Virtual Flex Sensor Issues

[HackAbuse]

Has anyone seen a factory virtual flex fuel setup go awry before?

I had a customer bring back a 6.0 GMC Sierra today that was tuned last year.

The engine was pinging, mostly under medium load, low rpm. After scanning it, we found the fuel trims way out (~-13%) and the ethanol percentage at 70%, so the truck was trying to run WAY more spark than factory when the customer had 87 octane fuel in it. Everything on the truck was stock otherwise and the only changes to the engine tune was more aggressive PE mode. Has anyone else had something similar happen? Could the MAF have caused this?

[5FDP]

Yeah, seen it happen before.

I've had to reset the composition with a scan tool to make the vehicles run good again. Never found the exact cause but it's a somewhat common issue if you search the internet enough. Many times on the tahoe/suburbans and sometimes the 1500 series trucks.

If he's not using flex fuel, just disable it. The virtual sensor is pretty poor at what it does but it's just using the fuel trims to determine the alcohol content.

[HackAbuse]

I'm worried it may have been too late and he's already shattered a ringland. After disabling flex fuel, it still has a rough idle, misfire and is down on power.

[jpierro79]

Ok run a compression check for an oil downed cylinder. When a cylinder has a failure it will read higher compression as the oil build up will cause higher compression. A ringland failure will cause low compression. Also maybe add a water remover if you dont find any compression down. If you dont find compresssion down you might have an injector failure causing lean on any single cylinder. let it idle with 1/2 quart Marvel mystery oil in case of locked lifter.This works well to in fuel but takes about 45 minutes if you have a jammed injector. dont go condemning ring land failures. The KISS method KEEP IT SIMPLY SILLY. 99 percent of all repairs are less complex than we make them out to be. 20 years of repair leads me here. Dont forget about DOD failure will cause same misfire and pinging as it will not allow ANY airflow and the misfire will shake the truck more violently than normal misfire. I am not saying your wrong but on any ls pull the valve cover and look for rocker movement during idle. He may have been pinging due to bad fuel or possibly DOD failure which unlike normal misfire causes severe fluctuations in vacuum.
This is just me id go
Check compression
Check Ignition wires
Check Fuel injectors
Check fuel pressure Could be lean causing pinging
Check DOD

Then condemn the bottoms end. I am not trying to tell you what to do but dont go to worst case scenario.
Yes ive been to worst case scenarios but they are more rare is gas vs diesel.
i had a guy w a 97 f250 with 7.3 powerstroke I kid you not after 320k nightmare of problems
Injector driver module bad no caused start
changed IDM then Bad connectors on passenger side injectors causing 157 failure codes under injector test
clyinder number 4 dead injector completely and after tear down 12 bad valves half sunk into cylinder heads. We were going to do a reseal but after inconsisten compression checks it was certain to need rebuild.
I also had 4 corroded connectors on IDM module which cause initial failures. These are RARE situations.

[Ben Charles]

Virtual system is crap in short...
Always use a sensor

[10_SS]

Has anyone seen a factory virtual flex fuel setup go awry before?

I had a customer bring back a 6.0 GMC Sierra today that was tuned last year.

The engine was pinging, mostly under medium load, low rpm. After scanning it, we found the fuel trims way out (~-13%) and the ethanol percentage at 70%, so the truck was trying to run WAY more spark than factory when the customer had 87 octane fuel in it. Everything on the truck was stock otherwise and the only changes to the engine tune was more aggressive PE mode. Has anyone else had something similar happen? Could the MAF have caused this?

The reason virtual sucks is because all it does is monitors the fuel trims. If it see's them adding fuel, upwards of say 15% then it will think it must be because it's running E85. So if you have something wrong with your tune or mechanical like a vacuum leak or headers not setup right, anything that causes fuel trims to start adding fuel, then you will see the Ethanol % start increasing.

In most cases the problems come in at idle when vacuum is highest, if you sit there idling then you may see fuel trims increasing meaning it's adding fuel, the more it increases, the more it starts increment the Flex Fuel %. Then, you stomp the gas, and you have all this extra timing that you don't need, then ping badda pingerooze.

Vacuum leaks can be something like the purge canister, etc. Header gaskets, manifold, etc. Anything that would cause it to run lean and add fuel to compensate.

Other problem is, you cant tell it's running that lean since you might only be adding 15% but that is with the say 70% ethanol reading so fuel is allready increased 20-25%, so really it's 35% adding fuel, not 15%.

