Struggling to find my error with gt350 throttle body

[Grim5.0]

just a weekend warrior here, enjoy getting to know the how and why and will invest the time to learn. By far the most complex ecu ive messed with to date. still reading and re-reading several years worth of posts on here. At least the labeling for the tq pid's seems to have become more standardized. Thank you for the clarification on the cause of the errors and what is going on/wrong. it is slowly sinking in.


Im raising my hand in math class here for a homework check. Would always prefer to calculate the answer vs shoot from the hip if possible. I am bad at math

The patent labels the atmospheric pressure as Pa and specifies it is to be measured in inhg for use in their equation. Is your labeling for velocity simply referencing theirs? or is it needing to be converted from inhg to pascals? i have it converting to pascals in the spreadsheet
here is a quick spreadsheet, effective area from the throttle angle area lookup has not been setup, will work it out if everything else is sound.
TB angle error to Load.xlsx

if i have this right, using the ETC throttle angle error will give us the TB angle we are off by and then using the manifold vac (ETC model) we have our x and y lookup on the effective area table in the TB model. From this we arrive at a known effective area error for a given angle/vac. plug this into the aforementioned mass flow equation and we get the mass flow error further converted to a load error. Using this load error, the rpm matching the angle error in the log and the mapped points we can cross reference in a given torque table what that tq error actually is, or approximate. From there that is applied to the DD table at the respective pedal % and rpm. If im wrong on this, im scratching my head on how the DD gets worked to a tb angle/tb area from tq

I copied the graph pictured above to calculate the angle error from actual and desired but the results are quite different when plotting vs using ETC throttle angle error. seems the polling speed and then count per cell can really throw this off, higher counts drift to zero . i realize i did not have the etc angle error selected though on that log and have added it going forward.

[murfie]

Yes it needs to be converted to Pa.
1 (pascal / kilogram) * (meter^3) = 1 m^2 / s^2

OR not converted:
1 (inch of mercury / kilogram) * (meter^3) = 3386 m^2 / s^2

Inch/ min is the velocity we are ultimately trying to get to. So whether you take the square root of it converted or not, it turns into an amount of m/s, and its just converting that to inch/min.

The whole torque model is usually based on a simple relationship like 1.0load = 370ftlb of torque, so .1 load = 37ftlb of torque, etc.. Don't need to make it more complicated than that.

[paulsmithsr]

Ok I have 2019 gt 10 speed and want to install a gt350 manifold and throttle,body what all should I change
I have a bullit tune for 2019 but it’s a 6speed. - should I just copy and paste throttle diff and leave everything else alone ?

[murfie]

Ok I have 2019 gt 10 speed and want to install a gt350 manifold and throttle,body what all should I change
I have a bullit tune for 2019 but it’s a 6speed. - should I just copy and paste throttle diff and leave everything else alone ?

There's little to no gain going from a 2018 manifold to a GT350 manifold, more of a PIA to get the IMRC to work properly. I would suggest forgetting about doing it and porting the 18 manifold if you want to modify something.

[paulsmithsr]

thanks for advise was gonna add cams and don't wanna do CJ intake and the gt350 is best for bigger cams and high rpms
I would just ;lock out the imrc

[murfie]

If you are doing cams in a 18+ I would be more worried about what to do about the DI pressure and flow rates, than the manifold/ throttle body.

[Grim5.0]

DD is a strike. Changes aren't reducing ipc error s and moves the oscillating TB around but doesn't get rid of it. Chased the ETC error and found etc error increasing dramatically at times after charges. Once the pedal would progress through that region quick enough the studder reduced a bit but never went away. Just to see how it would respond, I went barbaric on the DD and plotted diagonal line from 8% to 64%.
Not as expected!

Im going to give the Tq tables a few logs and see if it responds.

What is desired load based off of? I've only seen desired airmass talked about on an s197. Is that suitable data to compare air load too to calculate inverse table error and correction?
They mirror each other but start to drift apart during ipc error and quick throttle changes. Sometimes 0.1 or more at times.
Starting to map the inverse table airload in logs.

[murfie]

Your driver demand torque controls your scheduled torque that your desired load comes from.

