When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
It goes from octane and knock to posting pictures of our trip meters and MPG's.
Not sure if it's directed to me and my post, but I assure you my training plus ASE certifications didn't come by reading mechanics for dummies. I stated what I know through years of training and hands on experiences, I no longer care to get involved in fuel discussions that goes no where due to it being always one group being far left and another group being far right.
I posted up my readings as to contribute to what another member posted with his results of using regular at x speed and x miles. I really don't care anymore, you guys can put in whatever makes you feel better and post up whatever results and findings as needed. nfnsquared as well as some others are going to use regular and find nothing wrong with it as myself and other will use premium and see it our way. We can all agree to disagree. This is my last post in this thread. Have a good day.
Well, I got my ScanTool bluetooth reader. I'm using Torquescan to try to find the knock retard PIDs. Does anyone know what these are offhand? 6spd, does Hondata show what they are? I tried using the program in Windows, but I can't get it to link with my device.
Well, I got my ScanTool bluetooth reader. I'm using Torquescan to try to find the knock retard PIDs. Does anyone know what these are offhand? 6spd, does Hondata show what they are? I tried using the program in Windows, but I can't get it to link with my device.
Have you tried pushing the "power" button, on the device?
Well, I got my ScanTool bluetooth reader. I'm using Torquescan to try to find the knock retard PIDs. Does anyone know what these are offhand? 6spd, does Hondata show what they are? I tried using the program in Windows, but I can't get it to link with my device.
I'm not sure what the knock/timing retart related PIDS are for our ECU. In fact, I wasn't even sure that Torque Pro could read them, but I assume it must based on the fact that Accura-OC did get a reading off of the app...
Yeah, well, I left TorquePro Scanner running for an hour with no luck. This bluetooth OBD device isn't very good, btw -- the connection is not very solid, physically.
I did, but from what I read about the app it doesn't directly measure ignition retard. It's taking guesses based on analysis of other operating parameters. It's also not clear how to use it -- it wants me to step on the gas, full throttle?
I did, but from what I read about the app it doesn't directly measure ignition retard. It's taking guesses based on analysis of other operating parameters....
Yeah, like I said earlier, it's not clear to me how it does it. It says it uses a heuristic algorithm to find detonations based on analysis of timing advance, throttle, revs and speed. But it also says it will report ECU learned detonations. So, it's a bit confusing to me.
Originally Posted by RustyLogic
It's also not clear how to use it -- it wants me to step on the gas, full throttle?
3rd gear, start at low speed, WOT until redline or the app completes....
Well, what I've now done is recorded the ignition timing for driving in varied conditions for 45 minutes using 87 octane. I also recorded engine load, rpm, throttle positions, fuel trims, AF ratio, air intake temp, O2 sensors, etc.
Next I'll do the same recording for 93 octane, and see if there are any significant differences. The ignition timing for 93 octane should be the "baseline" behavior. Any consistent changes I see from this with 87 would likely signify ignition retarding due to knock sensor. I'll update with results and datalogs once I have them completed.
ECU on our cars has a memory even when you fill up with premium initial readings will be base on past performance. I do not know weather it takes certain number of miles or time but there is a re-learning curve. If you'd like you can reset ECU before running new tests with 93.
One thing everyone (on both sides of the argument) have failed to mention is HOW the engine determines its timing, and what effect that has on the engine's life.
No matter what anyone wants to believe, the statement made previously that, "if there is no knock, increased octane provides no benefit." End of story - you can't get around that. The goal is to use the lowest octane that prevents detonation. Octane is resistance to self ignition. Anything higher than what you need is wasteful and provides no performance gain.
This can be demonstrated in many boosted applications on E85 - the octane rating is so high, and the added fuel contributes to cooler cylinders allowing one to continue to increase timing well past MBT. I've seen instances of 10 or 15 degrees timing past MBT without knock on a boosted setup, with NO gain in power. Some people think that higher timing is better for some reason, but I'd say it is the opposite, to an extent. Higher timing means the engine is less efficient. Remember, timing is BTDC, so the higher the number, the earlier it is igniting before the piston reaches top dead center. In an ideal world, you'd wait until the cylinder was fully compressed before igniting it, or 0 degrees.