Your better off just disabling virtual flex fuel, even on stock setups. Add a real sensor and do it right.

[jpierro79]

agreed real sensor no bs virtual. you can change your commanded to 14.1-1 as a minimal stoich as well and disable it. its what its set at anyway 10 percent is about 14.1-1 i run 14.4-1 incase its low but i had to rebuild my VVE to do it but ive got huge injectors ect ect

[HackAbuse]

Your better off just disabling virtual flex fuel, even on stock setups. Add a real sensor and do it right.

That's exactly what this was. A 100% factory truck down to the air filter. Virtual Flex was set up from the factory. I couldn't even figure out what was wrong with it at first until I saw stoich at 10.86

What's bizarre is that the PCM seemed to be ignoring fuel trims entirely as it applied to the estimated alcohol %. It almost seemed like it was pulling a number out of it's ass. I saw trims as high as -22% at certain loads

[10_SS]

That's exactly what this was. A 100% factory truck down to the air filter. Virtual Flex was set up from the factory. I couldn't even figure out what was wrong with it at first until I saw stoich at 10.86

What's bizarre is that the PCM seemed to be ignoring fuel trims entirely as it applied to the estimated alcohol %. It almost seemed like it was pulling a number out of it's ass. I saw trims as high as -22% at certain loads

I think it's trying to average a bit so certain loads if not left there long enough will be ignored, so at some load it's close to 0% LTFT and others is where it's off, making the virtual calc % increment when it shouldn't be. There may be tables that let you adjust virtual. It's a hot mess. You will find more and more newer years run a real sensor. Funny thing is that virtual crap is probably patented by somebody and they look like a hero cause they saved $3 per car by not having sensor, LOL. Just like the Baro calc. Some/most GM's dont have a real Baro. What a nightmare that has caused in years past.

[HackAbuse]

I think it's trying to average a bit so certain loads if not left there long enough will be ignored, so at some load it's close to 0% LTFT and others is where it's off, making the virtual calc % increment when it shouldn't be. There may be tables that let you adjust virtual. It's a hot mess. You will find more and more newer years run a real sensor. Funny thing is that virtual crap is probably patented by somebody and they look like a hero cause they saved $3 per car by not having sensor, LOL. Just like the Baro calc. Some/most GM's dont have a real Baro. What a nightmare that has caused in years past.

It seemed like a neat idea to me initially years ago, but after dealing with this, knowing that a simple logic error could potentially ruin an otherwise perfectly working engine is absolutely infuriating. I'm really hoping for the best here.

One thing is for sure. I will never let another vehicle roll out my door with virtual flex enabled, factory or otherwise, ever again.

[HawkZ28]

Mine went to crap with a fuel trim tune before I had my wideband, and just before I installed an MBRP catback, and ceramic coated LT's a few weeks later. All trims were -.1 to .4. One day it just spiked straight to 30% on 91 clear. With e85 would only read 55%, then one day pegged at 80% when I re-enabled for grins and giggles. Was able to get it down to 3% on clear, so I turned it off again. Tried new GM O2's, different year O2 settings, nothing made a difference. One option I've read is replacing the ECM to make it work "like it should", but then you're on borrowed time.

I noticed it would read closer in the learning stage when I was accelerating more than cruising. Since most driving is done cruising, it would settle at some arbitrary high E% on clear most times.

I'm fixing mine by ordering a kit from DSteck/DSX once I'm done paying for our dentists international vacation or boat provided at the expense of my girls and wife.

[Area47]

the problem 1. He puts crap gas in it and it triggers the ecm to read way more ethanol than what is in it due to the relearn process as outlined in the owners manual. trigger the relearn by putting fuel in it. drive it till the trims stop changing. shut off for 30 mins, do it again. You HAVE to drive them after you put fuel in them. learned this the hard way years ago.

[HawkZ28]

The reset cycle is triggered by 3 gallons or more fill up, and you are supposed to drive approximately 15 miles at 45mph or more immediately following the fill up. I have seen the learn procedure continue after a KOEO cycle following a short drive/fill up. I've also seen it continue below 45mph.

I don't think it's from crap gas. Happened on mine with BP, Shell and Phillips66 91 clear. Those are the only quality stations in my area. Not trusting Casey's, QT, kum n go, or the sketchy smokes, liquor, cell phones and gas stations.