Canuck

I have just learned from a reliable source that there is a fault in all three ignition maps for the P1/STiV, and ask whether anyone else is aware of this?

Apparently each of these maps is programmed to respond under different conditions to detection of pre-detonation, by retarding the ignition. In all of the cars so far examined by this skilled mapper (sorry, he wishes to remain anonymous) none of these maps in fact changes the timing at all!

Is this a common/standard fault? I am in touch with Subaru at the moment to see whether they are aware of this and am awaiting a response. If true, how should they respond (particularly if some of us are running either outside of warranty or with the existing one rendered void by letting the cat outa the bag!)?

Canuck

Pavlo

Can you please elaborate on the "none of these maps changes" statement?

I take it you mean the high octane map, low octane map and ignition correction map?

The maps themselves will not change, the ignition timing changes in accordance to the fixed maps.

However, you do need for the maps to be appropriately defined.

I have read reports here that the ignition correction does indeed work, and you can observe this by using Deltadash or a select monitor.

Paul

Canuck

Pavlo, good question, mate - I'm afraid at the moment I am not clear on this issue, and definitely NOT technical. My vague understanding is that this has to do with the timing changes programmed into [/b]each[/b] of those fixed maps - that those changes in fact DO NOT alter the timing. This was picked up by mapping P1's with ECUTEK.

At the moment I do know that Subaru are aware of this post I've made here, and will be getting back to me with more information.

[Edited by Canuck - 11/27/2003 4:52:54 PM]
edited cos I'm in a hurry and can't type!!

[Edited by Canuck - 11/27/2003 4:53:40 PM]

johnfelstead

You have you knickers in a twist Cannuck and are not informing people of the real configuarion used in the STi5/6 and P1.

There is no Fault! Whoever told you that is being iresponsible and scare mongering, trying to cover their **** due to probably not knowing the configuartion properly before mapping an STi5/6/P1.

The STi5/6/P1 has a diferent configuration in use in its ignition timing maps than a UK/Euro spec car.

You have 3 ignition base maps in a JECS ECU (UK and Jap spec), Low Octane, High Octane and High OctaneB. You also have knock correction maps that are used to actively control ignition based on the detection of knock from the knock sensor, this can advance and retard ignition around what the base maps are set to.

On a UK spec Jecs ECU, the timing in the Low Octane and High Octane maps are diferent, when a UK spec ECU senses knock, it uses the knock correction map values to remove timing, if it cant remove enough via the knock correction values, it switches to the Low Octane Ignition map, which has a less advanced set of ignition values, this takes over ignition control and again, uses the knock correction tables to dial in even more retard if needed. This is how the UK turbo can survive on 95RON fuel.

With a P1/STi5/6 all 3 ignition maps are identical. When you map an STi5/6 or P1 you have to match all 3 ignition maps to keep that corelation, otherwise if you were to leave the Low Octane ignition map alone, as is common practice on the UK spec ECU remaps, you could actually leave your ECU in a state where the Low Octane map is more advanced than the high octane map, which could be disasterous to your engine.

Your mapper should know this, if he doesnt and doesnt remap your Low Octane ignition map to match the high octane ignition map, then he isnt doing the job properly and shouldnt be touching your car.

The above should explain why you should only ever run your P1/STi5/6 on SUL or Optimax and never use 95RON, no matter what the blurb you recieved said when you bought your P1. Remember ECUTEK is a relatively new product, and when the P1 was launched there was no facility to remap the JECS ECU.

P1's and STi5/6 use the same ignition base tables, they are slightly diferent in the active nature of the knock correction tables, being more active in controlling det, but they still should be run on SUL.

sg72

Does'nt matter to me now but

When I hasd my, at time standard STI 5 Deltadashed.

ECU was showing No knock correction.

This was on 2 occasions.

Even changed knocksensor after certain mods

Still no ECU correction.

Realised I could have just disconnected sensor and gt CEL light to find if original was being recognised.Doh!

Car had to be mapped to accomodate lack of correction.

Had SECS monitor fitted .

This also displayed ZERO knock correction.

I cannot explain what was going on .

I had also heard it was a fault in the STI 5 ECU

Now running mapped APEXI OK now.

Now able to pull 5 or 6 degrees ignition advance apparantly.

Optimax and NF (AS before).

johnfelstead

The ECU only dials in knock correction when it needs to, if your base map is spot on then you wont see any correction. The longer the ECU is run without an ECU reset the less likely it will need to use any knock correction, as it learns all the time and is storing that info in the ECU to apply against the base maps.

The early ECU's go deaf to knock above 6000rpm, the STi9 has a knock sensor capable of reading to the red line.

The STi7 onwards ECU's are very diferent in the way they work, they are far more active and you will see far more knock correction being used, both positive and negative, if you view this via the SECS display or Delta Dash.

sg72

I'm sure you're correct John.

