Extremely Weird Airmass Dip Under Cruising Steady State

[cmitchell17]

I made the necessary changes to this one for you to try. The VE may be slightly off. I changed your map boundaries to get smoother table transitions. Doesn't appear you have a sd os available. Wouldn't think the maf hanging would make any difference as it's still failed.

So I tried the tune file you sent. Everything seems to work good and smooth, except for when I started it up the idle overshot a little and went unstable then kind of damped out it seemed.

Now the problem is I am not making any boost haha. But I think it is because I just hooked up the vacuum port on my BOV, before I just had the vacuum port open to atmosphere. I got a tee fitting and tapped into the brake booster and ran that line to the BOV. Before I only had the wastegate bottom (not the top one since I think the top one is supposed to vent to atmosphere so the diaphragm can move up freely without compressing air inside of it) vacuum port hooked up, and it is tapped into the charge pipe coming from the intercooler into the throttle body so it sees boost and not vacuum. Before that I didn't have the wastegate hooked to anything and there wasn't any change in boost. I am thinking maybe the wastegate isn't even opening at all because if I remember correctly I think I only put one spring in the BOV and it was the weak one maybe, so I think all my boost is leaking out my BOV before it can build up.

But anyway I know that's not really a tuning question so I need to figure that out on my own haha, it's been months of issue after issue, but that's how it goes and how you learn your lessons. I've never really dealt much with forced induction stuff, besides diesels, but that had nice stock variable vane turbine boost control, until now so I am learning. White Truck Turbo Log 6 Flashed Cal File Ghuggins Made Changes To.hpl

Thanks Ghuggins for the help, really appreciate it.

[4wheelinls1]

I used to tune PD Superchargers 4-5 per weeks for about 3 years, I had this problem on a number of occasions and had to do the 2 Bar OS upgrade (all Australian OS have this) I had to re-tune maybe 5-6 cars before I identified the issue and used the OS Upgrade. Oddly I used to use the other product when they were first to market with VVE and don't recall ever having this issue, but that was 8-10 years ago.
I had customers cars for days as sometimes they might only fault like that once in 60 miles.

For the last 2 years I was giving support in Aust for HP for the local distributor, this came up as well with no solution from HP. You may have noticed in some OS the De-soot has now got a temperature variable, that was also because I identified that de-soot was not off when it shoud be off, that should be raised as well as de-soot turned off.

[07GTS]

did this still happen when the maf table was zeroed as well as filled with proper data even if maf not used ? thinking if it randomly looks at the maf and there is no value there it causes the dip in airmass but if u have a proper value there it was ok ? when SD was the prediction coefficients zeroed as well ? i zero them when full SD no maf at all encase they still play a role

[cmitchell17]

did this still happen when the maf table was zeroed as well as filled with proper data even if maf not used ? thinking if it randomly looks at the maf and there is no value there it causes the dip in airmass but if u have a proper value there it was ok ? when SD was the prediction coefficients zeroed as well ? i zero them when full SD no maf at all encase they still play a role

With GHuggins latest edits so far it has been good other than a slight idle overshoot at startup. He zeroed the MAF table, and slightly tweaked the prediction coefficients and changed the VE correction factor table and also returned the desoot mode to stock values.

Yeah you would think it would reallly mess it up if you zero the values but dosen't appear that way.

[cmitchell17]

I used to tune PD Superchargers 4-5 per weeks for about 3 years, I had this problem on a number of occasions and had to do the 2 Bar OS upgrade (all Australian OS have this) I had to re-tune maybe 5-6 cars before I identified the issue and used the OS Upgrade. Oddly I used to use the other product when they were first to market with VVE and don't recall ever having this issue, but that was 8-10 years ago.
I had customers cars for days as sometimes they might only fault like that once in 60 miles.

For the last 2 years I was giving support in Aust for HP for the local distributor, this came up as well with no solution from HP. You may have noticed in some OS the De-soot has now got a temperature variable, that was also because I identified that de-soot was not off when it shoud be off, that should be raised as well as de-soot turned off.

What is de-soot mode for, when I search for threads about it a lot of the ones with info don't come up anymore or they are lost.

[cmitchell17]

So bad news, it is still doing the dip and clunk thing. I took it down the highway, and that is where it does it the most after you start driving a little ways.

Good news is with GHuggins edits the idle is now as smooth as it was stock.

I also fixed the BOV, I accidentally put the lighter wrong spring in there so it was just blowing out under any boost, but unfortunately its still not making as much boost as it should, from the log it looks like its only making about 3 or 4 psi, when the wastegate spring is supposedly a 7 psi spring, but it is a chinesium wastegate so who knows.

