2012 Ram 1500 with Magnuson throttle issues

[tdebeer]

Hey guys. We fitted an 8 speed transmission in a 2012 Ram 1500 that had a 6 speed before. In the process we had to install a new PCM as the old NGC was locked by whoever supplied the PCM and supercharger in 2012. We could not read it either. We asked Magnuson to supply a base tune as it was their supercharger and paid $400 for it.

It seems to run fine on WOT 100% throttle down runs, but is a little rich on the AFR at around 10-10.5

But as soon as you lift off to approx 70% throttle it pulls timing and TPS starts closing itself. AFR goes to 10 and it backfires and splutters...

Really frustrating issue. I have tried copying over a whipple that we did but cannot get rid of this issue

Can anybody help PLEASE

I will upload a log and the tune files

[luciddeath]

What tune are you running on the log bud?

[luciddeath]

Looking at the run, one thing to consider is fueling is off pretty good all the way through. I would attempt some nn training on it to start. get fueling about 15% closer or maybe the injectors are off by a bit, didn't really look too closely.

[tdebeer]

Running "Magnuson with throttle lowered" tune on that log.... I have tried hundreds... pulling my hair out here.

Fueling is rich yes, Should I just change stoic or FA table to start with?

From looking at other posts it seems this is a fairly common issue to get them smooth after supercharging it... The issue seems to be the TPS closing because of some torque or other limit it is seeing

[tdebeer]

NN Training? how do you do that? Sorry if its a silly question but this stuff seriously above my pay grade...

[drumnbasslvr]

I have never run NN on a vehicle that began life as a naturally aspirated engine and now is boosting. I make things simple to start. Turn off that NN, blanket the entire PE to 0.0192 for 11.5 AFR, then simply tune the airflow VE table to get things in check, use the long term feedback to handle your trims part throttle, use wideband afr error to correct your wot fueling, (AEM OBD2 4.9 wideband works brilliantly for this), or use yours, fix your injector data, pretty stock for numbers at a glance, multiply your knock tables by 1.2 to start, ensure you are never below 5 degrees in wot ever, change your wot voltage to 2.7 ( I changed a 2015 challenger 5.7 whipple tune today to 2.0 which helped, they build boost instantly so all the better ), change that voltage in power enrich and torque mgmt, put your throttle voltage stuff back to stock. Log more parameters! Your log looks like a whole bunch of just enough but not enough to diagnose well.

PS I'm not touching that neural network trainer yet because, yes it looks cool, but comes with zero instructions.

[HaasExp]

The biggest issue here is the insane amount of spark retard that's being commanded especially at WOT. The reason seems to be related to knock as the total knock retard parameter maxes out at -16 degrees in some areas. But even then it doesn't make complete sense as you are only commanding ~0 degrees timing overall with severe "knock" apparently occurring at a PR of only 1.2-1.3.

You still have the neural network (NN) on so that would have to be trained for such a large change to the engines airflow due to the maggie. Having the NN still on might be causing these weird issues.

The most common method used here to tune with a S/C is to disable the NN and tune by the VE tables and associated correction factors.

[tdebeer]

Thanks for this! I will attempt this step by step and see where I end up. Your explanations makes sense. Its just frustrating that we pay guys like Magnuson for this tune and we are this far off.

[tdebeer]

I am definitely going to try using VE today. Thanks so much for the feedback so far.

The biggest issue here is the insane amount of spark retard that's being commanded especially at WOT. The reason seems to be related to knock as the total knock retard parameter maxes out at -16 degrees in some areas. But even then it doesn't make complete sense as you are only commanding ~0 degrees timing overall with severe "knock" apparently occurring at a PR of only 1.2-1.3.

You still have the neural network (NN) on so that would have to be trained for such a large change to the engines airflow due to the maggie. Having the NN still on might be causing these weird issues.

The most common method used here to tune with a S/C is to disable the NN and tune by the VE tables and associated correction factors.