But I have to say that my car was Deltadashed and Delta ECUed.

After ECU had been reset.

Due to work being done on car (NOT all Mods.)

It was in fact a reset that had caused my ECU to fail to compensate a faulty MAF and cause the destruction of engine and Turbo.(Really my fault as I new something was wrong but failed to address it!)

Cheers

Steve.

sg72

Also mean't to say.

It was after. New engine, different Turbo (VF 34), FMIC and Induction Kit.

When being Delta ECUed(ECUTEKED)

That it was spotted that ECU was failing in Knock correction.

Could Ignition map have still have been spot on after said Mods?

johnfelstead

A faulty MAF isnt going to be corrected for via the knock correction function of the ECU, so what you are talking about is a diferent issue. Sorry to hear the engine failed as a result, it's not a nice thing to have happen.

johnfelstead

I take it you decatted the car too. Those mods require a remap to be safe. An ECUTEK remap should be able to cope with those mods without any problems, with mods like the APS cold air kit you would need the MAF recalibrating via ECUTEK too.

[Edited by johnfelstead - 11/27/2003 7:21:11 PM]

sg72

Hi John

The mods done to my car now are EVERYTHING. But internal.(Have parts)

All mapped to suit.

The events I refered to in previous posts.
Happened earlier in the year.

It was during cars Tuning of new engine/Turbo FMIC Induction Kit that obsevations had been made as regards ECU.

Previously on Deltadash it was assumed car was spot on.

Yes Car had previously been Decatted and running K+N panel filter(YEP!).

Hadn't been mapped but running Optimax and booster at time.

sg72

Also aware that MAF and Knock are separate issues (Twas just an aside).

It's through reading many posts on Scoobynet.

By yourself and many others that I have acheived my ultimate SUBARU goals.(In such a short time).

I Think I can understand the technical issues.

And have indeed experienced many along the way.

Just I dont do Number crunching.

Seat of pants the way for me.

Steve.

johnfelstead

With a K&N you need to keep an eye on the MAF, as the MY99/00 MAF is pants, and unless you use dry filters like the STi filter i use, they are easily contaminated. When you decat a P1/STi5/6 you tend to get det spikes.

You should remap these engines even if you are just decatting them, the remap will control the waste gate duty and timing to work around the changes in eficeiency the engine experiences due to the changes in Breathing and exhaust flow.

Also bear in mind the P1 was designed around 98RON fuel, not 97SUL. The reality is though, that the timing maps in the ECU are designed for 100RON and the ECU is relying on improved knock sensitivity to make it safe for 98RON. Now you make the engine more eficient with a decat and you are pushing the ECU even closer to its designed in safety limits, it's not surprising that you could destroy it if everything isnt 100% on the engine or its sensors or you are running 97RON or maybe even worse. A remap done properly will put the safety back and not have the ECU running on its outer edges of control. A remap done incorectly makes matters even worse, which brings us back to the first post in this thread.

Bob Rawle

A few facts from my perspective.

JDM MY99-00 STi code is, as far as anyone has been able to determine (and I mean the best) identical to P1 EXCEPT that it has less sensitive knock correction. This can be clearly seen when mapping.

All JDM STi MAPS are identical to P1. Which means they are all based on high octane fuel, the P1 code then relies on knock sensitivity alone.

There is no fault, the ecu reacts to inputs according to its programming, on a P1 its more reactive to an input derived from the knock sensor. But that input is necessary to provoke a reaction.

All these engines are different and generate differing levels of "noise", accordingly one mans meat etc which means that when mapping you CANNOT RELY ON THE NUMBERS.

Det cans and a wide band are the order of the day. The factory ecu will react to "no det" and will ignore "det" however thats merely program function according to what it sees ... and no fault.

Ok so am I splitting hairs here .. no not in the least, I have, however, spent a long time driving my own car(s) with the oem ecu in place purely to try and appreciate the nuances.

The factory ecu has to be treated like any aftermarket ecu and mapped to suit the engine, relying on the ecu reducing ignition in defense is risky, which is why we fit knocklinks for driver awareness.

FWIW a Motec has no knock correction available to it as normal, so when you map that you have to get it right, a Link, on the other hand, has an almost too reactive knock correction system but is totaly dependant on the mapper setting the correct trip levels.

Whatever code is written in for this is, by and large, invisible to all of us Ecutek software users, its therefore simply not possible to state "fault".

It has been long since well documented on this and other BBS that the knock learning of the MY99-00 software is nefarious. Not a secret and no suprise.

Final comment, I have copies of the select monotor screens from the P1 supplement that shows knock correction under various conditions using ... 98 ron fuel, so P1 owners you should be using booster.

comments from other Ecutek tuners ... anyone ???

cheers

Bob

edited to add that there are more than three ignition related maps.



