TDStuart Tuning Adventure

[tdstuart]

Updated the Auto SD Calculator to include a 3d graph based on rpm, tb map, and cyl airmass. This should help show you how your data is looking for the rpm range you selected. Also added target rpm and range options. Set your target rpm and range (targetRPM - range, targetRPM + range), then hit these min/max buttons to update the min/max rpm. Yes I was that lazy haha

I've been thinking about maybe adding the ability to view the data with the map the car is using to see how it compares.

Sponsored

 

[markmurfie]

Let's take a snapshot of the last log:

At that moment in time, the SD-model MAP is 7.9 psia.
At the same moment, the TP Model MAP is 21.22 inHg, which is 10.4 psia. Let's assume the TB model, baro, and MAF are accurate, therefore this number is accurate.

The SD-model MAP is used to determine the throttle blade angle. At that time, the throttle body deltaP is 14.7 - 7.9 = 6.8 psid. So it moves the throttle to an angle where it expects to get 19.04 lb/min airflow at that dP. However, the resulting airflow is only 14.24 lb/min.

If the SD-model MAP were 10.4 psia to match the TP-model MAP, then the throttle body dP would be 4.3 psid. With only 4.3 psi pressure difference, the throttle body would be commanded to open more to get the desired airflow, and the MAF would match the desired MAF better.

Therefore, if the SD-model MAP were higher, to match the TP Model MAP, then the throttle would open more, and the MAF would match the Desired Airmass closer and you're make roughly 25% more torque at that pedal position.

Doing some quick and dirty flow math, if you ratio the square root of the SDdp and the TPdp, and multiply by the 14.24 lb/min, you get 19.3 lb/min, which is very close to the 19.04 it desires. That backs up the case that if the models match, then the flow would also match.

In summary, I think the SD model needs more work. I do believe the TP model and SD model MAPs can be tuned to match well up to about .5 load at low rpm and .8 or so at high rpm. At that point, opening the throttle further won't result in any more load.

The problem is you can increase the pressure the MAF translates too, and we have right up to baro metric and even more 32inHg vs fixed at 30 baro, but you can't tell it more MAF than the sensor is reporting is there. So the only way to correct it is bring the Des. MAF down via the torque request, or limit the airmass the SD covers, which is what Im recommending as we are trying to correct the issue with the SD model.

 

[engineermike]

The problem is you can increase the pressure the MAF translates too, and we have right up to baro metric and even more 32inHg vs fixed at 30 baro, but you can't tell it more MAF than the sensor is reporting is there. So the only way to correct it is bring the Des. MAF down via the torque request, or limit the airmass the SD covers, which is what Im recommending as we are trying to correct the issue with the SD model.

I see what you're saying, but that applies to the [many] times that the load won't achieve demanded even though the throttle is mostly open and actual MAP is at or near baro. The example snapshot I posted only has the throttle at 35% open and MAP is well below baro, so it's not limited by mere baro or terrible VE at that moment.

 

[markmurfie]

I see what you're saying, but that applies to the [many] times that the load won't achieve demanded even though the throttle is mostly open and actual MAP is at or near baro. The example snapshot I posted only has the throttle at 35% open and MAP is well below baro, so it's not limited by mere baro or terrible VE at that moment.

After about 3100RPM all points from both the throttle model and calculated MAP are from the 100% VE slope for the calculated MAP and Barometric pressure is limiting MAP for the TB model. This is what my post showed. Your snip shows 4000+.
It's a disagreement on the airmass side as it seems to be switching to the 100% VE slope not staying at Barometric pressures with the throttle body model. I'm not sure what the logic in the ECU that says that a lower calculated MAP from the 100% VE slope is more appropriate than using the Calculated MAP max or Barometric pressure is, but that's what it's doing.

 

[engineermike]

After about 3100RPM all points from both the throttle model and calculated MAP are from the 100% VE slope for the calculated MAP and Barometric pressure is limiting MAP for the TB model. This is what my post showed. Your snip shows 4000+.
It's a disagreement on the airmass side as it seems to be switching to the 100% VE slope not staying at Barometric pressures with the throttle body model. I'm not sure what the logic in the ECU that says that a lower calculated MAP from the 100% VE slope is more appropriate than using the Calculated MAP max or Barometric pressure is, but that's what it's doing.

I believe you are correct that it's switching to a different calculation rather than the quadratic output, which is why the outputs don't match.

So, using the same snapshot I posted earlier:

I plugged in tdstuart's quadratic coefficients into hptuners and used the SD calculator to see how its output compared. At that rpm, load, and mapped point, the hptuners output MAP is 15.68 inHg, or 7.7 psia, which I believe we can assume is close enough to the logged 7.9 to confirm that the hptuners SD calculator uses the same logic as the PCM.