[luciddeath]

I tune a ton of hemi and boosted pentastars. i can't afford to turn off the NN with pentastars.. though you could just easily do the adjustments on the ve with nn and vvt off.. but if its got a stock cam it wouldn't be worth it

[tdebeer]

I took your advice and also HaasExp input and started fresh with basics in.

I used the injector data supplied by Magnuson in their tune....

Truck is running WAY better now and transmission changes perfectly. No more weird throttle changes and spluttering. Its running quite rich throughout even at idle at 12.5 on the AFR and I can hear some ping at the bottom of the rpm range and can see it pulling KR when I hear the ping. I can also say we changed to a slightly colder plug (NGK 6509 - LTR6IX-11) and in the process found that the right side rear coil pack was touching the firewall (Its a Right Hand Drive Converted Ram and the firewall gets modified to fit the pedal cluster etc) and I am 100% sure this was causing a false knock reading under torque as the engine moved and touched the firewall.

So attached is the latest log and tune... How do I get this AFR to 14.5? Am I missing something? If I calculate my Stoich etc we should be there?

Fresh Tune no NN Log.hplFresh Tune no NN Ram Magnuson.hpt

I have never run NN on a vehicle began life as a naturally aspirated engine and now is boosting. I make things simple to start. Turn off that NN, blanket the entire PE to 0.0192 for 11.5 AFR, then simply tune the airflow VE table to get things in check, use the long term feedback to handle your trims part throttle, use wideband afr error to correct your wot fueling, (AEM OBD2 4.9 wideband works brilliantly for this), or use yours, fix your injector data, pretty stock for numbers at a glance, multiply your knock tables by 1.2 to start, ensure you are never below 5 degrees in wot ever, change your wot voltage to 2.7 ( I changed a 2015 challenger 5.7 whipple tune today to 2.0 which helped, they build boost instantly so all the better ), change that voltage in power enrich and torque mgmt, put your throttle voltage stuff back to stock. Log more parameters! Your log looks like a whole bunch of just enough but not enough to diagnose well.

PS I'm not touching that neural network trainer yet because, yes it looks cool, but comes with zero instructions.

[luciddeath]

To answer your response @drumnbasslvr the nn trainer is just virtual volumetric efficiency. It's just like tuning the ve.. but it's to adjust for different cam positions.

@tdebeer what injectors are you tuning for? i'll take a look at the file and let you know my 2 cents at least..

[luciddeath]

@tdebeer if you're going to tune the VE you need to turn the VCT off completely on all drop boxes under Variable Cam... Otherwise it will be running off the NN still.

[tdebeer]

I believe they are Deka Injectors supplied by Magnuson in 2012 when we installed the Supercharger... Hope that helps? I am trying to confirm this now but its midnight here now so can prob pull one out tomorrow and check.

Do you suggest we try and activate NN again? I had some serious issues running with that though.

To answer your response @drumnbasslvr the nn trainer is just virtual volumetric efficiency. It's just like tuning the ve.. but it's to adjust for different cam positions.

@tdebeer what injectors are you tuning for? i'll take a look at the file and let you know my 2 cents at least..

[tdebeer]

Magnuson V3 Stoich 720.hptMagnuson Stoich 720.hpl

@luciddeath thanks we seemed to have found the weird acting throttle / fueling issue after logging with more PID's. I realized that its running into an overtemp situation on the cat and exhaust temp sensors so I raised those (The truck has no cats)

Now we are back to the original tune with NN on and its ok, but its literally cruising with 5 - 8 degrees of KR at around 2000 rpm and part throttle... comparing the tune to a whipple truck we did, it actually has 5 to 10 degrees less timing in that range so its really bad. It seems to run at 14.5 to 14.8 on the AFR then and I would think 13.5 would be better as you are definitely getting some boost there? How do you raise that? I tried changing the Stoich to 0,0720 instead of 0,0688 but it made zero difference? Why does it not change? This truck might get burned soon

To answer your response @drumnbasslvr the nn trainer is just virtual volumetric efficiency. It's just like tuning the ve.. but it's to adjust for different cam positions.

@tdebeer what injectors are you tuning for? i'll take a look at the file and let you know my 2 cents at least..