[Edited by Bob Rawle - 11/27/2003 10:28:14 PM]

sg72

Great Bob.

That covers it.

Basically.

Everyone. Watch everything like a Hawk.

Btw. Sorry I didn't go the Link Route(I may be sorrier still).

Thanks.

Steve.

currymonster

Just a couple of questions specific to my car really. I've just bought an STI5. It is fully decatted, has a link ecu, three port boost solenoid, hks powerflow induction and a lambda and knocklink.

I need to refit the standard ecu, airbox, itg plate filter and 2 port boost solenoid. Until I can have the standard ecu ecuteked would fitting a uk centre cat in place of the decat centre section go some way to protecting my engine against boost spikes?

Cheers Cal

Bob Rawle

Currymonster, may even induce spikes as its more restrictive, you can adjust and minimise by tweaking the restrictor once you have refited the two port solenoid, or just try using the three port with no restrictors anywhere except in the vent pipe which goes from the bottom solenoid port to the air intake.

bob

currymonster

Ok will tell the garage saturday. Need to book in with you for an ecutek sharpish I think your into march bookings already? eek!

Cal

IPALMER

iTrader: (5)

"link ecu, three port boost solenoid, hks powerflow induction"

Any reason why you're removing these mods ?

Iain

Canuck

John and Bob, thanks for your authoritative replies.

It was not my intention for this post to serve as scare-mongering and apologise to the SN community if that was inferred. Please appreciate that I can barely spell ECU, and was reporting this observation as a second-hand conversation. I will endeavour to determine that I have got my facts straight and report back, pronto.

Meanwhile, I have had a phone call from Customer Services at International Motors who have made it very clear that if what I originally posted was in fact absolutely correct there would be no P1s running on the road since they'd all have blown up! So far as they are concerned, consistent with what John has stated, there is no fault in the ignition maps in the P1/STi5,6.

I will post later when I have clearer information.

Keith

[Edited by Canuck - 11/28/2003 2:12:38 PM]

Canuck

Ok, lets see if I have got this any straighter?? I am not talking about knock detection, merely the way that the ECU responds to it.

The point I am trying to make, which in fact no one seems to be disagreeing with, is that "In the P1/Sti5,6 the 3 ignition maps are identical, which means that the timing in the low octane map that is the fall-back position when detecting serious knocking, is the same as the high octane maps". Now, I must be missing something here, but I understood that it was this Low Octane map that is meant to retard the timing to prevent pre-detting, which apparently according to everyone so far does NOT change the timing?? Now, so far as I can see it, either this means that these cars do not rely on these ignition maps to make this knock correction or NO correction is being made.

John and Bob, there is no question that re-mapping requires appropriate changes to be made to the 3 ignition maps, and we are not even going to get started about a discussion over the various compensation tables. The mapper who brought this to my attention in fact alters the Low Octane Map to make appropriate timing changes different from the High Octane Maps!. I just wonder why it is that the three maps are identical in the standard car - How do you use ECUTEK to respond to pre-detting if not to change the ignition table in the Low Octane map?

I have a sense that Steve (a.k.a sg72) is making a point similar to mine, with some first-hand experience (just cannot get my head around his "sentence" structure to understand the point he wants to make! Sorry, Steve!).

Canuck

sg72

OH DEAR!

Sorry about the way I structure sentences. Was even told at school I tend towards a German structuring.
I cannot help it. It has caused misunderstanding and confusion all my life!

This issue is all behind me now. But facts are...
STI 5 Type R ltd. AF041 ECU Decat and K+n filter.
Deltadashed. Showed up NO Knock correction.
Fitted Scoobysport Backbox and mid.
Deltadashed Showed up No knock correction.
Engine replaced with STI 6 unit VF 34 and FMIC.
During mapping with Delta ECU showed up No Knock correction.
Replaced Knock sensor Again No knockcorrection on Delta ECU.
Mapped to accomodate.
Fitted SECS 3 UNIT Showed up No correction ever on monitoring function.

That's it.

I can give no expanation. Just observation.

Steve.

Canuck

Steve,
das ist goot, ja?
Herr Keith

[Edited by Canuck - 12/8/2003 6:54:19 PM]

sg72

I did it again, on my third sentence of reply.

Doh! and again.

And again. Doh!

john banks

I mapped Steve's car after his original engine let go and agree with what he is saying, to be totally clear, I will add that even with fairly good going detonation the ECU never retarded where it should have done, it was also quite sensitive to tiny overboost causing transitional detonation. Along the same lines as Bob says the knock control algorithm is pretty patchy especially on modified cars.

Some P1 ECUs are fantastic on some engines retarding and advancing appropriately in real time and keeping the det (even transitional stuff) in check whilst running good performance. Others like Steve's (reflashed with P1 code) just sit on zero all the time.

