VTT tune theory, can tuning experts confirm this?

[snives]

Hey all, Just want to get a spot check from the expert tuners. I've done the VVE tuning, and idle tuning, and VTT tuning and its running pretty smooth for a stock GEN V 5.3, dod delete, with a mild cam, locked out at 0. Maybe its just me but I'm thinking it should be making more power, so I'm questioning my method of VTT tuning.

Should a properly tuned VTT table contain the exact Engine Torque [Sensor](N m) value when logged at that particular [Cam, Spark, Airmass, RPM] (for airmass table, and MAP for map table) condition?

-or-

Should each cell in the VTT table be Engine Torque + Zero Pedal Engine Torque or some other calculation?

[GHuggins]

Delivered torque will always be different than what's in those tables due to the airflow variable. You can use delivered to dial in the airmass side on it's own, however the VE offset will greatly change the map side.

You've made it this far. Set up histograms or tables or whatever they're called now logging the individual spark cells vs cyl air and map and then log delivered torque in those. After that plug those values straight back into the torque model. You'll understand what I'm saying after doing this depending on the VE and MAF change.

[snives]

Thanks GHuggins, yes, that's how I was tuning VTT, I setup a bunch of histograms for logging as you describe. I think the minor issue I have now is that if I don't get good coverage during a logging session, I end up with areas of coverage, and areas which aren't changed. If I don't manually massage the unchanged cells, then during the coefficient creation process the extrema in these cells influence the coefficients so much that it ends up losing much of the granularity of the newly logged data. This seems like we end up back to where we start to a certain degree.

[GHuggins]

It'll change. Even the airmass side before you're all said and done. You'll also notice throttle issues going away in most instances. Yes hand shaping is required unless you copy and paste each spark to excel and force the engine via the timing tables to run in those particular timing cells wile gathering as much data as possible for each. However you'll have to do this several times and if you aren't plugging the offset created by the VE table back into the map model each time it will cause the airmass side to start skewing.

[GHuggins]

I should probably add this in here to. I'm uncertain if "base torque" is what's defined in the torque models, but I'm pretty certain "delivered torque" is what's actually used after all of the drive-line losses and other filters are ran through it. GM's software already calculates torque, but the base is what's changed with the engine mods and is what then has to be corrected. We can then use delivered to get the general shape as it's already got all of the filters and losses run through it, but you have to figure out how to shift it and by how much using the airflow tables as offsets and then some to get things right.

https://patentimages.storage.googlea...03010427A2.pdf

[snives]

Yes! that's an interesting distinction we should find out, begging the original question. Should we log Delivered Engine Torque or some other calculation?

What I hear you say is the VTT Airmass side is fine to log Delivered Engine Torque but logging Delivered Engine Torque into the VTT Map side will be affected by the offset created by the VE table. I actually don't understand what this offset is, how to see it in the scanner, or calculated it in a math function. Are you saying we need to account for the losses on the MAP side but not the Airmass side?

Reading elsewhere on the forum people have had some success logging Zero Pedal Engine Torque and backfilling the Idle External Load table. That would lead me to believe Zero Pedal Engine Torque is modeling the parasitic losses in the torque model.

So let me try to lay this out, anyone jump in to correct me.

"Base Torque" (not a loggable parameter) - The torque the engine makes without regard to losses from friction, accessories, oil viscosity, and pumping losses.
Zero Pedal Engine Torque -The calculated sum of all losses due to friction, accessories, oil viscosity and pumping losses.
Delivered Engine Torque - The sum of all torque produced, as the result of demand, including all losses (net engine torque).
Engine Torque - Delivered Engine Torque adjusted in accordance with SAE J1349 adjusted for 0% humidity at sea level at 25 degrees C.

So using that logic Delivered Engine Torque = Base torque + Zero Pedal Engine Torque.
Leading us to use Base Torque = Delivered Engine Torque - Zero Pedal Engine Torque.

As a test I reverted back to stock VTT tables and then reapplied previously logged data filling VTT Airmass with base torque and VTT Map with Delivered Engine Torque. I did it this way because I saw the stock values in VTT Airmass seemed higher than in VTT Map.

Drivability seemed good, transmission shifts like stock, idle is also good for having a cam. Give it a try and let me know what you find.