When the ECU determines through limits, IPC, or what ever other reason that the throttle is not doing what the driver is demanding from the pedal, it steps in and takes control of the throttle via scheduled torque.
These limits, some mentioned in previous post, you may need to identify others, need to be moved out of the way so the driver demand is in control of the throttle and the ECU doesnt step in.
Not moving these limits and adjusting either the TB model or torque model, is just going to skew those models and introduce inaccuracy.

Posting tunes with the adjustments you made, and logs of those tunes will help identify what you are missing. It will also take a lot of guessing from lack of info out of this thread, as so many others are plagued by. Help the forum, help you.

[Grim5.0]

sure thing, appreciate the continued time. attached is the log from shutting off anti-shuffle, maxing ipc max/min tables respectively and only the DD adjustments made that you previously recommended. i put the DD back to whats on this log for reference in the attached tune.
STOCKOS-GT350TB.hpt murfs changes.hpl

Would you recommend shutting of clip/add or reducing it to a less involved state? I believe this may be learning and adjusting the DD table from bills explanation HERE

The change log will show lots of small changes back and forth as i experimented to see how things responded. The long crank on hotrestart was helped by increasing starting friction loss or by bumping the the load on inverse table 14 at the cranking tq/sub500rpm range. DD had no perceived impact on startup and at idle, but does impact the path the tq reduction is taking on the way back to idle.
I reduced the channel logs to try and improve the polling of the TB specific channels. If there is something missing, please advise.

quick snapshot of the DD being increased 30ftlbs higher than stock and etc error increasing as well. DD increases, like the one pictured, push the EBT higher, even when matched 1:1 from error in a previous log. Stock DD table pictured bottom right for reference to see the scale of the changes and impact
DD & etc error increase.png IPC decreased but etc error increase.png


if desired is the end tq to load calc, cant we use this to also calculate how much we are off on the DD as well? The workflow shows the additions of the errors/losses but we are assuming all of those adders are accurate from the factory, and the only error is related to our start at DD.
eng control flow.png

Calculating backwards from TB angle to useful load adjustment gets crazy small as the pressure across the blade nears equalization with atmosphere. But the etc angle can be quite large! im wondering if it is best to use something closer to "actual pascal" for the velocity calculation vs the ETC model manifold from the logs. as the etc model calculates near zero pascal, the actual airflow is at or nearing 101325 pascal as it enters the TB, am i in line here?

[murfie]

The way you are looking at ETC torque in the channels list, compared to the calibration DD table is wrong. I see it following driver demand using a graph and that the only way anyone should ever compare the two. A technique I use for better graphs is the export function, I find it works way better to export and use the view all data rather than the view zoom data only. This is also why a snip of your scanner and the calibration together is entirely useless to help anyone with.

DD torque vs log.PNG

You were not logging anything to determine IMRC position, but looking at MAP vs load, and seeing torque increase above 4000rpm, it looks like the IMRC transition from closed to open is not smooth. Lower your optimal stability IMRC threshold to between 1200-1300RPM, And the hysteresis 200-300RPM. It could also be an adjustment needed in the the Best fuel economy load for opening/ closing tables. for opening Increasing at distance 6,7 and decreasing at distance 8. If you compare the GT350 to a GT tune, keep in mind the GT350 goes from closed to open at distance 5/6 and the GT does it at 7/8. Probably adjust the best drivability ones as well to keep things consistent. that mode is a very small window.

IMRC fuel economy.PNG

[Grim5.0]

Hopefully this is useful, turned CLIP/ADD off and had to adjust the max ipc error to 36000 (double) to avoid limp mode in the tune above.

no clipp-add.png ipc err.hpl

EBT is above DD and TT. can adjust the TB model to reduce EBT per thatwhite5.0 or increase both DD and then TT to match ebt.
3 degrees etc error, 0.018in^2 at 1189 pascals(ETC model) and 2100 rpm is roughly 2ftlbs error on the DD table, which seems terribly small.

edit: will adjust the imrc settings as indicated and add the imrc positions. i do believe that is the main area thatwhite5.0 zeroed in on for the tq tables as they have a dip from one to the next during the transition.

[veeefour]

thanks for advise was gonna add cams and don't wanna do CJ intake and the gt350 is best for bigger cams and high rpms

2018 and GT350 - identical manifolds peaking around 7200-7300 rpm.

No such thing as bigger cams for Coyote - longer duration and aggressive lobe profile is everything you can get.