I found a few data points in tdstuart's log to know that the MP19 SD model is accurate at low loads around .11 but inaccurate at .5 at the 4350 rpm row. I went on to iterate on the quadratic coefficients until the SD calculator output matched the TP-model output in the snapshot:

The coefficients it took to get this output are:
quad: -10,000,000
slope: 23,500
offset: 0

For reference, the starting point was:
quad: 12,039,820
slope: 12,436
offset: 3.6

@tdstuart what markmurfie is saying is that there are actually multiple calculations going on in parallel to determine the MAP, and there is some logic there that chooses which to follow. The breakpoints in the logic are unclear.

Yours isn't following the ax^2 +bx+c at the higher loads, which is why the TP-MAP doesn't match the logged MAP even though the raw quadratic output says it should. The HPTuners SD model appears to match the Ford MAP output. So, I used it to come up with the coefficients above. It wouldn't hurt to plug them in for the 3950 and 4600 rpm rows, then log it while varying the load in that rpm range to see if it makes the numbers agree better.

 

[tdstuart]

Not a super long log but it looks like the load limiting worked.

 

[K4fxd]

What do you mean by load limiting?

 

[K4fxd]

Desired VS actual load is way off at least at this point

 

[K4fxd]

Again I'd get rid of the cam channels, they are just slowing down the PID's you actually want.

 

[markmurfie]

Desired VS actual load is way off at least at this point

Correct, load, MAF, desired load, desired MAF they are not close, but look at calculated MAP and estimated MAP form the TB. You are not fixing those relationships with the SD, and I have tried saying that a while ago.

I'd call the SD for MP 19 good. Unless you hook up a sensor and find both are way out.
Move on to TQ tuning and then spark tuning.

 

[tdstuart]

What do you mean by load limiting?

Setting the SD maximum load tables to what @markmurfie suggested.

 

[K4fxd]

@markmurfie

So now that the SD is correct, what is the process for fixing the load discrepancy?

 

[markmurfie]

@markmurfie

So now that the SD is correct, what is the process for fixing the load discrepancy?

Modify driver demand to what the ECU is saying engine brake torque is or vise versa.

Driver demand should look like a dyno graph. I'm referring to the torque line, NOT THE HP line as it seems like people confuse what I mean by saying dyno graph. But at this point the basic "fix" is to just modify the DD table or the TTL/LTT calculated torque model, so they match.
Many people will argue that the proper way to do it is leave the DD stock and only modify the calculated torque model, which works, but I don't think it accurately represents reality, and that's what I like to set my cars up with unless table access or some limit doesn't allow it.

 

[tdstuart]

So I am thinking I should figure out how I want to setup my mapped points so I can know what mapped points to tune.

For example, I was just testing (with imrc locked open), having MP14, MP15, MP19, MP20, MP21, MP22, MP25, and MP26 enabled.

here was my settings for FE mode:

I copied the blending and just have it set to blend between MP0 (MP14) -> MP7 (MP19) -> MP0 (MP14). I like how it drives, I left MP15, MP20, MP21, MP22, MP25 enabled as they were either being used for blending while in FE or are used in OS/OP. I enabled MP26 to see what it was doing.

I haven't touched snap lines yet because I'm not 100% sure how they work. What does MPOP do and is there any reason to use it?

One thing I was thinking was to change MP25 from -20,20 to -20,30 and change MP26 from -20,35 to 0,30. That way the car should mostly blend MP25 and MP26 as it goes from -20,30 -> 0,30 during OP.

I wonder how hard it would be to keep imrc enabled. Would I just have to add MP0, MP2, and MP7 (The main MP that would be used during cruise and idle)?

So then the final would be:
MP0, MP2, MP7, MP14, MP15, MP19, MP20, MP21, MP22, MP25, MP26

Still a lot of points and maybe I could make some optimizations to cut down MP20, MP21, and MP22.

Let me know your guys thoughts or if there is an easier way to do all of this.

 

[tdstuart]

Modify driver demand to what the ECU is saying engine brake torque is or vise versa.

Driver demand should look like a dyno graph. I'm referring to the torque line, NOT THE HP line as it seems like people confuse what I mean by saying dyno graph. But at this point the basic "fix" is to just modify the DD table or the TTL/LTT calculated torque model, so they match.
Many people will argue that the proper way to do it is leave the DD stock and only modify the calculated torque model, which works, but I don't think it accurately represents reality, and that's what I like to set my cars up with unless table access or some limit doesn't allow it.

So my next step for tuning the mp would be that?

Sponsored