A New Analysis of Premium vs. Regular

gasoline

#1

Introduction:

A while ago I read CarTalk’s analysis on premium vs. regular fuel; specifically, questioning whether or not one can simply rely on the knock sensor to run a vehicle that requires 91 on 87. According to CarTalk, you can run a premium required engine on 87, so long as it has a knock sensor and you don’t hear pinging:

But the general consensus among CarTalk, my dealer, and the internet at large is mixed. So I decided to dig deeper on the matter myself, and came across some sources of information:

http://papers.sae.org/960497/

http://www.repairfaq.org/filipg/AUTO/F_Gasoline7.html

The octane requirements of an engine are determined not only by its design and timing parameters, but also on the operating conditions and environments of the vehicle. For example, ambient temperature and altitude influence octane requirements.

For temperature, it seems that, roughly speaking, every 7C +/- results in a change of +/- 1 ON. For altitude, it seems to be about a reduction of 3-4 ON per increase of 1000m (or 10 kPa). Finally, spark advance/retard of 1 degree results in a change of 1 ON.

My hypothesis right now based on articles that I’ve read is that vehicle manufacturers also use the words “recommended” vs. “required” according to both engine design and assumed operating environments, and whether or not spark retarding can prevent knock under all possible operating conditions. For example, manufacturer might assume the vehicle to be operated between -20 F to 120 F. Altitudes -300 ft to 12,000 ft. Weight loads +80lbs to +1,000 lbs. (For typical consumer).

If a high compression engine design with a knock sensor permits the vehicle to operate in the full range of conditions by adjusting timing to avoid “significant” pre-detonation, then the manufacturer puts the label “91 recommended.”

If a high compression engine design with a knock sensor permits the vehicle to operate without knock in a subset of the full range of conditions by adjusting timing to avoid “significant” pre-detonation, then the manufacturer puts the label “91 required.”

As an observation, if you examine the new 2015 TLX, there are two engine types:

Compression 11.6:1: 4-cyl.
Compression 11.5:1: V-6 & SH-AWD

The 4 cylinder engine has a fuel label “91 recommended.” The V-6 has a fuel label “91 required,” even though its compression ratio is technically smaller. They both use the same mechanisms to compensate for varying octane number and operating environments: knock sensor with spark retarding. However, the V-6 model is also 300 lbs heavier than the 4 cylinder model. I also assume its operating temperatures may also be higher than the 4 cylinder, since it seems likely that 6 cylinders generate more heat than 4.

So obviously, compression ratio is not the only mechanism at play. If added weight and heat means the vehicle can operate without knock only between -20 F to 80 F and easy driving (reduced heat generation), instead of the full range of -20 F to 120 F and intense driving, the manufacturer would write “91 required” as opposed to “recommended”. I’m sure there are other factors at play here as well, but along the same lines of restricting the environments in which the engine can safely operate.

So I believe that if your engine’s operating conditions (easy driving, high altitude, or cold temperatures) are within the range that either knock is altogether avoided, or the engine can compensate through timing and knock sensor usage, no long term engine damage occurs, and your vehicle is fine running on regular 87 octane.

However, in an effort to understand how my own car (2005 Acura TL) would react to lower octane fuel, I set out to perform a few quantitative experiments, detailed in the next section.

Experiments:

Ideally, to understand under what circumstances the knock sensor is retarding ignition timing, you should read the “knock retard” parameter directly from your ECU using an OBD2 compliant scanning device. However, my car’s ECU does not report this value, so I had to measure it statistically.

To do so, I drove the same car twice: one trip on 100% 87 octane fuel, and another trip on 100% 93 octane fuel, in a 2005 Acura TL. The gas cap states this car “requires 91 octane” fuel. I recorded as many operating parameters as my ECU reports, and plotted ignition advance as a function of engine load for both trips:

Figure: http://i.imgur.com/qQeo4v4.jpg

In order to record engine parameters, I used the following Bluetooth OBD2 device:

And the “TorquePro” Android OS application.

Results:

Comparing ignition timing between 87 and 93 for the 2005 Acura TL:

  1. Does knock occur when using 87 octane in cold weather?

Yes. The knock sensor retards timing in cold weather, using both 87 and 93. The engine load required to trigger knock retard for 93 is about 70%, whereas for 87 it is sooner: about 50%.

