Guess who's running 87 octane

[SVXer95]

PON = AKI (anti-knock rating) = (R+M)/2

[Rotorflyr]

Wow, did this thread get off track.......
So how bout a little more info on things when you can (fuel econ, performance changes, etc)

[n00b on demand]

93 octane just went up to 2.65 a gallon here. I just paid 37 dollars for gas. One of the guys at the pump saw my bill and asked if i drove an SUV and i said no its not called an SUV its an SVX

[PhatWeasel]

Does the ECU use an external switch to select fuel grades, a la Mustang SVO, or does it need to re-learn for the lower octane?


EDIT: Answer discovered in a ANOTHER THREAD.

[longassname]

Originally posted by dcarrb
Someone correct me if I'm wrong, but I'm thinking anything you do to enable an SVX to run 87 octane gasoline will worsen fuel economy to the point that you'll realize no savings over running premium.


dcb

Ok, consider yourself corrected . I don't have any mileage information on the chip yet to share (actually i'm going to go with whatever consesus users here who have the chip come up with) but the changes made should not "worsen the fuel economy to the point that you realize no savings." If you read the information on stage 1 you will understand what changes were made and how they work then you will know what i'm talking about. Yes the fuel tables have been richened but it's a matter of getting a much better return for the little bit more fuel added. I just did the math to give you an idea of the %'s were talking about here. The stage 1 code is 7.8% richer than the factory tsb rom. The 87 octane code is 9.3% richer than the factory tsb rom. Whether mileage improves due to the performance improvement or goes down slightly it certainly won't be greatly reduced.

[longassname]

Originally posted by svxsubaru1
Did you gain any power?

The performance increase of the 87 octane code over factory is very noticeable though probably not quite as good as the base stage 1 code. You can read the full explanation on our website at http://www.ecutune.com/svx-stage1.htm

The paypal link to purchase is now there as well.

[UberRoo]

I've run the cheap stuff for a month or so without any ill effects other than a slight loss of power. There was no pinking or hesitation of any sort. I continued to fill my tank at 330 miles as I always do, (because it runs out at 345 miles,) and it continued to guzzle the same amount of fuel at each fill up.

I decided it was worth the extra dollar per tank to have that extra get-up-and-go, so I've been back on the expensive stuff again.

I think I might find a test strip somewhere and take some benchmark drag runs with different grades. Maybe I can run the middle grade without giving up any ponies. For the extra dollar, it's really not worth my time, but I'm curious so I might do it anyway.

[longassname]

You might want to read the info we provided on our website about the factory tuning of the svx. You WON'T HEAR NOCK in an svx unless it's so outrageously bad that it's out of the ability of the ecu to correct for it by retarding the ignition timing. YOU WILL GET A MASSIVE LOW END PERFORMANCE LOSS from the corrective measures applied to prevent knock with the stock factory code which gets worse as your fuel grade does. Further, once you have driven the car with the wrong grade fuel without using ecu code that is propperly tuned for it performance with the higher grades of fuel is detrimented because the protective measures learned in by the ecu take much longer to learn out than they did in. If you read the website I think you will understand. Right now you are just comparing poor performance with poorer performance.

You should strongly consider taking your car to the dealer and having them clear the memory on your ecu using their select monitor. Of course, after you stop running regular unleaded on the factory code. All those premium unleaded only stickers are there for a reason. The US market cars without our chip run considerably lean especially in the low rpms which causes knock retard and performance problems even on premium unleaded. Those problems only get worse with lower octane. Your ecu undoubtedly has a lot of performance robbing adjustments learned in to prevent knock. To clear them out you must use a select monitor.

Originally posted by UberRoo
I've run the cheap stuff for a month or so without any ill effects other than a slight loss of power. There was no pinking or hesitation of any sort. I continued to fill my tank at 330 miles as I always do, (because it runs out at 345 miles,) and it continued to guzzle the same amount of fuel at each fill up.

I decided it was worth the extra dollar per tank to have that extra get-up-and-go, so I've been back on the expensive stuff again.

I think I might find a test strip somewhere and take some benchmark drag runs with different grades. Maybe I can run the middle grade without giving up any ponies. For the extra dollar, it's really not worth my time, but I'm curious so I might do it anyway.

Originally posted by longassname

You should strongly consider taking your car to the dealer and having them clear the memory on your ecu using their select monitor. Your ecu undoubtedly has a lot of performance robbing adjustments learned in to prevent knock. To clear them out you must use a select monitor.

Just a note

The SVX ECU was programmed at the factory to run on high octane. A Select Monitor can erase data your car has learned since leaving the factory, but it can not erase factory programmed data! It also can not reprogram your ECU (in case anyone was wondering).


Dave

[longassname]

Originally posted by red95svx


Just a note

The SVX ECU was programmed at the factory to run on high octane. A Select Monitor can erase data your car has learned since leaving the factory, but it can not erase factory programmed data! It also can not reprogram your ECU (in case anyone was wondering).


Dave

That's correct. Ideally you should install the stage 1 chip then go to the dealer and get your ecu's memory cleared using the select monitor. If you don't get your ecu's memory cleared it will relearn to match the new code but it will take longer...particularly the learned ignition retard.

[UberRoo]

I don't doubt that what you say is true, or at least somewhat true, but I noticed an almost immediate increase in power once I switched back to the higher grade, and I believe it has completely returned to it's original map on less than one tank. Perhaps it has not, but I can't tell the difference. (It can still get from the stop sign to the red mailbox in less than 21 seconds.)

The US market cars without our chip run considerably lean especially in the low rpms which causes knock retard and performance problems even on premium unleaded.

