Do you NEED Premium gas in a Cooper 'S' at Altitude?

In my TT I was told, and occasionally ran regular in
my Audi TT at Denvers altitude with less oxygen,
with little if any performace change . . . .

 

Yes, premium is highly recommended, regardless of location. Why chance it?

 

... and you realize our premium is hardly. While you can buy 93 east of here, 91 is the so called premium we can buy.

 

i ran premium in my 03 mcs and i still use it in my 08 shell gas only! they carry 93 and i no they dont water down their gas

 

If you're a Broncos fan put diesel in it.

 

The manual says to not use lower than 87 octane .. And .. if you use the 87 expect lowered performance. Here's a link to an article on the subject in today's on-line New York Times.

http://www.nytimes.com/2008/08/03/au.../03OCTANE.html

In the Times article, BMW stated that on their cars premium will result in highest performance level but 87 octane won't hurt anything.

That said, for every 2,500 feet or so in elevation, octane can be dropped by one and still maintain sea level spec. So, in the mile high city, 85 octane is the standard regular gas (equivalent to sea level's 87 octane).

But ... The octane rating change comes because of air the pressure difference between sea level and higher elevations in the intake manifold. The MCS is a turbo. It's intake manifold pressure is defined by the turbo, not by outside ambient air pressure. So .. Even at my elevation of 5,700 feet I don't plan to take the chance of using octane less than the recommended minimum of 87 (our mid-grade).

For non-turbo MCs there shouldn't be any reason at all why 85 octane could not be used at elevation.

The Times article, among other things, mentions that Hyundai says that their 375 HP high performance engine (requires premium but regular is OK) loses 7 HP with a switch to 87 octane, or a loss of less than 2% in performance.

I'm now using premium in my MCS but not so long ago I tried mid-grade and didn't detect any real difference in the way the car drove. Maybe I'll try it again now that my knowledge level has been expanded thanks to the NY Times.

 

One thing to remember is that the ratio of oxygen to other gases in the atmosphere doesn't change with altitude. The only thing that changes with altitude is pressure, and while that results in fewer oxygen atoms in a given volumn, the overall mix of gasses is the same.

While the lack of pressure lessens the combustion chamber pressure in a normally aspirated car, this actually increases a turbocharger's output, because there is a greater difference between exhaust pressure and ambient pressure. This raises the turbo's speed and increases the boost it puts out. Basically, while the outside pressure might be half of what it is at sea level, the cylinders see a pressure that's much higher (sea level or greater) due to the turbocharger's compressor, which gives the engine the same performance as if the car was at a lower altitude.

QED: use premium in your Mini S regardless of altitude, your car will thank you for it.

 

is to log your timing. See if it's getting any pull. If it's getting any ECU induced timing retardation, higher octane is beneficial.

All these articles are nice, but cars are getting tuned closer and closer to the edge. While all the philosophy may be educational, take them all as guides to the eye, for data logging is the only way to know for sure.

Matt

 

Is there a way to log the ignition timing that will let you positively tell the difference between fuel-related timing retardation and the normal timing adjustments that the ECU constantly makes to adjust for engine rpm and load?

An idea I've been playing around with is hooking a Keithley analog-to-digital converter directly to the engine-knock sensor and driving around for a while while I log the knock sensor output to a laptop. I figure if the knock sensor isn't triggering, then any ignition timing changes from the ECU won't be because of the fuel.

 

My ScanGauge II will monitor ignition timing if I ask it to. The problem for me is, it doesn't stay steady; it's dynamic; it's always changing. Need something that will store the data in real time and then allow you to graph it on your computer. I think I saw something like that on the market for around $100. But all it does is store data - no in-car display.

 

Even if you can graph the ScanGauge's recordings, you'll still see a lot of timing fluctuation that has nothing to do with the ECU pulling timing in response to engine knock - that's why I'd rather log the output from the knock sensor directly.

