87 vs 91 Octane: LC vs LX570

[Redbull]

I use whatever the manual tells me. For my Lexus it says 91+ and my Land Cruiser and Explorer recommend 87 so that's what I use.

My personal thing is I change the oil every 5k miles on all of my cars regardless of what the manufacturer says - so I don't always listen

 

[jefo13]

Reviving this thread here... [totally anecdotal and not scientific) - 2016 LC

I often take a drive to the mountains to ski, 750 miles RT. I have always used 87 and avg about 14 MPG
This weekend because I had a 25 cent discount from Tmobile I filled at Shell with 91. I immediately noticed a less "tacky engine" and smoother acceleration. But what was hard to believe for be is that I averaged 15 MPG. I will continue to use high octane for the rest of this ski season to see what is real or not...

 

[bjorn240]

In reality, it's incrementally more cost to fuel with premium, but it's also incrementally better performance AND better fuel efficiency. It's not just throwing money away by splurging for that premium.

There is no reason to believe that premium fuel would give better fuel efficiency than 87. In fact, the higher-octane hydrocarbon blend to prevent knock is actually likely to have 1-2% less BTU/gallon than a standard blend. And unless we have evidence that the ECU tuning is actually different between the two cars, it's unlikely we'd get better performance either -- unless the timing is allowed to be more aggressive on the LX than the LC, there is no way for it to adjust for more knock-resistant fuel.

 

[TeCKis300]

There is no reason to believe that premium fuel would give better fuel efficiency than 87. In fact, the higher-octane hydrocarbon blend to prevent knock is actually likely to have 1-2% less BTU/gallon than a standard blend. And unless we have evidence that the ECU tuning is actually different between the two cars, it's unlikely we'd get better performance either -- unless the timing is allowed to be more aggressive on the LX than the LC, there is no way for it to adjust for more knock-resistant fuel.

From a BTU/gallon standpoint, you're absolutely right. Octane does not add energy content. In fact, octane is a combustion inhibitor.

How this plays out is that a higher octane fuel allows the air/fuel mixture to be compressed further, thereby allowing the extraction of more power. Octane is a key variable in allowing for increased power per unit of fuel, and thereby MPG. This is a concept that is highly accepted, and industry has been pushing for higher octane at the pumps to enable them to reach new targets - SAE Powertrain Panel: Higher-Octane Gas Could Improve Fuel Economy

Between the LC and LX, the engines are mechanically identical. Where the extra compression can come from is dynamic compression in the form of timing as you suggested.

Sure, there's no proof here. There is strong anecdotal evidence that suggests it though.

The LX is rated for more HP - 383hp vs 381hp, 403tq vs 401tq. The manual also requires high octane fuel only for the LX. Calibrations are easily and readily changed. I've tuned cars for many years and I can do it as an end user. I've seen the base maps be different in the tri-state region where we have 91 octane, vs rest of the nation on 93 octane. Not saying the LC/LX, but many models use to have CA specific calibrations and parts to meet their more stringent requirements. Tailoring calibrations is pervasive and common, and highly likely between models.

The Tundra also shares the same driveline. There's talk about tunes being developed for the LC, from Tundra tuner shops. There's a reason they can't just slap it on, as the calibrations there are surely different. It's not hard to imagine vehicles of shared drivelines to have different calibrations, and as a tuner having done this for many years, I've seen the evidence.

That said, I wouldn't bother running an LC on anything more than it's required 87 octane. Though more can very possibly eek out more performance under highly loaded situations (e.g. towing) For LX, it's required 91 octane.

 

[linuxgod]

I run 87 but I did notice my mileage jumped over the last month or two. Over the summer and in Sept I got ~11MPG cruising at 75-80 on the highway. Recently I did the first half of the same drive and got almost 14. I was actually wondering if something in the gas formulation has changed a bit recently (less ethanol?)

 

[swatski]

Somehow, I missed this discussion, and I actually looked! Very interesting, I will post my response as I have actually given this quite a bit of consideration, more or less around these two issues (in quotes):

Can you provide evidence of this "performance" advantage based on fuel in the LX? I ask because I've never seen anyone substantiate this. If someone can, I'd love to see it. Until then, I remain convinced it's all in the head (and wallet).

There is no reason to believe that premium fuel would give better fuel efficiency than 87. In fact, the higher-octane hydrocarbon blend to prevent knock is actually likely to have 1-2% less BTU/gallon than a standard blend. And unless we have evidence that the ECU tuning is actually different between the two cars, it's unlikely we'd get better performance either -- unless the timing is allowed to be more aggressive on the LX than the LC, there is no way for it to adjust for more knock-resistant fuel.