If you have a 1970s SBC with 8.5:1 N/A motor, running 93 will only empty your wallet faster, it won't provide any boost in performance, at all, ever. (If you add a 80mm turbo with 25 psi, that changes things a bit )
However, with the TL, and 11:1, the timing maps within it are already designed for MBT, which in most instances cannot be achieved with even the garbage 93 E10 that most people have access to. So, what happens? It knocks, the knock sensor pulls timing. What I have had trouble determining, (I think every tuner is in the same boat here) is how LONG/what algorithm does it use for adding timing back?
Subuarus have a well explained knock control strategy as well as 'timing recovery' strategy, so it's easy to figure out.
The TL on the other hand, may constantly be trying to add timing back, detecting knock, and then retarding timing. If that is the case, then running a lower octane fuel can decrease engine longevity.
Will running 87 octane in an 11:1 N/A motor with a knock sensor put a hole through a piston? No. Will it add unnecessary bearing wear from increased knock/detonation.
In that case, it would be wise to run the lowest octane available to prevent detonation, which in the case of the TL's programming, is 93.
Finally, a reasonable and informative post!
While we can get 93 here in the East, unfortunately, out west it's 91. Can anyone here explain why that is?
While we can get 93 here in the East, unfortunately, out west it's 91. Can anyone here explain why that is?
.
.
91 here in Colorado is due to the altitude. The higher the altitude the less oxygen that's in the air. Hence no need for higher octane with reduced oxygen available.
Allright guys, here we go -- even though I couldn't directly measure knock retard, I was able to infer it by comparing data. See below:
Scatter plot
Green = 93, Blue = 87, Red = Difference
In both charts, we see ignition timing (degrees) as a function of engine load (%). First chart, 87=green (83 is typo). Second chart, 93=green. The second chart is an average of the first chart, except that blue/green is flipped (thanks matlab).
Three observations:
- At higher loads, engine timing is retarded, regardless of 93 or 87 octane.
- 87 is clearly retarded more than 93 as a function of engine load.
- According to second chart (red line), average retard 87 vs 93 is around 8-10 degrees.
Last edited by RustyLogic; 02-01-2015 at 12:58 PM.
Old analysis assumed distribution of RPMs across trips was the same.
New analysis: compensation for variations in RPM between trips.
I analyzed timing as a function of RPM and engine load (2D histogram), measured the difference between octane ratings, and compare only parts of this histogram that were sampled in both trips, for loads above 45%.
The measured delays (retard 87 vs 93) are plotted above. Smaller than the first analysis. Still averaged over RPMs, but equally weighted across RPMs ranges that were experienced on both trips.
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 about 50%.
2) Does 87 retard timing more than 93?
Yes. By about 8-10 degrees on average after a load of 50%.
3) Does relying on the knock sensor cause engine damage?
Well, 93 relies on the knock sensor under higher loads, as shown in the above charts. So "relying on the knock sensor" to adjust timing does not, in itself, lead to engine damage.
4) Will using 87 long term damage my engine?
As shown in the chart above, knock retard using 87 only leads to additional retards outside the realm of 93 when the engine load is above 70%. If your engine load stays below 70%, we can show conclusively that you're not going to cause any damage, because it's still "in spec".
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 -- perhaps using the automatic shifting prevents this. 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.
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 fine. 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, you're probably fine.
5) What was your intake air temperature for your experiment?
During motion, the IAT was about 50-60F on average.
6) Any other observations?
Fuel trim values appear to be higher using 87 compared to 93. Will try to do a more thorough analysis on this eventually. AF ratios are the same, however. O2 sensor readings were the same.
Last edited by RustyLogic; 02-01-2015 at 02:08 PM.
Good info Rusty! MATLAB!! I haven't seen that for a while, actually haven't seen it since I graduated, great program.
I did one log today on 91 octane and didn't encounter any knocks but I didn't exceed 50kPa or 50% load very much. When I did the ECU was adding knock retard, by small amounts, 5 to 6 degrees at most an for very brief periods. My IAT's are significantly lower than what you are taking readings at, I am at 33 deg F in town and low 20's deg F on the highway. Cruising on the highway at 70mph I was at about 46% load and running at 29 deg of IGN timing. Weather forecast is cold for the rest of this week and I will be filling up with 87 soon to see what happens.
