I can't quite wrap my head around GDI engines and fuel dilution of the engine oil

I guess the engine makes a pour refinery at separating the oil and fuel, once they are mixed together in harmony and synergy.

I have a feeling that the amount of fuel might be excessive. I don't want to believe the automakers didn't notice this with their GDI tuning. They also didn't notice LSPI, IVD.... and the list goes on.

The question you should ask... HOW MANY GALLONS OF FUEL MAKES IT INTO THE OIL SO THAT THE UOA shows that 5-10 ounces not evap'd at any given point? and would you notice it?

My wife drives 4000 miles and uses 200 gallons of fuel. Would I notice if it was really 199 gallons or 201 gallons? I won't guess at the accumulation or evaporation rate. How many gallons of fuel do you think you're oil and PCV are processing continuously? Then, be scared. Gallons might be blown past your rings over your 10k or OLM interval.


That is mind boggling.


Poor fuel economy too.
 
That's mostly true, but there are still nice gains NA. Many NA engines have switched to DI and it's not because it's cheaper and easier. Your example is a bit odd using two completely different engines.
I'm using two pushrod V8's, that's hardly odd Jimmy.

If we want to look at other examples:
The Pentastar V6 is 3.6L, PI and 305HP; 84.72HP/L
The Honda J35Y4 is 3.5L, DI and 280HP; 80HP/L
The GM HFV6 is 3.6L, DI, and 301 to 323HP (depending on application, Camaro has the highest perf version): 83.6-89.7HP/L.

As I said, it's when you add forced induction that you get the real bump, because you can keep the high compression ratio, which you can't do with PI.
 
I'm using two pushrod V8's, that's hardly odd Jimmy.

If we want to look at other examples:
The Pentastar V6 is 3.6L, PI and 305HP; 84.72HP/L
The Honda J35Y4 is 3.5L, DI and 280HP; 80HP/L
The GM HFV6 is 3.6L, DI, and 301 to 323HP (depending on application, Camaro has the highest perf version): 83.6-89.7HP/L.

As I said, it's when you add forced induction that you get the real bump, because you can keep the high compression ratio, which you can't do with PI.
Since I have no desire to go with forced induction I'd gladly give up a few HP and have PI.
 
I'm using two pushrod V8's, that's hardly odd Jimmy.

If we want to look at other examples:
The Pentastar V6 is 3.6L, PI and 305HP; 84.72HP/L
The Honda J35Y4 is 3.5L, DI and 280HP; 80HP/L
The GM HFV6 is 3.6L, DI, and 301 to 323HP (depending on application, Camaro has the highest perf version): 83.6-89.7HP/L.

As I said, it's when you add forced induction that you get the real bump, because you can keep the high compression ratio, which you can't do with PI.

Compare the same engine. First that comes to mind because I own one:

A 2018 Forester 2.5 (PI) made 170 hp and 174 lbs-ft. MPG was 24/32
A 2019 with the same engine (DI) makes 182/176, MPG was 26/33

That's just one example and is not even an engine that was designed with DI in mind in the first place. Typically designed ground up they also can run much higher compression ratios.
 
Compare the same engine. First that comes to mind because I own one:

A 2018 Forester 2.5 (PI) made 170 hp and 174 lbs-ft. MPG was 24/32
A 2019 with the same engine (DI) makes 182/176, MPG was 26/33

That's just one example and is not even an engine that was designed with DI in mind in the first place. Typically designed ground up they also can run much higher compression ratios.
Problem with that line of though is that it ignores that you can get those kinds of gains just with architectural upgrades, with no requirement for DI. The original Pentastar was 260HP for example, later bumped to 305. Still no DI.

Your Subaru example went from 68.8HP/L to 72.8HP/L, either figure well within reach for a port injected mill as I've already illustrated. The fuel economy gain seems to be the most significant improvement attributable to DI in your example, certainly not the anaemic power figures.

I maintain that the your argument:
Jimmy_Russells said:
DI makes so much more power
Really only applies to forced induction applications.

It is more efficient however, I agree with you on that.
 
Problem with that line of though is that it ignores that you can get those kinds of gains just with architectural upgrades, with no requirement for DI. The original Pentastar was 260HP for example, later bumped to 305. Still no DI.

Your Subaru example went from 68.8HP/L to 72.8HP/L, either figure well within reach for a port injected mill as I've already illustrated. The fuel economy gain seems to be the most significant improvement attributable to DI in your example, certainly not the anaemic power figures.

I maintain that the your argument:

Really only applies to forced induction applications.

It is more efficient however, I agree with you on that.
To point out how anemic the FB25 is (and I’m a Subie fan obviously) the 182HP/2.5L equates to a ~365HP 5.0. Ford’s latest NA 5.0 makes 480HP… meaning Subaru’s implementation makes 33% less HP/L than what is fairly easily possible. To repurpose a Subaru funny, their world actually is pretty flat… power wise 🤣

It’s a good thing they haven’t given up on AWD.
 