White Truck Turbo Log 7 Flashed Cal File Ghuggins Made Changes To and Put in Stiffer BOV Spring.hpl

[GHuggins]

I haven't looked at the log, but your fueling is dialed in?

[cmitchell17]

I haven't looked at the log, but your fueling is dialed in?

So I started with the stock VVE (I have a 100% stock engine) and scaled a 210 Kpa VVE table and took the 100 Kpa row and pasted it to the last 210 Kpa row and multiplied it by 1.33 and interpolated it. So from there the EQ ratio error from the wideband was at most probably 8% or maybe 10% in the WOT areas, I had done about probably 4 or 5 iterations since then of copying and pasting by half to the VVE table.

With that said it seems I never get anywhere with adjusting VVE tables, its almost like there is some feedback loop and it gets more unstable and you're error just gets farther and farther away. It was the same thing with tuning GEN V VVE's. I kind of feel like once you get it to where its no longer cutting out and seems like it drives normal, then the further you go with trying to make it better the worse it gets.

Sometimes it feels part of the problem is I could never get the filtering functions to work correctly, so say I was trying to eliminate certain areas and decel cells, and have tried to make filtering functions that will look at certain PIDs and shift them and calculate slopes and stuff but nothing seemed to work, and since there is no way to really check what you're doing is right, you kind of just have to look at it and see if it makes sense.

It also seems like when you paste error to a VVE table, you end up just negating what you did because it ends up getting so smoothed out or lost in the coefficient generation that you don't end up making any progress.

So to answer the question, I feel like it is dialed in I think, but maybe not, I guess I could look at the narrowband trims and just see what they are doing. This stuff just feels like black magic, especially with no way to check your work haha.

[GHuggins]

The second it hits 100kpa going into boost the mg/kpa calc drops to 0 just as 4wheels said.

I made some changes to this one to correct torque model, fueling in boost and some minor other changes to try and get the OS to read up into boost better. Doubt it will help at this point, but worth trying.

I went through about every 09, 10, 11 and 12 model yr vehicles I had on file. Could not find a single one with this OS that had a 2bar custom OS option. It's like hpt just forgot about the trucks??? I could find them under other OS such as cars and what nots, but they only offered trans OS swaps. Was trying to find one that I could just swap over for you, but no luck. I know hpt isn't doing this anymore either.

You may just wind up having to go MAF, but like I said I've only ever seen this issue running the mixed systems. With the MAF it would always do what your seeing coming to a stop or with light throttle - like cutting the key off.

[GHuggins]

Make sure to do a write entire with the last tune too.

[cmitchell17]

The second it hits 100kpa going into boost the mg/kpa calc drops to 0 just as 4wheels said.

I made some changes to this one to correct torque model, fueling in boost and some minor other changes to try and get the OS to read up into boost better. Doubt it will help at this point, but worth trying.

I went through about every 09, 10, 11 and 12 model yr vehicles I had on file. Could not find a single one with this OS that had a 2bar custom OS option. It's like hpt just forgot about the trucks??? I could find them under other OS such as cars and what nots, but they only offered trans OS swaps. Was trying to find one that I could just swap over for you, but no luck. I know hpt isn't doing this anymore either.

You may just wind up having to go MAF, but like I said I've only ever seen this issue running the mixed systems. With the MAF it would always do what your seeing coming to a stop or with light throttle - like cutting the key off.

Thanks for the help GHuggins, really helpful.

So I did apply the TCS patch and then did a full flash, which I see in your last tune file you had the TCS applied, so I guess I don't need to full flash? So I haven't noticed the issue again, but I haven't really driven it far at all. So I tried a boosted launch and I think I blew the drivers side exhaust manifold gasket out. I put remflex gaskets in when I installed my turbo manifold and installed the supposed "high flow" driver's side exhaust manifold, but while installing that I noticed the rear most driver's side exhaust manifold bolt broke off in the head (apparently this is the most common thing in the world and the most common location for it to happen). So I drove it back to my house and put my smoke machine in the exhaust pipe, but I didn't see any smoke coming from the gasket on the driver's side manifold. But I did notice that my crappily rigged intake elbow boot had blown off my charge pipe, so maybe I didn't blow the exhaust manifold gasket which I hope is what is going on, but it sure sounds like a blown exhaust manifold gasket.

I am going to try to fix the charge pipe and pray that is the issue and I didn't blow the exhaust manifold gasket. My hands and back are killing me messing with this thing so much, but maybe I can get by, at least temporarily, with one of those brackets that screws into the bolt holes on the head above the manifold bolt holes and holds the manifold on.