Originally Posted by 04WDPSeDaN
Not sure if it's directed to me and my post, but I assure you my training plus ASE certifications didn't come by reading mechanics for dummies. I stated what I know through years of training and hands on experiences, I no longer care to get involved in fuel discussions that goes no where due to it being always one group being far left and another group being far right.
I posted up my readings as to contribute to what another member posted with his results of using regular at x speed and x miles. I really don't care anymore, you guys can put in whatever makes you feel better and post up whatever results and findings as needed. nfnsquared as well as some others are going to use regular and find nothing wrong with it as myself and other will use premium and see it our way. We can all agree to disagree. This is my last post in this thread. Have a good day.
It was directed at you, I didn't mean to get you bothered by it but Rusty was looking to see what would happen with knock when running lower octane fuel in cold ambient temperatures. Posting trip meter pictures doesn't help Rusty get to what Rusty's looking for.
Everyone argues over what fuel to use in the TL. It didn't seem to me that Rusty was trying to sway anyone from running 87 or 91 or 93. Rusty asked a very valid question, do you think you could run lower octane in specific scenarios? I figured since I live in a cold area and have FlashPro I could help Rusty out, that's what these forums are all about. I run 91 octane in my car because that's the most commonly available premium fuel in my area and it doesn't contain ethanol, at least where I fill up. I prefer to avoid ethanol if I can, unless I had a FI application, E85 . I also use 91 in my GSR
I should mention that it was extremely windy when I was doing this, very heavy cross winds.
Ah, well, that could very well account for the difference: cross winds/head winds do put an extra load on the engine. I'm surprised that it would cause 5 degrees of difference though.
I'll be able to get a direct comparison between a headwind / tailwind along the same stretch of road later this week.
Last edited by nfnsquared; 02-02-2015 at 12:03 AM.
This was the last 2 gallons of a second consecutive tank of 87, so if the ECU requires time to adjust to an octane change (and I don't believe it does), it had ample time to do so.
Interesting. Today, I was running at 34.5 deg of IGN timing, cruising at 74 mph on 87 octane. Temp was -5F.
I refilled today with 91 (16 gallons) and will check timing tomorrow while cruising at the same speed.
Hi nfn, do you have a bluetooth OBD2 connector? Do you think you might be interested to also log your engine load, ignition timin, and IAT on a couple of your trips and compare 87 vs 91? We can see if we get approximately the same results just by eyeballing it.
If a few of us do this at varying IAT and altitudes, we can get a better sense of how the octanes compare under varying scenarios.
It would also be really helpful if we could share our raw data measurements. I could put these up somewhere, they're in CSV format from TorquePro.
Hi nfn, do you have a bluetooth OBD2 connector? Do you think you might be interested to also log your engine load, ignition timin, and IAT on a couple of your trips and compare 87 vs 91? We can see if we get approximately the same results just by eyeballing it.
If a few of us do this at varying IAT and altitudes, we can get a better sense of how the octanes compare under varying scenarios.
It would also be really helpful if we could share our raw data measurements. I could put these up somewhere, they're in CSV format from TorquePro.
Yeah, I can do that. How big does that CSV file get? I have 64GB storage on my phone, so I doubt that file size will be an issue... I'll data log on this tank of 91, then run a tank of 87 with data log.
Good data points.. indeed it does take time to adjust. Inacc confirmed several years back in one of his write-ups. I think 2 tanks would be sufficient. Keep it coming.
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 about 50%.
How can you be certain that the timing retard was due to the knock sensor being activated?
Originally Posted by RustyLogic
...Clearly, the knock sensor is being used even when using 93 to keep the engine in proper operating order.
Again, how do you/we know that the knock sensor actually detected knock? Timing retard under load isn't always due to knock, correct? It could just be due to standard mapping tables, correct?
Anytime load is increased, timing will retard. It's not necessarily due to knock.
And one last question: What's the difference between "Engine Load" and "Engine Load (Absolute)" on TorquePro?
Originally Posted by Majofo
... Inacc confirmed several years back in one of his write-ups. I think 2 tanks would be sufficient. Keep it coming.
Correct me if I'm wrong, but I believe Rusty ran back to back readings (e.g. 87 tank, then 93 tank), so apparently the ECU can and does react immediately?
Last edited by nfnsquared; 02-02-2015 at 10:26 AM.
Good data points.. indeed it does take time to adjust. Inacc confirmed several years back in one of his write-ups. I think 2 tanks would be sufficient. Keep it coming.