Some of it will but not all of it. Plus the presence of the fuel can cause permanent degradation of the viscosity index improvers used in the oil. Gasoline in the oil is not as benign as people like to believe.

Which is why I opted for a 10W-30 without viscosity improvers. 3.5 HTHS, and fluid enough for any temp I'll ever see. Very low noack (3.3%) and very high flashpoint (260°C) which should help in evaporating just the fuel/water and not the lightest fractions of the oil. Also has esters to hopefully mitigate deposits on the intake valves by either dissolving carbon or by keeping valve stem seals pliable.
 
I guess the engine makes a pour refinery at separating the oil and fuel, once they are mixed together in harmony and synergy.

I have a feeling that the amount of fuel might be excessive. I don't want to believe the automakers didn't notice this with their GDI tuning. They also didn't notice LSPI, IVD.... and the list goes on.

The question you should ask... HOW MANY GALLONS OF FUEL MAKES IT INTO THE OIL SO THAT THE UOA shows that 5-10 ounces not evap'd at any given point? and would you notice it?

My wife drives 4000 miles and uses 200 gallons of fuel. Would I notice if it was really 199 gallons or 201 gallons? I won't guess at the accumulation or evaporation rate. How many gallons of fuel do you think you're oil and PCV are processing continuously? Then, be scared. Gallons might be blown past your rings over your 10k or OLM interval.

Well, if the volume between the pioston and cylinder wall makes up 2% of the total chamber volume at TDC, an the rising pressure after ignition pushes about 5% of the mixture in there, I would expect between 1 and 3% of the fuel consumption ends up in the crankcase on a healthy engine, possibly all 5% or more when compression is compromised with stuck rings.

The majority of the fuel will go through the pcv, but considering how smelly short trippers get a good percentage actually gets in the oil.
 
To point out how anemic the FB25 is (and I’m a Subie fan obviously) the 182HP/2.5L equates to a ~365HP 5.0. Ford’s latest NA 5.0 makes 480HP… meaning Subaru’s implementation makes 33% less HP/L than what is fairly easily possible. To repurpose a Subaru funny, their world actually is pretty flat… power wise 🤣

It’s a good thing they haven’t given up on AWD.

You guys are into weird comparisons. They aren't competing with Mustangs, they are competing with Rav4's and Corollas and the like in a world where 90% of those buyers care much more about fuel mileage than power (my wife included).

I mean Audi was making 100 hp/L out of an NA DI engine... 15+ years ago. That makes the Mustang look pretty sad.
 
Which is why I opted for a 10W-30 without viscosity improvers. 3.5 HTHS, and fluid enough for any temp I'll ever see. Very low noack (3.3%) and very high flashpoint (260°C) which should help in evaporating just the fuel/water and not the lightest fractions of the oil. Also has esters to hopefully mitigate deposits on the intake valves by either dissolving carbon or by keeping valve stem seals pliable.
Is it possible to make a multi-viscosity oil without using viscosity index improvers?
 
Is it possible to make a multi-viscosity oil without using viscosity index improvers?
Sure, Dave says they’ve got a full line without VII; I just can’t tell you how it works. For sure their 10w20 does not have any VII, not sure which other ones fall into that category. But 10w30 seems to be a fairly easy one to hit; you’ll notice most 10w30s have much lower viscosity indexes, and significantly lower Noack which are both good indicators of how much VII is present.
 
Which is why I opted for a 10W-30 without viscosity improvers. 3.5 HTHS, and fluid enough for any temp I'll ever see. Very low noack (3.3%) and very high flashpoint (260°C) which should help in evaporating just the fuel/water and not the lightest fractions of the oil. Also has esters to hopefully mitigate deposits on the intake valves by either dissolving carbon or by keeping valve stem seals pliable.
What oil?
 
You guys are into weird comparisons. They aren't competing with Mustangs, they are competing with Rav4's and Corollas and the like in a world where 90% of those buyers care much more about fuel mileage than power (my wife included).

I mean Audi was making 100 hp/L out of an NA DI engine... 15+ years ago. That makes the Mustang look pretty sad.
Honda made 124HP/L out of an N/A port injected engine, 24 years ago ;) (S2000)
BMW made 100HP/L out of an N/A port injected engine 19 years ago with the E60 M5 (S85 engine)
 
Not true. DI allows for greater efficiency, power, and cleaner emissions.
That's questionable.
The gain in thermodynamic efficiency may help the manufacturer in the CAFE game, but it hurts the user when the added initial cost, and added long-term maintenance costs outweigh the fuel savings.
"Cleaner emissions" as in greatly increased emissions of carcinogenic soot?
 
With gasoline as volatile as it is, and engine operating temperatures as high as they are, and PCV systems doing their thing, how does gasoline manage to dilute the engine oil? It seems like it would get vaporized as soon as the engine warmed up. How come it doesn't? Does it combine with the oil forming a different solution?
Aside from the higher fuel pressure and where it’s being injected (directly into the cylinder), you also have to consider lower tension piston rings.