If I can fix that I will try your changes. I have also ordered a bung and a Hitachi LS7/LS9 MAF as well, but I need to decide if I will use the IAT sensor in that MAF, or continue to use my "broken out" IAT sensor. I guess it depends on where I put it, which to me it would seem to be most logical to put the MAF as close to the throttle body as possible, this would theoretically avoid the problem of not metering the "missing" air that gets blow off through the BOV. If I put it right before the throttle body I shouldn't need my stand alone IAT sensor anymore.

[cmitchell17]

So finally got my blown exhaust manifold gasket fixed. I wasn't about to pull the engine, or pull the head off to try to pull out the broken bolt, so I got one of those bracket things that bolts to the head and pushes up against the exhaust manifold and so far its working.

Anyway, I full flashed the latest file Ghuggins made. So far it hasn't done anything weird, but now I have a different problem. Now its pulling an extreme amount of torque management advance. I also got a service trailer brake message, but I scanned it for codes like 20 times and I couldn't get anything to show up besides the permanent mass or volume air high code that has already been there.

Ghuggins so it looks like you made more torque model edits in the latest file, and the torque source while the torque management advance is coming on at WOT is "axel". This is the same dreaded thing that happens to me on GEN V stuff, except this is pulling timing down to almost zero at WOT.

So I don't get why we really even have to make edits to the torque model. I mean obviously GM didn't calibrate the torque model outside of the torque range of the engine, but I would assume the virtual coefficient model would kind of extrapolate someone accurately outside the stock torque ranges?

Also I just got my Hitachi MAF and bung in today, so I guess I will try to put it right before the throttle body. I'm not sure what calibration I will put in it, I guess ill start with a LS9 values. I guess this is all dependent upon how good I can braze the bung into the charge pipe haha.

White Truck Turbo Log 8 Flashed Cal File Ghuggins Made Changes to Number 3.hpl

[smokeshow]

Seems like its always 10 people all at once that run into the air calc mode issue. Seen it several times today already lol

[GHuggins]

Well, I didn't think it was going to cause that, but it was done as a test and if you check your TM settings - that's what it's hitting now because it's over calculating your engine torque output. Just call it a weird hunch. Go back to stock torque model and see if the problem returns. I just put a completely different fi caddi's tm in yours.

Oh and to save you from burning out your tranny down the road, your going to need to tune in your torque model. This controls all of your steady state pressures. This is why "we need to dial it in". GM doesn't interpolate anything on it's own in the lower throttle regions.

[cmitchell17]

Well, I didn't think it was going to cause that, but it was done as a test and if you check your TM settings - that's what it's hitting now because it's over calculating your engine torque output. Just call it a weird hunch. Go back to stock torque model and see if the problem returns. I just put a completely different fi caddi's tm in yours.

Oh and to save you from burning out your tranny down the road, your going to need to tune in your torque model. This controls all of your steady state pressures. This is why "we need to dial it in". GM doesn't interpolate anything on it's own in the lower throttle regions.

Bad thing with me about airflow and torque models is I tend to go all or nothing, meaning I want it 1000% perfect or I just want to give up haha, so I will end up wasting a lot of time for not much output.

Anyway, I got the MAF bung braised in, pretty crappy job, but its in. So I had a breakout harness, and for some reason I just decided to leave my broken out IAT plugged in, but I plugged in my MAF and checked the wire outputs from a LS9 to ensure everything was in the right position (MAF signal, ground, 12V power). So it seems to be working initially, but then the state changed from No Fail to MAF Failed. So I turned it off and started it back up a few time and kept trying to clear codes (since I had that permanent MAF code (which seemed like it was a circuit high but now its saying performance)) then it didn't find any DTC's anymore, the frequency outputs a signal and it defiantly responds to throttle so it seems to be working.

However, it is not calibrated of course. Once it went from the MAF Failed state to the No Fail, it immediately went super rich like .6 lambda, and the trims pegged to -25% and it was trying to die. So I pasted in a stock LS9 MAF calibration, but I am running a 3 in charge pipe tube, so I would imagine it will need quite a bit of calibration from the stock LS9 values. I am running a Hitachi LS3/LS9/LS7 MAF which was surprisingly cheap, only about $23 or so I think.

One thing I didn't think of was the straight (laminar?) flow, I put it in a straight part of the charge pipe but its only about 3 or 4 inches from a 45 degree bend, I guess it won't effect it too much.

And as for the torque model, couldn't I put in LS9 values? at least to start with? Or does everyone just extrapolate the rest of the curve at the higher airmass/airflows?

[4wheelinls1]

Seems like its always 10 people all at once that run into the air calc mode issue. Seen it several times today already lol

Yep, its not new.