According to Inacc posts it usually took him 8 miles on average to see timing reverse back to "norm" at the same time he confirmed that timing was not pulled across all conditions. For example if timing was pulled at 50% he could read same adjustment for next 8 miles at the same time there wouldn't be any adjustment at 45% load. Not sure if it make it clear, anyone who's interested can google his thread.
Correct me if I'm wrong, but I believe Rusty ran back to back readings (e.g. 87 tank, then 93 tank), so apparently the ECU can and does react immediately?
It appears to be a real-time feedback loop but it stays conservative. The adjustment is weighted on the safe side, so it may stay conservative from one tank to the next. I think two tanks is sufficient though.
I learned something valuable for you this morning. I remember reading in the "Valve Chatter/Pinging/Knocking in 2007 TypeS" thread that you wondered how quickly or slow the ecu would replace timing once the knocking was gone. I have some helpful info.
For my morning commute and evening commute yesterday, I had my pseudo sensor in a position to retard the timing.
But last night, I moved the sensor to a place that should give me no retard (yet to be confirmed). On purpose, I did *not* reset the ecu to see how the ecu would react.To my surprise this morning, the ecu was still remembering from yesterday and severely retarding the timing. I thought to myself "oh crap, I have to drive this wet noodle all day." There is a big difference in part-throttle power between full timing and the retarded timing. Easy to tell the difference.
After I had come to a stop after traveling 8 miles at a steady 55 mph, I was shocked to feel that most (but not all) of the power had been returned.
The ecu returned most of my timing after 8 miles. So, somewhere between getting on the main road and 8 miles later, the ecu returned most of the timing. It takes me about 2 miles from my driveway, thru the subdivision, smaller road, before I hit that main road, where I traveled those 8 miles. The car was a wet noodle (poor bottom end) for those first 2 miles.
I just ran a quick 16-mile highway run (8 miles each direction N/S). The winds were west/nw at ~18 mph, so I had a quartering headwind going N and a quartering tailwind going S. I could tell no change in timing vs a highway run using 87:
This is with 1 gallon of 87 and 16 gallons of 91 in the tank and 10 miles driven on that mixture before running the test. I'll keep checking after more miles have been driven on this tank to see if the timing changes any.
For one, not ALL preignition is audible….one time, when I had my race GSXR 1000 apart - for other reasons- I noticed pepper like small holes marking the pistons (especially by the valve cutouts). Puzzled, I asked a pro mechanic friend (with lots of experience) what was up. "Preignition", was the simple answer. NEVER heard a thing! Right away, I changed my fuel from 50/50 race gas/ pump fuel I had recently gone to (I'd read somewhere it was a cool thing for power) to a blend of two race fuels. Never saw the situation again.
I will go with the general outlook that 11:1 CR is not for 87, no matter where, or when.
LOTS of folks have weighed in on this thread, but I will add a side story. I have a '97 4wd T100 pickup. 3.4 V6. Runs as nicely as when I bought it with 83K miles (298K now). Found a great mechanic, when I moved here 2+ years ago. He was amazed it was still running. He told me every other 3.4 he'd seen blew at 150K. Well, despite the hefty price differential, I feed it premium. Anything else, and it has audible preignition. My guess is all those other folks were going with 87, despite the warning from the engine. The preignition must be eating holes in pistons= Kaboom!
How can you be certain that the timing retard was due to the knock sensor being activated?
Hi nfn,
It appears to me the engine was designed to always try to be squeezing for top performance, using the knock sensor to avoid knock. In this manner, the engine is always running at the upper cusp of performance, no matter what fuel you put in it.
You'll notice that timing never goes above 40 degrees -- that's obviously rule based. Below 40 its a spread -- that's likely based on tables and operating parameters. At higher loads, past some threshold, timing begins to delay linearly (on average statistically) with the load... that type of behavior looks like knock sensor to me. And the behavior is amplified with 87, which makes me suspect its the same knock sensor process in 93.
6spd also mentioned that he detected some knock with 91 above 50% engine load, which lends credibility to this conjecture.
We need to do 1 of 2 things to conclusively validate and prove that it is knock retardation occurring even at 93:
- Run 93 in a hot climate and see if the curve shifts.
- Get more knock sensor readings.