[GHuggins]

I personally use a tool that I had commissioned or built just for this. If the author of said tool wants to share it, he's more than welcome to. These are some screenshots of the tool applying the correction factor I'm referring to. Problem is there's another side of this that the tool doesn't add in and I'm not even sure where to begin to do that. The MAF creates it's own offset. The MAF kind of plays straight into the Airmass calcs, so I log delivered and then apply the offset to that created by the VE table only. I then work the models by hand to make delivered the same map to map. Problem as we've discussed is the parasitic losses. You have to work both maps at that point to account for them. If you're sure 0 pedal is this particular "offset" then all you would need to do is add that into delivered and then the offset from the VE table on top of that...

Stock VE - VE Offset Tool.jpg
Tune VE - VE Tool.jpg
% Change.jpg
Stock VE Table with Offset.jpg
Changes Applied to Stock Ve.jpg
After Calculating.jpg

[GHuggins]

What I wound up with was this - So you can see how much things were moved to balance and get timing and torque where it was happy.

Deliverd Tq Map.jpg

Delivered Tq Airmass.jpg

Map Torque Model.jpg

Map Torque Model 2.jpg

Airmass model.jpg

[snives]

That's pretty cool. So you find the % difference between the stock VE table and the tuned VE table, and then use that as a % multiplier on the stock VTT tables? That is interesting.

I'm not sure Zero Pedal is all of the parasitic losses, its a more of a hypothesis. Learning from another tuner who says these L83 engines like to run with roughly ~0 ZeroPedalTorque, which conflicts with my hypothesis of what that variable is.

I've noticed my idle timing advance is still up at 24 degrees or so, so i'll have to take a deeper look at it.

[GHuggins]

So in looking at stock engine logs what I'm finding is 0 pedal is around 5 to 10 lbft and delivered is 10 to 20 lbft at idle... FWIW. The models are also really close to being balanced delivered tq wise within a few lbft for all spark groups.

Admittedly, I've had a lot of problems with the above posted truck getting it to idle right. If I shift delivered and 0 pedal down any at all then timing goes through the roof. Gets up in the 30's. I've had to really work the adaptives and external load tables A LOT to get it to smooth out as good as it currently is. Problem is it's running a ported 95mm tb. I honestly think the airflow isn't correct for throttle plate position. I wound up having to work the "background" etc area/rate table to get it halfway right, so might be why "its" numbers are a little off.

Do you have links for the 0 pedal write ups? I've never had much luck with 0 pedal being at 0 lbft. That always made them feel clunky because delivered starts getting into the negatives a lot and when you take off it's like they get a kick in the backside. Plus if trying to backup with any sort of load it might cause the tranny to slip.

[snives]

First want to reply to your issues (to see if I could help return the favor in any way) and then I'll get back on topic.

When I look at the stock VTT tables, I see the trend that there is more torque produced at higher spark advance, makes sense right? So, how I understand it the ECM does an inverse looks up of current [Airmass, RPM, Immediate Engine Torque Cmd] and gets a Spark Advance. It can do the same thing on the MAP side, and also gets a Spark Advance, and either chooses one, or averages or something but then there's your Spark. This is just my understanding and may be wrong, anyone please jump in here to correct me. So if the torque tables at each spark advance degree are similar, like you say, that would be a big problem, it would likely hunt and swing wildly from -10 to 40 in an instant because there is no gradient (does that describe your symptom?).

Regarding the 95mm TB, I'm sure you've checked it but there's Engine->Airflow->Electronic Throttle->Max Area.
If you're concerned about the airflow vs plate position maybe you could set up a histogram of MAF vs Commanded Throttle Actuator and see if its linear? There's Max % Area vs RPM where you could profile it, but for a single TB, MAF should really handle all that for us.

I have decent idle and have not touched my External Load tables. Okay back on topic.


Here's a link for that Zero Pedal tip, it was in a YouTube video from Matt Sanford at Skye Tuning.

So I realize the two ideas are not compatible, if we A) intend to have Zero Pedal Torque be ~0 and B) that Zero Pedal Torque models the parasitic losses. Since adjusting our VTT tables should not change Zero Pedal Torque if it were modeling parasitic losses. So, yeah, I guess try it both ways and compare????

[GHuggins]

The factory models are the ones that agree and are completely balanced spark group to spark group. Also what I've tried to strive for and what putting in delivered and correcting delivered should be doing for you. There are other controlling tables that go with spark control such as closed throttle torque request. Both models should agree spark to spark. What I believe is the OE only dials in the idle areas and then how they shape vertically from there is how they shape which is why you see them go so high in the higher loads. If you look at the offset application pics you'll notice they have a "swing" to them where the idle area is higher. To keep the idle area right I would have to take a straight edge and draw out the models from there.