So, I started experimenting with low (87, regular) vs high (93, premium) octane a few months ago, after reading about new-er Toyota engines being controlled by sensors - almost entirely. The way I understand it, Toyota has decided to allow the software (algorithms) to take raw data, in real time, to calibrate and run the ECUs so there is no actual "tune", more like running within a "constant variable" depending on readings from sensors - pre-drivetrain, post drivetrain, and everything in the middle...

One of the things I read that made an impression on me was the DAP tuners claim (made with regards to Tundra 5.7 engine tuning) that "These engines pull 8-12 degrees of timing due to knock retard on 87 oct, 6-10 deg on 89oct, 2-6 deg on 91oct, and 0-2deg on 93oct. That is on a completely stock truck with factory tune." Wow! (if that's true).

I decided to look into this some more. Over a period of time I ran my 2017 LC200 switching half-tanks of either 87 or 93 gas (running on empty in between to minimize mixing the two grades). Consistently, my engine was smoother running 93 under various loads. Most notably, the engine did not down shift as much (compared to 87 gas) - most apparent at highway speeds, but similar when driving a local route - going over the same hills and keeping the same speed.
(I didn't notice any better/changed MPGs.)

The results of running 93 were most obvious on one particular stretch up the hill - a nice, gradual incline where on 87 the truck would not maintain speed (without a downshift) but with 93 it stayed in higher gear all the way up the hill. Done this enough to know it was not a placebo effect.


I conclude that with these trucks 93 octane provides for more power (through more timing?) under a given load.

Does it run perfectly fine and reliably on 87? Yes.
Does it run better on higher octane? Yes, and I can tell.

So, I run on 93 all the time now.

(Again - never seen any fuel economy gains with 93)


--

 

[linuxgod]

There's one sure way to settle this... Anyone have access to a dyno?

 

[bjorn240]

There is strong anecdotal evidence that suggests it though. The LX is rated for more HP - 383hp vs 381hp, 403tq vs 401tq. The manual also requires high octane fuel only for the LX.

The only verifiable evidence is the difference in the manual. The SAE protocol for power measurement (J1349) states, "The results are considered valid if the repeatability of at least three measurements is within 1% of the mean." If you combine that with the fact that two identical 3UR-FEs would vary by a percent, there's no user-discernible evidence they ran more aggressive timing during the (LX) 3UR-FE dyno test. I agree it's likely that they advanced the timing ever so slightly on the LX (why else would they make the recommendation for 91), but it's also possible that the difference is a result of dyno and engine differences.

In any case, the Toyota ECU is likely sufficiently advanced to have both a timing retardation and a long-term ignition learning function (heck aftermarket ECUs have had this for years). So if you run the Lexus on 87, it'll identify the instances where knock happens, retard the timing immediately, and then retard the timing in the same load/rpm scenario going forward -- effectively running the LC map, if such a difference exists. I run 91 in my LX, but given the above and the known conservatism of Toyota, I'd have no problem running 87 if I needed to. I judge the risk of damage to be infinitesimal.

This should be easy enough to find out though - there must documentation in TTIS about the advance on the varieties of 3UR-FE or be an OBDII plug that can tell us ignition advance. Let's get an OBDII and run it on an LC and an LX on a particular protocol. Should settle it once and for all.

 

[Markuson]

If you think there is $4-5k of performance improvement value every 200k miles (the cost of forking out for premium) makes sense....then by all means, keep buying premium gas.

As for me...I’ve got better things to do with that $$ than throwing it at a mythological improvement only a Dyno has any shot at discerning (or proving false entirely).

 

[bjorn240]

The issue really isn’t whether the LC will perform better on 91. It’s whether there are any scenarios where the LX knocks on 87. One knock is one too many. TeCkis300 and I are just debating that likelihood.

 

[Taco2Cruiser]

There's one sure way to settle this... Anyone have access to a dyno?

I did, and I’ve done testing with both. About 15 tanks with each, back and forth, from 87 to 91 to 87 to 91. Computer was reset between them. I also tested each tank for ethanol levels as they were a very consistent 5% ethanol.

No real difference in power output. Sometimes a coupe more HPs or torques one way or another. Leans to 91 giving more about 70% of the time.

What matters most is fresh fuel, and 87 has a better chance of being fresher at random stations.

To through a better wrench in the mix, I am now playing with (and trust me, I am. It a fan RedLine fan boy) RedLine SI-1 as a maintence dose. About 1/5 ounce (approved is 1/6 to 1/3) ouch per gallon.

Or about $0.16 more per gallon. That has been giving me about 1/3 mpg increase over the last 4,000 miles. That includes three separate 300 mile pulls with a 4,300 pound trailer and two 250 mile 9,000 pound pulls.