This process has been really interesting to me. 87 gives you a certain operating point. 93 gives you another. 91 (the designed octane) would probably put me somewhere in between (though I cannot purchase it here). What this tells me is that this particular engine was designed to always be trying to deliver the driver the very best performance possible -- in order to do that, it's always putting the fuel, whatever fuel you use, on the edge of knock conditions past a certain load. That would mean the knock sensor is not "optional" equipment on this vehicle. If it breaks, you have to fix it ASAP or you risk engine damage, essentially. UNLESS (and I wouldn't be surprised), if the ECU was designed to, by default, go to full retardation if the knock sensor stops functioning.
Last edited by RustyLogic; 02-02-2015 at 05:08 PM.
That latter question can be easily tested by simply unplugging the knock sensor and measuring the difference in timing, before and after. I can't imagine Acura's actually would be designed to not, by default, go to full timing retard.
... but I'm not willing to do that experiment. Maybe someone has experience from it actually going bad?
-------
P.S. Found a video from EricTheCarGuy, looks like that's what the default behavior is, even in cars that don't require 91 (see video at 4:10):
Btw, note that hitting the engine with a hammer does not trigger the knock sensor. It's because the computer is designed to recognize a particular frequency, tuned to the engine, at thresholds below levels that cause engine damage. (See chapter 4.2 of book below):
It appears to me the engine was designed to always try to be squeezing for top performance, using the knock sensor to avoid knock. In this manner, the engine is always running at the upper cusp of performance, no matter what fuel you put in it.
You'll notice that timing never goes above 40 degrees -- that's obviously rule based. Below 40 its a spread -- that's likely based on tables and operating parameters. At higher loads, past some threshold, timing begins to delay linearly (on average statistically) with the load... that type of behavior looks like knock sensor to me. And the behavior is amplified with 87, which makes me suspect its the same knock sensor process in 93.
6spd also mentioned that he detected some knock with 91 above 50% engine load, which lends credibility to this conjecture.
We need to do 1 of 2 things to conclusively validate and prove that it is knock retardation occurring even at 93:
- Run 93 in a hot climate and see if the curve shifts.
- Get more knock sensor readings.
This process has been really interesting to me. 87 gives you a certain operating point. 93 gives you another. 91 (the designed octane) would probably put me somewhere in between (though I cannot purchase it here). What this tells me is that this particular engine was designed to always be trying to deliver the driver the very best performance possible -- in order to do that, it's always putting the fuel, whatever fuel you use, on the edge of knock conditions past a certain load. That would mean the knock sensor is not "optional" equipment on this vehicle. If it breaks, you have to fix it ASAP or you risk engine damage, essentially. UNLESS (and I wouldn't be surprised), if the ECU was designed to, by default, go to full retardation if the knock sensor stops functioning.
So if what you are saying is true, and the engine does it's best to squeeze every ounce of performance out from the fuel it is sipping, is it safe to conclude that on a lower octane gas such as 87, the power output of the engine is lower?
So if what you are saying is true, and the engine does it's best to squeeze every ounce of performance out from the fuel it is sipping, is it safe to conclude that on a lower octane gas such as 87, the power output of the engine is lower?
Yes. However, I'm not necessarily convinced you can conclude it lowers mileage. From my own experience I haven't seen that in my data yet.
It appears to me the engine was designed to always try to be squeezing for top performance, using the knock sensor to avoid knock. In this manner, the engine is always running at the upper cusp of performance, no matter what fuel you put in it.
You'll notice that timing never goes above 40 degrees -- that's obviously rule based. Below 40 its a spread -- that's likely based on tables and operating parameters. At higher loads, past some threshold, timing begins to delay linearly (on average statistically) with the load... that type of behavior looks like knock sensor to me. And the behavior is amplified with 87, which makes me suspect its the same knock sensor process in 93...
Not to be combative, but that is all pure conjecture. And my tests using the TorquePro Knock detector plugin would SEEM to prove that there is no detonation for the given conditions in my tests using 87.
Couple that with the fact that I see ZERO difference between timing advance at constant speed highway driving for 87 and 91, I'm still not convinced.
You're right about one thing though: we need a scanner that reads knock sensor output, but I don't even see a PID for that.
.
01-30-2015, 12:44 PM
)
. I also use 91 in my GSR 
Maybe someone has experience from it actually going bad?