  1. Does 87 retard timing more than 93?

Yes. By about 8-10 degrees on average after a load of 50%. Below 50% load is about the same.

  1. Does relying on the knock sensor cause engine damage?

Interestingly enough, 93 relies on the knock sensor under higher loads (>70%), as shown in the chart. So “relying on the knock sensor” to adjust timing does not, in itself, lead to engine damage, if the engineers are to be believed.

  1. Will using 87 cause long term damage my engine?

As shown in the chart, knock retard using 87 only leads to retards outside the realm of 93 when the engine load is above 70%. If the engine load stays below 70%, I think we can say conclusively that you’re not going to cause any damage, because the timing retard is still “in spec” for this vehicle.

Outside of spec, it’s hard to tell. During normal highway driving up a hill, I was hitting engine loads at high gear of around 70-90%. I drive with steptronic. The amount of time I spent above 70% load was roughly 17.6% of the time, or ~1/6. So that’s the amount of time I spent out of timing spec. But I could just as easily delay my shifting up in gears to prevent my engine from hitting those higher load values.

In conclusion, for my vehicle I tend to side with the CarTalk guys on this. So long as the heat & the load combined are not so great that the knock sensor cannot do its job, you are likely fine. You can also adjust your driving as well to prevent your vehicle from loading the engine too much. Clearly, the knock sensor is being used even when using 93 to keep the engine in proper operating order. As long as the engine is prevented from knocking, I do not believe engine damage will occur. Since environments, conditions, and operating states change based on geography, Acura cannot guarantee that in all ranges of expected situations the knock sensor will be able to do its job with 87 – but they do guarantee that with 91. So long as the knock sensor works with 87 in your environments, you’re probably fine.

Appendix: A Discussion of Fuel Detergents

Regarding fuel detergents, some top-tier gas stations put the same amount of engine cleaners in their 87 octane as in their 93, as well. Mobil is one such station:

http://www.abc15.com/news/let-joe-know/gas-test

Mobil even advertises that the only difference between their 87 octane fuel and 93 octane fuel is the octane level:

http://www.exxon.com/our-fuels


#2

That’s a pretty long thread, so far

I’ll get right to the point . . .

Are you advising people to use 87, even when the owner’s manual specifically says 91 is required?

Regardless, I don’t think you’re going to change anybody’s way of thinking, or break anybody’s habits

No offense intended

And what you had to say was interesting

But you are a sample size of 1


#3

While you might get away with cheaping out on your gas, I’d expect less performance and possibly driveability issues like surging and hesitation. And possibly over the long term, some engine damage–knock sensors can only react when they hear knocks, and knocks can cause damage. Why would you bother? Even when gas was around $5/gal in the US, the difference between grades remained the same for the most part, so you weren’t paying any more overall for the premium gas than you were before.

I just don’t understand why you would buy a car that specifies premium and then not meet the specs. Why not just get a different car? All the appliances and electronics in your house are probably spec’d for at least 115VAC, but if you could get electricity cheaper by only running 100 volts, would you do that too? I’d be frightened to use your toilet paper :smile:


#4
And possibly over the long term, some engine damage--knock sensors can only react when they hear knocks, and knocks can cause damage.

Sorry for the long post, I know its hard to read. But if you do, you will find that even when you use 93 octane in the 2005 Acura TL, the knock detector still “detects knock” and retards timing at high engine loads. The degree of “pre-ignition” it detects, however, is within tolerable limits designed to not lead to engine damage. The same process occurs with 87, just shifted to lower loads (~15-20%).

Using 87 octane, so long as you drive the car with the engine below 75% load, you see that the timing retard range is not different from running 93 and going to 90% engine load. In these circumstances, 87 cannot lead to engine damage.

My point is twofold:

  1. Even when running 93, your knock detector retards timing due to detecting pre-ignition at high engine loads. Most people don’t realize this. Your engine is still “knocking” using 93 – you never know it because your knock sensor is working.

  2. For the 2005 Acura TL, using 87 under reasonable circumstances long term will not lead to engine damage, because the degree of pre-ignition is similar to that seen using 93 octane.