[4wheelinls1]

Bad thing with me about airflow and torque models is I tend to go all or nothing, meaning I want it 1000% perfect or I just want to give up haha, so I will end up wasting a lot of time for not much output.

Anyway, I got the MAF bung braised in, pretty crappy job, but its in. So I had a breakout harness, and for some reason I just decided to leave my broken out IAT plugged in, but I plugged in my MAF and checked the wire outputs from a LS9 to ensure everything was in the right position (MAF signal, ground, 12V power). So it seems to be working initially, but then the state changed from No Fail to MAF Failed. So I turned it off and started it back up a few time and kept trying to clear codes (since I had that permanent MAF code (which seemed like it was a circuit high but now its saying performance)) then it didn't find any DTC's anymore, the frequency outputs a signal and it defiantly responds to throttle so it seems to be working.

However, it is not calibrated of course. Once it went from the MAF Failed state to the No Fail, it immediately went super rich like .6 lambda, and the trims pegged to -25% and it was trying to die. So I pasted in a stock LS9 MAF calibration, but I am running a 3 in charge pipe tube, so I would imagine it will need quite a bit of calibration from the stock LS9 values. I am running a Hitachi LS3/LS9/LS7 MAF which was surprisingly cheap, only about $23 or so I think.

One thing I didn't think of was the straight (laminar?) flow, I put it in a straight part of the charge pipe but its only about 3 or 4 inches from a 45 degree bend, I guess it won't effect it too much.

And as for the torque model, couldn't I put in LS9 values? at least to start with? Or does everyone just extrapolate the rest of the curve at the higher airmass/airflows?

You need to set the dynamic airflow to 500 rpm and dial in the MAF without the VVE impacting it, once the VVE and MAF return the same mixture and fuel trim set the predicted coefficients to 0 and return the dynamic to 4000 rpm.

Make sure if the MAF pipe size is differnet the frequency is within the MAX/MIN limits for faulting

[cmitchell17]

You need to set the dynamic airflow to 500 rpm and dial in the MAF without the VVE impacting it, once the VVE and MAF return the same mixture and fuel trim set the predicted coefficients to 0 and return the dynamic to 4000 rpm.

Make sure if the MAF pipe size is differnet the frequency is within the MAX/MIN limits for faulting

Thanks, and you mean that the MAX/MIN limits might need to be changed for the different (smaller) tube to make sure you know when the MAF fails right?

I am honestly thinking about just using the narrowbands to turn VVE and MAF at low throttle points, I feel like I am chasing my tail with a wideband. I mean the wideband output looks plausible and good, but when I run it through the histogram and plot EQ Ratio Error (the default HP Tuner's Math Parameter) it just seems like there is too many discontinuities and errors never seem to converge.

[cmitchell17]

So I disabled closed loop by chaging the ECT limits for the o2 readiness and moving up the dynamic airflow limits to force MAF usage and was logging the wideband error from the MAF. But the weird thing I noticed is I don't have any DTCs (I returned the MAF/airflow DTCs to stock) but the MAF sensor state PID still reports MAF failed but I still think it's using the MAF for fueling. I don't have hard proof of that it just seems like it is. The MAF signal is outputting fine it seems.

The air calc mode seems to bounce from "unknown" to "hi speed" and nothing else. I don't think I've seen any other state indicated. Does anyone know what unknown means? Does that mean that HP Tuners didn't know what the output meant so they put unknown? Or is it an actual output defined by GM/SAE and it literally dosent know how the ECM is calculating airmass? Seems kind of odd that the PCM wouldn't even know how it's calculating it's own airmass, but I guess humans don't understand how our brains work either haha.

[cmitchell17]

So I had to disable the P0101, P0106, and P0121 tests to keep the MAF from failing once I hit a random 4600Hz. Not sure why though.

I am almost convinced I am off by some order of magnitude on the error percentage. I feel like the smaller errors end up converging (within +/-3%) but the large errors (up to 50%, since I started with a random LS9 calibration) just sit there at the same 30-40% iteration after iteration. I can even manually blindly adjust them by 20% and nothing seems to happen. So it makes me think I am only applying a very small correction factor. This makes even more sense when you look at a MAF curve, since they are exponential curves if I am multiplying by a small factor that is an order of magnitude off, then I will converge on the very small values, but as the curve grows exponentially it gets a lot bigger at the higher frequencies and I will never converge then.

I am displaying the built in HP Tuners EQ ratio error and selecting the units as "percentage", so the histogram shows numbers between 0 and 100. Then I copy this error and past it by multiplying by percentage half.

I feel like I made this same error over and over again one time but apparently didn't learn my lesson, but I reviewed all the guide and posts and videos.