAP, rpm, MP5 weight should be logged as well.
4. Start tuning the MP5 SD model. This will be iterative and each value only applies to that specific rpm. It's best to try holding engine speed constant and varying the load with the throttle as best you can. Specifically, I would start by modifying the slope values, auF33197, as this will scale the entire curve up and down by roughly the same % change. My bet is the slope needs to be scaled up and I'd start in 10% increments and see how the MAF vs desired airflow error responds. If there is no discernable change, I'd try larger adjustments. I would try to get the error under 5% by modifying just the slope values.
5. Put auF0080 back stock to re-activate the feedback correction.

Once this is done, MP5 SD model is fairly well calibrated and you'd have a decision to make. You can elect to just leave auF33355 as it is and just let it use MP5 at all times not idle and WOT. Lund often takes this approach, except they don't bother using a high efficiency point like MP5. Or, if you are successful and want to continue on, you can proceed to other Mapped Points that you might want to use. At this point, you can begin to see why many, such as myself, try to reduce the number of active mapped points as much as possible. I'm running only 5 points, and others have reported similar strategies.

So with the imrc would you want to do another non wot mapped point for imrc open?

 

[engineermike]

So with the imrc would you want to do another non wot mapped point for imrc open?

If you're going to be a part throttle with IMRC open at 40/50 cam timing (MP17) then it needs to be calibrated as well, but you're going to want to simplify this process as much as possible. I have ideas around this, like not allowing IMRC to be open at 40/50, but don't want to confuse the issue. The bottom line is if the mapped point is used, then the SD model needs to be calibrated. The problem with the coyote is that at WOT it doesn't "land on" any mapped points or even snap lines, so it's interpolating data from a dozen mapped points that now all need to be in good calibration. This is not the case with ecoboost, Roush SC, GT500, and others. You can see why some of us have elected to reduce the number of mapped points, and even using only 2 or 3 mapped points at WOT using OP cam timing and snap lines.

 

[Pistol_91]

If you're going to be a part throttle with IMRC open at 40/50 cam timing (MP17) then it needs to be calibrated as well, but you're going to want to simplify this process as much as possible. I have ideas around this, like not allowing IMRC to be open at 40/50, but don't want to confuse the issue. The bottom line is if the mapped point is used, then the SD model needs to be calibrated. The problem with the coyote is that at WOT it doesn't "land on" any mapped points or even snap lines, so it's interpolating data from a dozen mapped points that now all need to be in good calibration. This is not the case with ecoboost, Roush SC, GT500, and others. You can see why some of us have elected to reduce the number of mapped points, and even using only 2 or 3 mapped points at WOT using OP cam timing and snap lines.

Mike... Easy fix. He can change the OP tables to match that of 1 or 2 or even 3 mapped points and it will blend between those 3 and use those weights at 100%. It's blends between so many while under WOT because the values they have in the OP table are in between so many different mapped points so it's picking a little bit here and little bit there.

 

[markmurfie]

Directly from the MAF sensor and RPM it gives you cylinder airmass, from that calculated MAP using your SD quad, slope and offset coeffecients.

simply:
MAP = (Quad *(airmass^2))+ (slope*airmass) + offset

MAP estimate (TB model) comes from that same airflow through the throttle model. which is more than the predicted throttle angle and effective area inverse lookup tables, which only describe high MAP ETC vacuum.

These two PIDs in the logs you post agreed up until calculated map went above calculated barometric pressure. The throttle body model can't have a pressure higher than the barometric pressure, unless you fill out the delta above baro table or boost maximum limit-MAF tables as PCMtec calls it. Naturally aspriated cars dont need this. They calculate the barometrc pressure if they dont have a sensor. When you fixed the calculated barometric pressure it introduced a bunch of error between these two estimated MAP pressures.

Im not sure you will make any progress in correcting des MAF to MAF by blindly adjusting one of the three coeffecients in the clyinder airmass to MAP SD model. In fact you are going to pull your hair out if you have any from the error you cause in fueling and MAP in regions you were not tring to touch.

If you dont want to add a MAP sensor, but want to correct something, you would be better off matching your airmass values to the MAP estimate(TB model) in a scatter plot and regressing coefficients out of that relationship so that your MAF sensor and calculated MAP comes out to what the TB values are saying. Fix barometric pressure to a fairly accurate estimate.
Will it be right? who knows with out a MAP sensor. Will it make the ECU happy? Probably.