As an example, if you're cruising along at steady highway speeds, the ignition timing is going to be *very* advanced, possibly as much as 70 ° BTDC, because the lean air/fuel mixture in the cylinders doesn't ignite very quickly. But if you step on the gas to accelerate, the air/fuel mixture richens up and ignites more quickly, so the timing needs to be retarded to compensate. So in this case, the ECU is "pulling timing", but not because it's detected knocking from low-octane fuel.

 

Well said Matt. An R53 with just a 15% pulley will benefit from 100 octane unleaded 100% of the time. An R56 MCS also runs very close to the edge with its 16PSIg and 10.5:1 static CR. Anyone that even thinks about tipping into their throttle more than what a really slow grandma would do is foolhardy to use anything less than 91 octane.

$0.02,
Ryan

 

The only time you need the octain or to know the timing is at WOT, other times the load is low and the volumetric efficiency is lowered by the throttle body not being wide open.

Here's a graphic of timing pull, with some knock sensor output as well. Trouble is that knock sensor data isn't easily available (not an OBD datum) but timing is...



The 100 octane runs show nice smooth advance as the RPMs climb. The 91 octane shows a bunch of pulls, with the ECU then creeping up on more advance until it pulls again!

One thing that you can see here, is that the timing pull is very aggressive, and that making the amount that's pulled a bit smaller (50%-60%?) would result in even more average advance without really adding any risk to the engine.

The ECU is very smart, and does timing pull and advance on a cylinder by cylinder basis. So these aren't tons of runs, just two, with the bluish curves from 100 octane, and the reddish ones with 91. And OBD loggers can't get to the individual cylinder timing either! I think the AutoEnginuity can, but I don't think it supports the R56 yet. Other than that, it's expensive tools for sure.

Matt

 

Matt,
Thanks for the post - that's what I figured (that "timing pull" from low-octane gas is really only an issue at wide-open throttle).

On the chart, what are the units for the vertical axes? I'm supposing the left vertical axis is degrees of timing advance, but I can't tell what the right-hand axis is (the one that goes from -75000 to +75000). Also, what are the blue/red lines (R2 KW & R9 KW) at the top of the graph? Is that the knock sensor output? If so, do I interpret it correctly that starting at about 4.2 seconds, the knock sensor was triggering almost constantly with the 91 octane, but not at all with the 100 octane?

Also, in the lower half of the graph, is the blue line with all the data points on it an overlay of the timing information from all four cylinders, showing that none of them were pulling timing with 100 octane, while the pink/yellow/red mess underneath are the four cylinders showing various amounts of pull with 91 octane?

 

For those without data logging, try running a tank or two of lower octane gas, reset the mpg log and record what you get: also run the math for miles driven on the tank divided by gallons filled. Then switch back to premium, reset, and record. When I tried this, both the computer and my math indicated 89 octane reduced my mileage by 4 mpg, which completly wiped out the "savings" of paying less per gallon.

 

I did that for eight tankfuls last summer when I drove from Virginia to California, alternating at each stop between the lowest octane I could get and the highest. I averaged 360 miles per tank during the trip, so at each fuel stop, there wasn't much "old" gas left in the tank when I refilled.

The fuel economy averages were all between 28.3 MPG and 31.3 MPG, with no real correlation between octane and fuel economy (I averaged 28.3 on two different tankfuls, once with 87 octane and once with 91). The tank where I averaged 31.3 was with 87 octane also.

Of course, there could be some correlation hidden down in the data if I'd also recorded altitude, temperature, prevailing ground-level winds, etcetera during the trip, but I suspect that steady-speed highway droning just doesn't put a heavy enough load on the engine to require high-octane gas. Obviously, if you do a lot of city driving and/or you're hard on the gas pedal, your experience may differ. Honestly, with gas at $4.50/gallon here, the extra $0.20/gallon for premium just isn't that big of a deal.