I'm willing to listen to others as I'm always willing to learn, but what I find is usually they don't dial half what they need to in and have their own problems as far as the youtube vids goes. Plus as you're talking about, no one knows for sure how all of this is defined other than maybe Smoke

Yes, I'm well aware of the etc area max table. This is the one I'm referring to having to adjust for this particular truck. Basically defines throttle body airflow based off of it's tps position. I lowered it about 4% for the idle tps area as even the OE seems to lower this for bigger tb's. Plus the throttle needs to close more to flow the same air vs a stock tb that idles at 20% opening.

Throttle area.jpg

EDIT - Watching Matt's vid right now. I can almost guess who tuned the truck originally by what he's finding so far Of course I've never seen him set the coefficients to one value like that, so who knows for sure.

[GHuggins]

OK, a few minor comments on Matt's vid. Stock trucks idle between 15 to 20 degrees of timing, not 0 to 5. Not sure where he got that unless he was thinking of a car instead? Map was pretty high in this one with his tune. He's got the torque model set up to hold the throttle open which is why timing is low. This does a couple of things. Helps fight reversion and it helps fight idle dips. Problem is, it can shift and often for the worse when doing this but I think he would have had some feedback on that. One thing he didn't point out is when you make a big change to things like that it can often take a little bit of driving to settle out to it's "final" numbers. I don't like 0 torque at idle. I do a lot of tunes for truck owners some of which tow heavy loads. 0 torque or even negative torque causes problems. I like something around 10 lbft in neutral or park more like the OE has them. I understand a cam won't need this much, but it seems to work out. Then if you know how to adjust it, you can indeed get fueling to nearly perfect all the time on these. Otherwise, the man knows his shit

[DSteck]

It needs to be edited as spark vs airmass and spark vs MAP at each relevant RPM breakpoint.

That said, without a dyno and knowing all of the intermediate losses, there’s no prayer of getting it right. The sad reality is torque modeling is not something that’s going to be accurate without equipment and a hero sharing the gray area that happens between the raw models and delivered axle torque.

[GHuggins]

All too true. The tools I use help, but still requires a lot of hand shaping. I originally "thought up" that tool based off of what Jake shared with how the models are developed and how he "helped" his. It still only "somewhat" puts the models back to a "factory" setup. Doesn't actually dial anything in due to all those other players but I'm happy with how it helps. Hand shaping is required from there. Plus there are a butt load of background tables that affect the idle. I will say plugging delivered back in really helps with trans life especially on the pressure controlling airmass side. FWIW the trans temp multiplier for line pressure has been found for the 6speeds. Don't know if hpt will ever put it in? BUT yes, after lots and lots of research and digging, torque can only be plugged in via a dyno.

[DSteck]

The dyno is only part of it. There’s still the gray area.

Once HPT accepts the CAN output of my dyno, I’ll be able to do it in totality with high confidence as opposed to just making one model match the other. I need a number of DMRs added to the scanner though first.

[GHuggins]

What about the other controllers or are those the gray areas you're referring to like pedal commanded torque? I assume you'll be using an engine dyno with a special system running the engine to get around all of this?

As for the average Joe, I know you can get wot numbers fairly accurately. Just requires logged time, tire size, gear ratio, trans ratio, tcc slip, trans slip and rpm. That's how virtual dyno's work, but this isn't good for anything other than wot. Delivered is good enough for that. Idle and off idle is where everything's a pain.

[DSteck]

I will be able to go from axle to airflow torque with high confidence because I know the gray area. I’ll be doing it with a chassis dyno (Mainline Prohub). The airflow models aren’t putting out delivered torque; delivered comes after friction/inertia losses from the accessories, blower, and internal forces.

[GHuggins]

From all my logging and investigations airflow adds to or takes away from delivered, but otherwise yes to the other factors.

So you're going to be using driveline losses then adding friction losses to that to get the base model back correct based off of airflow. Must be nice having all of those other numbers

OK, so just asking here. Is there anything that can be logged to show "just" the friction losses?

OR could you or Jake better explain what the different torque readouts are that the scanner is showing such as 0 pedal, predicted, axle and so on and maybe perhaps what might directly control these?

[DSteck]

Nothing essential is available to log yet. I’m not using anything already in the scanner, either, as it isn’t explicitly useful.