The TB model is saying the engines VE is much less than the stock SD. I guess thats expected from larger cams moving it up in the RPMs. Maybe an indication of a hurt engine, or a bad throttle body. Of course isolating the MPs and RPM ranges you would get a much more accurate representation of where VE is less and where its greater than stock.

 

[engineermike]

Mike... Easy fix. He can change the OP tables to match that of 1 or 2 or even 3 mapped points and it will blend between those 3 and use those weights at 100%. It's blends between so many while under WOT because the values they have in the OP table are in between so many different mapped points so it's picking a little bit here and little bit there.

That’s more of less what I did for mine but the existing points weren’t exactly where I wanted to be at wot so I had to move them a little. In order for this to work properly, you also have to add snap lines between the points you intend for OP.

Really the Roush sc cal to me was the best example of what good looks like for an aftermarket or homebrew tune.

 

[Pistol_91]

That’s more of less what I did for mine but the existing points weren’t exactly where I wanted to be at wot so I had to move them a little. In order for this to work properly, you also have to add snap lines between the points you intend for OP.

Really the Roush sc cal to me was the best example of what good looks like for an aftermarket or homebrew tune.

Agreed but I meant more of a temporary thing if he needs to tune SD for a certain mapped points for WOT

 

[engineermike]

…The TB model is saying the engines VE is much less than the stock SD. I guess thats expected from larger cams moving it up in the RPMs. Maybe an indication of a hurt engine, or a bad throttle body.

I’ve posted this and the reason why twice already in this thread. The comp intake cam adds all its duration (20+ deg) to the closing side of the lobe, so at engines speeds below which ram-tuning can work, you reverse flow out of the cylinder on the compression stroke. Therefore, it takes higher map to get the same load. As such, for any measured, known load, the map calculation from the sd model will be erroneously lower (more negative) than actual.

 

[markmurfie]

I’ve posted this and the reason why twice already in this thread. The comp intake cam adds all its duration (20+ deg) to the closing side of the lobe, so at engines speeds below which ram-tuning can work, you reverse flow out of the cylinder on the compression stroke. Therefore, it takes higher map to get the same load. As such, for any measured, known load, the map calculation from the sd model will be erroneously lower (more negative) than actual.

By the ECU limiting the calculated barometric pressure, the SD gets to the 100%VE line sooner(under it is over 100%VE), but leaves WOT unable to get to a high enough load.

Letting barometric learn lower

Fixing barometric at 30

The engines VE just doesn't ever seem to get where it should be, even at higher loads, according to the throttle body model.

 

[GregO]

Is there a possibility the cams are not degreed correctly ? I’m sure they were installed “dot to dot” and not validated with a degree wheel kit to the cam grinders spec.
The OP mentioned installing a steel cam sprocket, there’s a possibility the sprocket keyway isn’t exactly in the OEM location.
The heads have been cut and I believe he used thinner head gaskets thus bringing the cams closer to the crank axis which slightly changes the cam to crank sprocket timing.
I see a lot of different tolerances stacking up here on the motor build.
“Dot to Dot” really only works with OEM components within OEM spec.

 

[bankyf]

Is there a possibility the cams are not degreed correctly ? I’m sure they were installed “dot to dot” and not validated with a degree wheel kit to the cam grinders spec.
The OP mentioned installing a steel cam sprocket, there’s a possibility the sprocket keyway isn’t exactly in the OEM location.
The heads have been cut and I believe he used thinner head gaskets thus bringing the cams closer to the crank axis which slightly changes the cam to crank sprocket timing.
I see a lot of different tolerances stacking up here on the motor build.
“Dot to Dot” really only works with OEM components within OEM spec.

I suggested that 30 or so pages ago, but nobody wanted to entertain that possibility. I really feel like there is a mechanical issue at play here. A simple compression test might give him a clue.

 

[GregO]

I suggested that 30 or so pages ago, but nobody wanted to entertain that possibility. I really feel like there is a mechanical issue at play here. A simple compression test might give him a clue.

You did, I went just a bit deeper in the weeds as most here don’t know the heads have been resurfaced and I’m most certain he used thinner head gaskets chasing a higher compression ratio.

 

[tdstuart]

You did, I went just a bit deeper in the weeds as most here don’t know the heads have been resurfaced and I’m most certain he used thinner head gaskets chasing a higher compression ratio.

Stock head gaskets

Sponsored