If you're curious, here are the numbers. First column is octane, second column is average speed, and the third column is fuel economy:

93 59.5 29.7
87 66.1 30.9
93 71.3 29.7
87 65.2 28.3
91 75.3 28.3
86 77.6 29.4
91 64.3 29.0
87 64.9 31.3

 

Left timing in DBTDC, right is in something I have no clue about! I'm guessing it's just what BIM-COM did with the data field from the knock window value interpreted as an integer, but really I don't know. For the knock values, this is what the ECU reports for just one of the cylinders for what's known as "knock window". Our ECUs don't just listen to the knock sensor, it's the processed knock sensor value in a certian amount of cam angle. That's how it can control each cylinder independantly with a single knock sensor. From the data I'd looked at, it seems that it's a heavily processed number, with some threshholding and gain. Because the sensor is at a different realitive position for each cylinder, I'd assume that these processing algorithms can be per cylinder as well. But really, all I've seen is the number goes wacky if it's detecting knock, and stays flat at the max value if it's not.

So yes, your interpretation is correct. You can also see that because the timing for the WOT runs with 91 never get to the max timing that the 100 gets to.

Matt

 

...so for those of use without an engineering background, what is the conclusion you can draw from this knock data? Is 91 that important? Is important in high-RPM driving like city and not so on highway? Is there not enough data to draw the conclusion?

 

is needed if you drive hard. If you're a driving pusssy, or very conservative, it's probably overkill.

Also, in some places like CA, you can't get really good gas anywhere but the few stations that sell 100 octane, or race tracks. Personally, I find it funny that the largest market for high performance cars in the US can't really get the full benefit of all the high end engineering that goes into them. Sad for me and the rest of the residents of the state.

But to the original question of do you need it at high altitude, the answer still is, it depends! I don't have data from the new engine at high altitude... And engine condition (spark plugs and carbon build up) will effect the results as well.

But it is also true that with this ever wider use of feedback controlled timing advance, the manufacturers are tuning closer and closer to the "edge" and letting the electro-nannies keep us from trouble.

Sorry, like most really important things, there is no simple answer.

Matt

 

Matt, are you using "High Octane" to mean 100 octane or 91+ octane?

 

or 93 if it's available where you live. That's what the cars are tuned for. But during some other parts testing, I found that most of a mods potential was "taken away" by timing pull. So if all you can get is 91 and you keep modding your car, you can't get all there is to be had from the parts. Tuning helps somewhat, but there are some fundimental limits to what you can stuff into a cylinder before bad things happen, and then the electro-nanny steps in.

If you're really interested in this stuff, there are lots of books on the subject of tuning forced induction cars. While our cars are a bit "newer" than the tech in the two books I point to, I wrote a couple reviews of some books in the reviews section that are good educations on the basic ideas that govern engine operation.

FWIW, none of this stuff is new. When I was digging around on water injection, I found some nice graphs from 1840 or so about ideal A/F ratios vs compression ratio that has data that's as insightful today as it was then.

FWIW, I just used the 100 as a "control" to see what would happen if timing pull were not an issue. I actually don't know what octane my car would show no pull at, but I'm guessing about 95 or so...

Matt

 

Matt,
One last question to put your timing/knock graph into context - what kind of engine mods are on the car that was used for the test? If that's your personal car, I'm assuming there's at least a pulley on it.

I'd be curious to do a similar test with a bone-stock engine using 91 or 93-octane gas and see if the timing pull is still there.

 

with pulley and some other minor stuff, but pre head install.

I've seen timing pull with 91 on bone stock cars as well. I don't know about 93. And I have no data on the new engine or R50s. But like I said, plug and cylinder condition affect this too.

Matt

 

Those are really bad highway numbers for an R56 (you don't list what you're driving). My stock R56 gets at least 34mph highway every time I've measured it. I use 91 gas. Do you know why your car gets such low mileage?

 

I think that covers it....

Matt