#5

I thought that was an outstanding post. I would far rather see analysis than opinion. And I was pleased to see links to reputable impartial organizations that actually have the knowledge necessary to formulate an opinion.

My own final conclusion remains unchanged; while the knock sensor might under specific sets of conditions protect an engine running fuel of lower than recommended octane, doing so entails risk… Environmental conditions in most areas, especially here in New England, can change rapidly and unexpectedly. The goal in making a fuel selection should in my opinion be to protect the engine regardless of the circumstances that the driver faces. The risk/benefit simply isn’t IMHO conducive to taking the risk.

Interestingly, we recently had an excellent illustration of the unintended consequences of changes in ambient temperature, pressure, and environmental conditions. Namely, “deflategate”. While the league politicians and the media were stumbling all over themselves trying to cover their rears and to create glorious news stories, an independent engineer ran his own tests. He got some league balls, filled them to the league standard in a dry room at the standard temperature (I believe the league uses 72F with no regard for ambient pressure or humidity, but I’m not certain). He then put them in a chamber at a controlled 50F temperature and let them stabilize. As the math predicted, the pressure dropped, but not enough to throw the balls out of spec. However, when he wet the balls (remember that it was raining the day of the NE/Indiana game) the leather, being hygroscopic, absorbed the water and expanded, further dropping the pressure for a total of 2psi dropped, matching the game balls. The NFL and the media both ignored his findings. They had agendas inconsistent with the truth.

The bottom line is that those using other than recommended fuels are taking a risk. It may run without incident, albeit perhaps with lowered performance. But it may not. And if it doesn’t, the cost just might be painful. But good people often do foolish things. As long as they can get the pump nozzle in the fill hole, there’ll always be some that will try to save a buck. Truth is, they probably ignore the wise advice of the engineers who designed other things they own too. And some people will do this for their entire lives without incident. Other people, like me, would blow a hole in a piston the first time they tried it. 'Cause that’s how my life works. {:slight_smile:


#6

During your testing, could you ever detect any knock or detonation with your ears?? Or did you rely 100% on the knock sensors response to determine when pre-ignition occurred?

As an interesting side-bar, many old radial aircraft engines used water injection when asked to deliver take-off power to cool the incoming fuel mix and suppress detonation…Maximum power was noticeably reduced if this modest water spray was lost…


#7

Did you happen to put the cars on a chassis dyno and measure the amount of power that was lost when running 87 vs. 91? How about changes in fuel economy?


#8

There are other factors involved in something like this. Camshaft profile, any potential carbon buildup issues, spark plug gap, engine operating temperatures, etc, etc, and probably the most important of all; the EGR system.


#9

Caddyman: Never heard a knock or ping. Knock sensor catches them while they’re still inaudible. That’s pretty cool! I would like to learn how to fly at some point.

FoDaddy: No I did not. Mileage is about the same – I used the Automatic app to measure economy on my routes to work.

ok4450: Yes, but so long as the retard doesn’t get more severe than operation with 93, would those really play a role? As I said, only 16-17% of the time does the retard extend more severely than in use with 93. If you stay below a load of ~70%, you never do.


#10

@rustylogic this thread, and your original post, is largely academic. The reason is this: I can find no reason not to use premium fuel when recommended. Show me how it benefits the driver in any meaningful way. The actual cost of the fuel is irrelevant, and by that I mean that for a monthly budget where saving $4 a week is a significant amount, the ownership of a premium car–or any car at all–is a mistake, probably in addition to many other things like cable TV and prepared food.


#11

There could be trouble if the knock sensor fails.


#12

@rustylogic, pinging can occur even with knock sensors and the sensors working as they should.

I’m of the opinion that the EGR system has more of an effect on pinging than the knock sensors do.


#13
My point is twofold:
  1. Even when running 93, your knock detector retards timing due to detecting pre-ignition at high engine loads. Most people don’t realize this. Your engine is still “knocking” using 93 – you never know it because your knock sensor is working.

  2. For the 2005 Acura TL, using 87 under reasonable circumstances long term will not lead to engine damage, because the degree of pre-ignition is similar to that seen using 93 octane.

The flaw in this logic: duty cycle.