UK ECUs are similarly variable, my own car has been non-sensical throughout its mods. On TD04 hybrid and big top mount it was fairly appropriate except for at high RPM where we all know they are deaf and to transitional det. With the bigger turbos it still let transitional det through and some real bad det too. With Scoobysport headers it was completely deaf to everything. On the 2.0 and 2.5 with Gruppe S headers it occasionally takes out a token 1 degree of timing when it is not detting, sometimes it does take out a bit and keeps it quiet though. I am moving on to individual cylinder closed loop knock control with a J&S Safeguard which will be an alltogether different ballpark I expect. Results to follow of course.

Interestingly the MY99 ECUs also have all three maps set to the same: High octane ignition map (+B), Low octane ignition map, but they are OK on 95RON because the timing is so safe. MY00 AE802 ECU had different high and low octane maps and that is why I recommend it to people wanting to run higher boost on manual boost controllers without mapping (with appropriate detonation monitoring of course).

When I was mapping other people's cars I tended to create a more cautious low octane map whether it was a UK or STi so that if it was switched to by the ECU it would be fairly safe.

Of note all the MY99/00 UK and STi/P1 ECUs seem to add the high octane ignition map to the matching ignition compensation map to get their timing assuming there is no detonation. They then seem to report dynamically as "knock correction" what timing they are taking off this total plus some sort of hidden learned adjustment, but since we don't know what that is and how often it updates the reports can be confusing. They seem to able to retard at least as much as the contents of the compensation map, and when I had my slappy engine a lot more - ie -8 degrees it would peg straight to under more than light load.

What is frustrating is the unpredictable and often inappropriate nature of the knock control. I reckon a charge temperature compensation to ignition timing would be far preferable to JECS knock control at least on modified engines.

Unfortunately when mapping on the road or the dyno you are only simulating a few scenarios and the amount of reserve dialled in for ignition timing away from the detonation threshold is a matter for judgement. Get it wrong and the car is underpowered, get it wrong the other way and if the car is ever ragged on a hot track you may regret it hence all the recommendations for knocklinks etc.

Whilst the knocklink is a very primitive device and the JECS knock control is undoubtedly technically far superior, in reality on a modified car the knocklink is far more useful. I would fit one on any Subaru modified or not personally.

johnfelstead

Incorrect asumption. With the stock settings on the ECU, with all 3 maps identical, you still do get knock correction. The amount of correction available is limited to what is available in the knock correction tables. Once the ECU has taken out its full available amount of retard, it will switch to the low octane map. If the 3 maps are the same, then the effect of that is that you dont get anymore retard than was available in the high octane map, but you still did see some retard. So even with 3 identical maps (P1/STi5/6) you do get active knock control, and this can be both positive and negative as the ECU is learning around the base map.

Bob Rawle

As John says, there is knock correction, the amount is controled by the ignition_compensation_maps of which there is low and high versions. Also, dependant on knock sensing, fuel is added as well from the fuel_compensation_maps again low and high versions.

Bear in mind that these cars are designed as high performance varients for Japan, reason the maps are the same is that using higher octane fuel there is no need for anything else, and the Japanese want the cars to hold their performance, however we bring them over here and run them on lower grade stuff, clever eh, oh and, to boot, we then delimit them and run them at high revs in high gears, which is not what they were designed to do.

So there is still no ecu fault or issue as such but, you have to treat the cars with TLC which means use booster for preference.

cheers

bob

sg72

Ta Bob!

Of course Everyone forgot about the Delimiting factor.
This would explain why even after fitting my aftermarket (Japan only ECU.),

I still have issues.

It would seem like another avenue.We have to investigate.

It is certainly something not all consider(Including myself!)Thanks for the Incite Bob/John/John.

This is Very important as regards the tuning of JDM. Imprezas.

I suggest this should be a mainline Topic.

Steve.

Steve

Canuck

So, how important is it when re-mapping to keep all 3 octane maps the same, or is there added value in dialing in some safety into the Low Octane Map?

Keith

edited to add this thought:

If there is no need for a Japanese car to worry about a functional Low Octane map, thus explaining why it is no different from the others, and if there was benefit in having a safety feature in the Low Octane map to retard timing, then especially for the P1 where some of us are making careers/hobbies out of mapping these damned things then why, oh why, has Subaru/Prodrive allowed this to slip through without dealing with it?? Particularly when these cars are apparently so bloody fragile!! (If we need to rely on accurate sensors to compensate for detonation, etc, then who'd put their trust in that other MY99/00 foible that Subaru are failing to deal with in the MAFF sensor ... !!??)

Exasperatedly,

Canuck

[Edited by Canuck - 12/11/2003 2:32:06 PM]