It's reassuring to know that the ECU normally relies on the knock sensors to generate the correct mappings. I see no reason why the ECU should take any longer to unlearn one map than it should to learn another, and from my experience it's actually quite prompt about doing either. If it's normally dependent on the knock sensors, it should be rather competent at making these adjustments both ways.

The most valuable piece of information is exactly how long it takes. I'll put my money on less-than-a-tank.

[longassname]

Originally posted by UberRoo
I don't doubt that what you say is true, or at least somewhat true, but I noticed an almost immediate increase in power once I switched back to the higher grade, and I believe it has completely returned to it's original map on less than one tank. Perhaps it has not, but I can't tell the difference. (It can still get from the stop sign to the red mailbox in less than 21 seconds.)

It's reassuring to know that the ECU normally relies on the knock sensors to generate the correct mappings. I see no reason why the ECU should take any longer to unlearn one map than it should to learn another, and from my experience it's actually quite prompt about doing either. If it's normally dependent on the knock sensors, it should be rather competent at making these adjustments both ways.

The most valuable piece of information is exactly how long it takes. I'll put my money on less-than-a-tank.

That's incorrect. The ECU takes longer to unlearn knock retard than it takes to learn it in as part of a safety measure. The whole system is very competent at everything; however, it is not designed to operate in the manner you are expecting. While I don't doubt that your performance returns to what you consider normal I do still believe that your normal is considerably less than if you were to clear out your ecu's memory. I don't know what to tell you to convince you except that you should take into consideration that you are basing your conclusions on your personal assumptions while I am basing mine on an actual knowledge of the ECU software.

[UberRoo]

Originally posted by longassname
I don't know what to tell you to convince you except that you should take into consideration that you are basing your conclusions on your personal assumptions while I am basing mine on an actual knowledge of the ECU software.

So then exactly how long does it take to unlearn this behavior?

The ECU takes longer to unlearn knock retard than it takes to learn it in as part of a safety measure.

The ECU clearly unlearned a very large portion of the knock retard very quickly, as it takes no less time to reach the red mailbox than it did before. I don't have a precision timing device, but it's doubtlessly within half a second. Why would it unlearn only 90-something percent of the retardation?

Logical extrapolations and real experience are a far cry from personal assumptions. It is hard to believe that Subaru would design an ECU that was so nearly incapable of unlearning a low performance map, especially one that is so competent and heavily dependent on it's knock sensors. Who has the ECU's source code?

[longassname]

Originally posted by UberRoo

The ECU clearly unlearned a very large portion of the knock retard very quickly, as it takes no less time to reach the red mailbox than it did before. I don't have a precision timing device, but it's doubtlessly within half a second. Why would it unlearn only 90-something percent of the retardation?

Logical extrapolations and real experience are a far cry from personal assumptions. It is hard to believe that Subaru would design an ECU that was so nearly incapable of unlearning a low performance map, especially one that is so competent and heavily dependent on it's knock sensors. Who has the ECU's source code?

I think the main error in your reasoning comes from the presumption that your ecu obviously unlearned 90% of the knock retard. That is not at all obvious to me. I'd be more inclined to reason in the reverse direction and conclude that the previous performance you are comparing your current performance with also shows the performance losses from learned ignition timing retard. I'm more inclined to reason in that direction for a # of reasons. First as we all know the car comes tuned to run on 93 octane. Second tests done while on the dyno and monitoring with a subaru select montior reveal that the cars run lean from the factory particularly in the lower and mid rpms and that this condition made the engine prone to knock and triggered the knock sensors causing drastic reductions in ignition timing. Third the Subaru TSB specifically stating that while the car can be run on regular unleaded in time of emergency when no other fuel is available that the ecu will learn this knock retard in and that it will be very hard to ever learn it out without clearing the memory with a select monitor (please note that while the subaru/unisia jecs software engineers who provided the technical information for this tsb ar the ones with the actual "source code.") Fourth my personal knowledge of the functioning of the sofware in the svx ecu. While I do not have the original "source code" from before it was compiled into the bin loaded into the ecu, I do have the bin from code in the ecu and the manual for the mitsubishi m37xx processor the code runs on which specifies the opcodes allowing me to reverse engineer the software. 5th I'm driving a svx on 87 octane now running the code which alleviates that low rpm leannes and is optimized to run 87 octane on an ecu with the memory cleared of it's previous learning and DO know the very significant power differnce between it and the factory code. When you describe only a slight performance loss when going to 87 octane on your stock ecu programming the logical summation IS that you already had a significant amount of ignition retard learned in. Which isn't really suprising considering a very significant amount of ignition retard is frequent in many driving conditions on the factory code even when driving on 93 octane. When you press the pedal to the floor and for a second or two nothing happens....I think you know what I'm talking about here......that's ignition timing retard. Now throw in a bad tank of gas somewhere before your original perfermance assessments and you have great reason to think you might have a bunch of retard learned in. Say the gas station you go to doesn't sell the best gas in the world all the time...another good reason to consider you might have a bunch of retard learned in. Maybe you don't but to me it doesn't seem so certain that you don't that you shouldn't consider it.

I don't want to force my opinion on you. I just don't aggree with what you are thinking. It happens. Maybe you're right and I'm wrong. Maybe I'm right and your wrong. Personally, since I think I'm right, I hope sometime you get the opportunity to find out I'm right and get back some of that horespower I think you're missing out on.

[RojoRocket]

Michael, Any idea the going rate at the Dealer to do this "select monitor brain-wipe"? As the third owner of the car I wonder at this point what I may be missing.
Apologies for slipping into the middle of your technical vs "time to the red mailbox" debate with such a thing as a $$ question, but inquiring minds, and all that.

Glenn