As for smoothness, my 5.7 is smooth no matter what it’s got running through it.

When I scope the valves from inside the block and the top of the cylinders and look at the intake valves , they barely look like they are showing a tad less gunk. Versus my usually things I do, which I’m not going to get into because it will distract.

As for pinging, no, my tubing software doesn’t show any difference between the two fuels.

So I finish this with what @Markuson said, if you want to spend the extra cash on premium, do it, it will make you happier maybe.

For me, I’m spending my money on more fuel, to go more places. And keep my retirement saving fat, so I can keep using regular fuel, to travel more places.

Both are good, but there is more to vehicle mainenance than a small percentage change of stuff that replicates octane in street fuel. (That’s why is 87 or 93, it’s either 87% or 93%, as good as 100% octane. (Come talk to me when you guys run leaded 112 race fuel in a 1700hp 454, then fuel starts to matter.)

Edit: please don’t take this as me being mean or condescending. If I had the Lexus, I’d probably do what the manual said. But it doesn’t negate what I’ve found when it comes to the Toyota.

 

[TeCKis300]

Happy to debate the science behind this as it's a much deeper topic than what's evident on the surface (more on this later). I'll agree with those looking for a discernable ignition timing, HP and/or efficiency differences, that it’s all ultimately “in the noise”. The variability in samples is sufficient to make up the difference. Yet I personally strongly believe there are differences between the LC and LX.

For those looking for a smoking gun readout. The problem is much harder than that. Calibrations are proprietary and not openly shared outside of Toyota, unless one has a tool to dump the maps, like what exists in many other platforms. That is where I am drawing my conclusion, because I have seen maps and differences that exists elsewhere, to know that Toyota would do the same.

Chassis dyno’s are really too crude of a tool to really measure this level of fidelity. MPG, too course. And then there are the litany of environmental factors that would make exact repeatable performances suspect. Even a dyno run on the same vehicle back to back shows more noise than this. ECU’s learn, engines/drivetrain/tires get warm, even on a dyno. Throw in a different vehicle, with different state of wear and other variables…it’s really hard to look for this level of output difference.

Timing is not an exact science even in an ECU. There’s literal ton of variables that weigh into the algorithm and compounded strategies. Much beyond just throttle position and fuel octane. Ambient environmental conditions, barometric pressure, engine temp, water temp, short term trims, long term trims, trims for each cylinder, fuel quality, fuel octane, vvti maps, O2 feedback, knock logs, catalytic condition, interaction between each bank in the V8, mechanical condition, hysteresis of all those variable… in other words – it’s not a simple back to back comparison that an end user will figure out.

Yet octane is a significant variable to the engine that the calibrators will want to account for. Higher octane fuel has a slower flame front than lower octane fuels. The optimal strategy always involves advancing timing for octane. Not necessarily just in regards to knock. But from an optimal combustion parameter standpoint, because the pressure rise within a cylinder based on octane is important to consider. Someone earlier said engines run on algorithms. Absolutely true. Yet these algorithms are always seeded with base maps that are pretty darn close to optimal. Trims are largely used to compensate for negative conditions, based on closed loop feedback from oxygen, knock sensors, etc. Sure, it’s possible these can advance for more fuel quality. That generally requires a more aggressive probing of the knock boundary. Meaning knock occurs when it tries advance, then it backs off if there’s not enough octane. This is generally not aligned to conservative OEMs like Toyota, where the general impact is more knock events that don't contribute to long term durability. Pair this probing strategy with a bad batch of gas, and the knock may not even be subtle. Enough for the user to feel, and perhaps even damage the engine. That’s why it’s important for base maps to be optimal for the expected octane.

Sorry if this is long winded. I hope this is not taken as arguing either as it's sharing of knowledge I have in this area.

 

[razzyrazz]

Learning a lot here. Without derailing this thread, anyone have thoughts on 89 octane? Often significantly cheaper than 91 or 93 and not much more than 87.

 

[Markuson]

In honor of this 2-year-old thread...I offer this...Cliff’s Notes Version of this thread:











!!!
!!!
!!!!!!!






Heehee

 

[Taco2Cruiser]

I think @TeCKis300 and I are at the same juncture that we always arrive at.

He looks more at the science of tuning.
I look more at what I see under real world test conditions.

Either are right, neither are wrong. They actually require both to be the best.

That said, I’ll still beat @TeCKis300 around a track. Muhahahhah!!!! Just playin’ (or am I?)

Alright... I'm going back over to bob is the oil guy. Read about molecules some more.