So why doesn’t it all burn off in the crankcase? My theory is this, because it doesn’t really stop happening. It’s not a one time event, it’s constant, with huge variations depending on cold starts, commute distance, driving conditions, the driver, outside temps, etc.

I’ve Owned 2 direct injection vehicles, one (2018 Chevy Silverado) I could smell the fuel in my oil and my oil catch can. Not only could I smell it but several ASE fellow techs could smell it. We’re are all flabbergasted looking at each other, asking...is that fuel? I smell fuel, do you smell fuel? That’s fuel, right? 🤣🤣🤣 Full grown men sniffing my oil catch can and drained oil from the crankcase. But that was more in the colder months. Warmer months not as much. Can’t comment much further on that one, traded it in at 50,000 miles...random misfires (direct injectors), oil consumption. The other was a Lexus LS460. No issues with that one until around 100,000 miles, then oil consumption got enormous to the point I traded it in at 179,000 miles. Dual system, port and DI. Rings seemed wiped out...oil tracking into intake, even through the air inlet snorkel up into the throttle body.
 
I scanned a few pages of replies and I didn't notice anyone with this question / observation about D.I.:

I was under the belief that computer-controlled injection of fuel directly into the cylinder would assure a more accurate mixture (for that specific cylinder), resulting in a more complete combustion event for each cylinder. That would imply that LESS fuel should enter the crankcase, rather than more. Isn't this how D.I. is supposed to work? Why doesn't it? It seems that D.I. somehow uses MORE fuel than is needed by the combustion process, yet D.I. is touted as providing better gas mileage numbers? I realize that more fuel is often needed for cooling, but it shouldn't be more than other types of induction. . . .what am I missing?
 
I scanned a few pages of replies and I didn't notice anyone with this question / observation about D.I.:

I was under the belief that computer-controlled injection of fuel directly into the cylinder would assure a more accurate mixture (for that specific cylinder), resulting in a more complete combustion event for each cylinder. That would imply that LESS fuel should enter the crankcase, rather than more. Isn't this how D.I. is supposed to work? Why doesn't it? It seems that D.I. somehow uses MORE fuel than is needed by the combustion process, yet D.I. is touted as providing better gas mileage numbers? I realize that more fuel is often needed for cooling, but it shouldn't be more than other types of induction. . . .what am I missing?
Fuel is being used to control exhaust temps and emissions, along with being used as a bandaid for higher compression and turbo boost.
And, the fuel isn't allowed enough time to fully atomize. Its like any product. Overhyped too much but not even close to living up to the hype hence why you shouldn't have any 'beliefs' in marketing.
 
I scanned a few pages of replies and I didn't notice anyone with this question / observation about D.I.:

I was under the belief that computer-controlled injection of fuel directly into the cylinder would assure a more accurate mixture (for that specific cylinder), resulting in a more complete combustion event for each cylinder. That would imply that LESS fuel should enter the crankcase, rather than more. Isn't this how D.I. is supposed to work? Why doesn't it? It seems that D.I. somehow uses MORE fuel than is needed by the combustion process, yet D.I. is touted as providing better gas mileage numbers? I realize that more fuel is often needed for cooling, but it shouldn't be more than other types of induction. . . .what am I missing?

There's much less time to vaporise the fuel. With port injection you spray into the intake port and fuel sits there on a hot intake valve evaporating off. Then the valve opens and fuel + air rushes into the cylinder, chasing down a piston that's moving towards BDC and then back up. Intake valve closes an the mixture gets compressed adding more heat until finally the plug sparks and burn starts. The flame front moving from the plug compresses the mixture more, into the metal and between piston and cylinder wall. Where combustion is quenched

With DI, you only inject the liquid fuel when the piston is nearing TDC, just before the plug sparks. There's simply much less time to evaporate, and you are pushing liquid fuel into the areas that quench burning, and past the piston rings. Now, if we assume liquid fuel is 100x more dense than vaporised, your fuel dilution issues could be up to 100x worse.
 
Kind of an aside on the DI, it was being experimented with in the early 60s. I believe it was Mechanix Illustrated that had an article about 2 guys who adapted DI to a flathead Dodge. They conjured up a high swirl combustion chamber and essentially a MSD ignition. The found out that it would run on any combustible fuel and achieved better fuel economy. By the time the article was written Texaco had purchased all rights to the design. 1962 or 3
 
So your oil level was down over a quart and you're attributing that to fuel dilution?
Well it was good just before the trip. No leaks.
Consider the possibility of a pint burnoff of fuel and water and the rest?
Gummed up rings? The oil was all full of solubilised varnish and soot. See picture above.
The change to a 10W30 fixed the burn off and sooty appearance. I am sure Summer season helped also.

What is your speculation? - Ken
 
Back
Top