Now imagine you mani peaks @7200 rpm and your cams want to peak @7500-7800 rpm. Manifold shapes the power band.
Your engine want's to suck more air at rpm where your power is already screwed up. It's like revving your stock mani beyond 7000 rpm - you can but there's no gain to use.

To summarize: aftermarket cams in coyote + any other manifold besides Cobra Jet = complete waste of time and money.
This is where those "cams in yote sucks" stories are coming from.

[paulsmithsr]

Comp cams make stage 3 cams and they add more power -I agree the cobra jet is better, but the gt350 (larger runners) ,larger volume And bigger opening for throttle body is better then the 2018 - with stage 3 comp cams stock- 2018 made 592 hp at 7500 rpm
The cobra jet made 609 hp at 8000 rpm twin 69 mm throttle body
the gt350 will be in the middle of them
With stock cams the cobra jet made 550 hp
so cams added 69 hp —- 232 (intake), 234 (exhaust)

[Eric@HPTuners]

The Ford Racing cal attached is for a GT350 TB on a GT. This data seems to work well.

[paulsmithsr]

Wow thanks need it for 2018

[Grim5.0]

paul the bullitt tune someone posted for you has a different TB table compared to the known gt350 ones like what eric posted. cant find a clean/stock Manual my18 tune, but compared to the tuned ones floating around, appears ford tuned the tq tables and base spark tables as well as SD. dont recall what DD was but you might want to confirm the bullitt TB is in fact a gt350. the 350 TB's are analog and im not sure what they do to use it in place of the digital TB on the 3rd gen 5.0s.

dd wtiming 2.hpl dd wtiming 3.hpl 3DDSTOCKOS-GT350TB.hpt DD changes.png

2 logs attached, one before and then after changes. Also attached is the adjustment to the DD from the first to second log. Turned off CLIPP/ADD and pushed the ipc max out of the way, this was done prior to both these logs. have not touched the imrc yet as after turning off the clipp/add it revealed etc/ipc errors in low pedal positions prior to the imrc opening. i can certainly feel the changes to the DD in the pedal, and getting throttle hangs but struggling to see the changes in the etc angle error. IPC does appear to decrease but instead of anti-shuffle/oscillation, torque control steps in and cuts timing! Have been applying etc errors to the respective gear tables using speed/gear but finding they are really really similar regardless. calculated tq from angle error using the previous discussed formula and reduced by 50% to avoid overshooting, used the etc model PA value from the logs.

There is a small hesitation if the throttle is stabbed but if driven like a normal person, there is almost no perceived stumble/oscillation with clipp/add off even prior to the DD changes. Will DD clear up rev hangs though? and the funky hot start that feels like vapor lock for a split second?

what is Throttle learn correction? with no clip/add its -100%, the frpp2 tune settled to -194%

[paulsmithsr]

The tune was for a 2017. Not a 2018. — the 87mm Throttle body ,I don’t think u can use on 18 and up
Did the 2019 bullit use a stock 2018 GT throttle,body ?

[Grim5.0]

im referring to the past thread where you were asking for a bullitt tune, or at least i thought that was you. thats a good question, ive not had the chance to inspect one under the hood

[veeefour]

The tune was for a 2017. Not a 2018. — the 87mm Throttle body ,I don’t think u can use on 18 and up
Did the 2019 bullit use a stock 2018 GT throttle,body ?

Bullit has 87mm TB similar to GT350 yet with different electronics - pre 2018 and post 2018 TB's are not compatible.

[Grim5.0]

Pretty sure its resolved. Logging and confirming.

It is DD. But frpp tunes the tables through ETC tq request on the Y axis. Freaking clever

Zero ipc errors and etc errors are decreasing the more I drive. Is that clipp/add doing that? Or leftover KAM or similar learned values being relearned?

95% there, long decel when in gear at throttle lift, 0% pedal and slowwww closing of TB blade.. any suggestions?
also idle is about 80%, hits 450-500rpm when returning to idle and never gets to target N. Thoughts?

would not have connected those dots with out all the good info these boards have from you guys. the DD ftlb (etc tq) is the ftlb on the tq inverse Y axis, match it to frpp. After being frustrated by the DD not responding as expected, applied the airload inverse correction that murfie outlined and noticed each time i was making a change to the inverse the actual airload was always ending up near one full row lower or close to it.

gt350tb-tq tune.hpt 95% there!!.hpl inverse scale.png