 

[bjorn240]

Timing is not an exact science even in an ECU. There’s literal ton of variables that weigh into the algorithm and compounded strategies. Much beyond just throttle position and fuel octane. Ambient environmental conditions, barometric pressure, engine temp, water temp, short term trims, long term trims, trims for each cylinder, fuel quality, fuel octane, vvti maps, O2 feedback, knock logs, catalytic condition, interaction between each bank in the V8, mechanical condition, hysteresis of all those variable… in other words – it’s not a simple back to back comparison that an end user will figure out.

All that is true for designing an ECU program (I’ve programmed everything from piggyback controllers for OEM ECUs for motorbikes to MoTech M880s for turbo awd cars - I made my living in motorsports for 5 years before I determined banking was safer and more lucrative) but I think you’re massively complicating this - you don’t need the inputs to tell you anything about the resulting ignition advance, you just need the max degrees before TDC - a single output - for a given protocol from an LC running 87 and an LX running 91. If, over the course of a similar 50-mile protocol, the LX shows a higher degrees before TDC, it’s running more aggressive spark advance. At that point, you’ve proven your hypothesis that the LX is tuned for higher octane gas. Then it would also be interesting to see what the LC does on 91. You could obviously map them against RPM/load/GPS location etc and get very fancy, but max degrees is the only measurement you need.

Like I said, I think that’s likely what you’d find, but I wouldn’t be surprised to learn they were the same either.

 

[NYC570]

You don't necessarily need a dyno to test this. You'd only need to check your KCLV with both 87 and 93 to see how well the engine is running with each. Toyota is very conservative with their tunes which are known for retarding timing. This is why a lot of guys with isfs run a mix of e85 to prevent this. I bet people running 87 in their lx will have lower KCLVs than people running 93 or higher.

 

[TeCKis300]

I think @TeCKis300 and I are at the same juncture that we always arrive at.

He looks more at the science of tuning.
I look more at what I see under real world test conditions.

Either are right, neither are wrong. They actually require both to be the best.

That said, I’ll still beat @TeCKis300 around a track. Muhahahhah!!!! Just playin’ (or am I?)

Alright... I'm going back over to bob is the oil guy. Read about molecules some more.

1700hp vs 700 hp. I concede...

 

[TeCKis300]

All that is true for designing an ECU program (I’ve programmed everything from piggyback controllers for OEM ECUs for motorbikes to MoTech M880s for turbo awd cars - I made my living in motorsports for 5 years before I determined banking was safer and more lucrative) but I think you’re massively complicating this - you don’t need the inputs to tell you anything about the resulting ignition advance, you just need the max degrees before TDC - a single output - for a given protocol from an LC running 87 and an LX running 91. If, over the course of a similar 50-mile protocol, the LX shows a higher degrees before TDC, it’s running more aggressive spark advance. At that point, you’ve proven your hypothesis that the LX is tuned for higher octane gas. Then it would also be interesting to see what the LC does on 91. You could obviously map them against RPM/load/GPS location etc and get very fancy, but max degrees is the only measurement you need.

Like I said, I think that’s likely what you’d find, but I wouldn’t be surprised to learn they were the same either.

Pretty cool! Glad to hear someone else that's been in the tuning world. What kind of turbo AWD cars? I can guess what it might be.

I've tuned a bit myself. Haltech, AEM, and piggybacks tuning Lexus and Toyota's 2JZ's. I find the OEM stuff more interesting because of the complexity and lengths they go (Bosche DME in Porches), but have never seen the encrypted workings of Toyota's implementation. I have a team currently building engine management for aircraft, but that's potentially an easier problem without emissions concerns.

I agree your suggested 50-mile protocol could work to suggest if there is advance or not. One would have to carefully control for other variables. Driver, weather, etc, could just as easily influence the outcome. Though if one did it enough, sure, perhaps could see a pattern. Perhaps better would just be straight up dyno pulls or drag racing as it's WOT, eliminating driver as a variable. It would even be better if we could just read-out the maps. Even in that scenario, depending on their control strategy, they could use variable outside the primary ignition maps to influence for octane.

 

[TeCKis300]

You don't necessarily need a dyno to test this. You'd only need to check your KCLV with both 87 and 93 to see how well the engine is running with each. Toyota is very conservative with their tunes which are known for retarding timing. This is why a lot of guys with isfs run a mix of e85 to prevent this. I bet people running 87 in their lx will have lower KCLVs than people running 93 or higher.

This is interesting. I haven't tuned any of the recent Toyota/Lexus and their control strategy seems to now give more readouts, including this KCLV (Knock Correction Learn Value). Any idea how to get at it and if it's a PID that can be read in OBD-II? Or is there a proprietary software that can read this out on the IS-F? Perhaps on other Toyota's?