How much CUMULATIVE time do you suppose an Acura spends at “high load?” It’s a luxury car–and, like most modern, non-economy cars, it’s got excess power. Add in that it’s unlikely to tow much, and you can only spend a few seconds at a time in this portion of the performance enveolpe and remain at legal speeds!

Now, on lower octane gas, you’re bumping up against the knock sensor MUCH more often! As a result, any possible long-term deletrious effects might be “relevant” at a high duty cycle, and not at a lower duty cycle. (To illustrate, the liver effects of my drinking a case of beer a decade aren’t “relevant,” as I’ll be dead of old age before cirrhosis kicks in. At a case a DAY, it becomes quite “relevant”!)

So, if we cannot rule out engine damage…what CAN we say with certainty? Well, running “less timing” than the ECU would prefer means less power and worse economy. Running overly-retarded timing trashes both of these performance criteria, which is why the ECU only pulls timing as a last-gasp attempt to avoid engine damage! These effects MAY well eat up much, or all, of the money saved from the slightly cheaper fuel.

So, to sum it up–running too low of octane WILL increase fuel burn, WILL reduce performance, MAY or MAY NOT save any money…and MAYBE, JUST MAYBE can lead to long term engine damage! (And, remember that more fuel burnt means a larger “carbon footprint,” if that matters to your thinking.) Sure sounds like a losing proposition to me!


#14
So, to sum it up--running too low of octane WILL increase fuel burn, WILL reduce performance, MAY or MAY NOT save any money...and MAYBE, JUST MAYBE can lead to long term engine damage!

@meanjoe75fan makes a good point. Pinging can lead to engine damage. And I’ve seen engines where the ping detector failed and destroyed an engine. If you don’t think there’s a problem…fine. It’s your vehicle and your money.


#15

I will just say this…our local police department tried to “cheap out” by using regular gas in their high performance cruisers several years ago. The end result was holes burned in the tops of pistons even though all the vehicles had knock sensors. When the owner’s manuals “require” 91 octane then use it. Why second guess the people who actually designed and built the engines?


#16

@rustylogic - thanks for the great post, with lots of actual data. In my opinion, it confirms the standard advice to follow the manufacturer’s recommendations. You noted that cars where premium is ‘required’ have engines that cannot fully adjust to regular under all conditions of load, temperature, etc. To hope that damage would be avoided by careful driving is a risk we can’t take when advising strangers, and it’s a risk I wouldn’t take myself. Why buy a performance car and then be forced to drive it in a way (keeping the engine from being under too much load) that eliminates that performance advantage?

Another great reason to use premium is this:

"2) Does 87 retard timing more than 93?

Yes. By about 8-10 degrees on average after a load of 50%. Below 50% load is about the same."

This is HUGE.

The fact that the knock sensor is used with either gas doesn’t matter. The engine control is optimizing performance, I’d expect it to be constantly adjusting timing based on input from the knock sensor, which is what it’s doing. It’s getting more power and efficiency with premium by avoiding that 8-10 degree hit when using regular.


#17

I like the original post from @rustylogic. I like a post with data. It is a single example so I wouldn’t expand it to cover all cars but the testing and logic are good. Thanks for posting this. I may have to try a version of this on my Mustang with a aftermarket 93 octane calibration and my 11:1 compression ratio Honda S2000.


#18

“I like the original post from @rustylogic. I like a post with data.”

+1 here. I’ve been flamed a bit on this forum too for conveying information that doesn’t stay within the strict conventionality that a few posters here insist on.

Another aspect of rustylogic’s post is that under certain conditions some engines rated for 87 octane can benefit from higher octane fuel.
They presumably retard the timing sometimes under high load.
When the ambient temp is above 90F my 87 rated engine runs a little “peppier” with 89 octane fuel.
However, I saw no noticeable MPG increase the one time I filled with 89.


#19

Let me ask about this octane issue in reverse, if I may, please. All l.pthree previous cars I’ve owned over the past forty years had manufacturer recommended 87 octane. But mechanics always insisted that a full tank of premium octane be run through several times a year as beneficial to the engine. When I asked about that with the 2014 Camry I bought last year both the dealership mechanics and independent mechanic shop firmly said that is not needed on current new cars. So what has changed???


#20

Nothing has changed. There never was a reason to run premium through your cars ‘several times a year’.