Are you suggesting that the pressure spike from detonation which may be ten times that of normal combustion may be inadequate to overcome say 6 bar of oil pressure from the bearing and cause damage? If it can knock out a ring land, why can't it damage a bearing?

I would suggest that the phenomenon you describe in your second paragraph is more characteristic of pre-ignition. That is not the sort of damage I've seen from Subaru engines that have simply had detonation which they can live with for quite some time. On these you tend to see damaged ring lands, slight piston crown damage - often at the edges, and bearings that look well used even on young engines that have been hard used with new oil pumps.

 

Nice one John - The engines you have seen with bearing damage that you can definetly attribute to detonation - what form does this damage take?

localised 'pock' marks or flats in the upper shell from impact? , scoring that you also get from oil related damage which seems worse in the lower shell? etc..

i ask simply because my limited viewing of stripped down blown subaru engines (basically mine) was definetly not detonation related IMO, but have seen loads of blown bike engines that have either detonation damaged pistons, or badly scored crankpin and shells - but never seen both in the same engine IYSWIM i.e proof to myself that det can knock shells out.

Thank you for the replys on this thread By the way

 

I am best commenting on my own engines since I know how they have been treated - pushing the boundaries often results in some detonation even if not intended, and the EJ257 seems to avalanche into det in the way the EJ205, EJ207 and 4G63 don't - larger bore maybe? I've seen detonation knock out the headgaskets and crack ring lands, and minor damage to the corners of pistons. My bearings have always looked OK because I don't let them det much, and I think the ring lands and headgasket seal were the weak links on these examples. With forged pistons on Subarus I know the bearings can take a clattering because the rest holds up so well, but this could as well just be from very high cylinder pressures in high power engines than it is from detonation, or more likely a bit of both.

Maybe the big end bearing damage with detonation occurs over a longer time. I don't know, but I don't think it is controversial except to the idiots on here that we should avoid detonation, esp in the 150 to 200 BHP/litre range where it tends to cause a lot more damage than at the level most road engines operate. It does appear that with apparently good oiling a Subaru engine can lose a big end bearing, and it can also do it with no apparent detonation. It clearly isn't a one dimensional story, but Subaru have cross drilled the late EJ207 cranks, and surely not just for fun.

 

mate, 10 times normal cylinder pressure? several bar more than normal operating pressures due to det are enough to destroy an engine, 10 times would equate to around 10-15,000psi+ that'd probably lift the head off before the pistons had chance to melt or the big ends suffered....

Again, i reiterate people talk crap on these forums, i'm not even going to entertain comments like this anymore. I suggest you get some actual development experience using proper equipment such as combustion chamber pressure transducers before you make outragous statements like that.

....trems signing off

 

John - Does this guy trems know your background/experiences

I think not

Rob

 

It don't sound like it suggest the use of the search facility and to add insult to injury that **** pslewis is spouting ****e again

Cheers

 

"The pressure spikes of detonation can reach several thousand psi, and the pressure rise is rapid enough to be considered an impact load. These temperatures and pressures are almost ten times higher than those accompanying controlled combustion."

Perhaps you'd like to debate this with Corky Bell And before you state that this book is a boy racer handbook, I've seen it on the shelves of some rather impressive engine tuners with motorsport successes under their belts. This guy was a successful motor engineer before you were even born "mate"

 

You'll notice from my posts I don't spout about stuff I can't back up. I was also tuning Subarus before you could even drive What experience do you have of tuning Subaru turbo engines? How many 200+ BHP/litre turbo engines have you personally tuned?

 

"As a result of this very rapid release of energy, cylinder pressures briefly spike to several times normal, causing a characteristic sound which gives engine detonation its common names."

Here we've got "several" times normal. Perhaps you'd like to edit this one as it is in a WIKI format. You could educate them a bit.

 

John,
have you found any marked variation between different pistons/heads and their resulting squish/quench areas in regard to their det characteristics?...I'm thinking of building a 400 plus motor based on the 04 sti and basically would value your input regards the best piston to go for on such a build...
Or do you think the pistons/heads don't affect this much due to turbo engine characteristics?

 

crusher, my interest is in mapping, not engine design so I'll not comment except to recommend Roger Clark Motorsport if you want an engine built as they have huge success in making sweet running, reliable, high power 2.0 engines. I do think pistons/heads are important on a turbo engine, but for 400 BHP I would attempt it on the standard unopened engine. If primarily a road car or for a bit of fun, the expense and downsides in driveability/reliability in chasing 500 BHP aren't worth it IMHO. Or do what I did and just make 400 BHP on an Evo - it is so much easier on standard parts except the exhaust, possibly the fuel pump and a remap. Again, basic knowledge here, but the 4G63 seems to have a faster burning combustion chamber which requires much less ignition advance. This gives less time for detonation to develop. Long stroke, twin scroll and ace stock turbo also help of course in terms of overall driveability - feels like a TD04 on a Scooby - 1 bar at 2500 RPM etc. Sound like a washing machine though.

 

John,
wasn't thinking of going over to the dark side just yet!...Seriously though, it's interesting waht you say about the 4G63...even though your not into engine design it was observations like that re scooby that I was obviously looking for...wondering how best to spec build to try to build in some sort of det resistance if you see what I mean...I have built many engines in the past for other marques..and by trial and error ( was rallying at the time) came across various good combos...This time around was looking for a shortcut using the valuable knowledge of others!...as I don't want to spend the time blowing up motors if I can help it!...
Roger Clark has popped into my head a few times though!...

 

Three Subaru engine setups would have interested me more had I carried on with Subarus... Type 25 although I've not driven it, from speaking to Iain about the work they've done on the heads and the bottom end it sounds just brilliant - strong and efficient. And then RCMS work is just sweet as well. And then of course the Crawford 2.33 which spools like a 2.5 but doesn't seem to break when you throw stupid boost and nitrous at it. Just plain scary! May well be Andy's setup as well, but those engines seem to be so damn efficient and take silly boost on pump fuel without detting, there is something just right about it. But I really wanted a Supra engine in an AWD car I like factory standard engines that just do the business.

 

I use my knocklink i a sleightly differant way as i've taken a output from it just befor the signal gets to the led driver IC.
I then log the output with the Autronic so i get to see were in the maps its causing a problem.
I'm using a stock subaru knock sensor in the stock location with the knocklink gain turned right up.
The other day i got a bad bunch of fuel from bp and the result was as follows.


The bottom red line is the recorded knock noise from the knocklink and my crude drawing is how it shows up on the led's. Its is possable to have knock within the engine noise higher in the rpm range as you can see small amounts on the spectrograph.
Remembering every engine will have a differant amount of noise, thats why they give you a gain control.

Here is the wav file of the knock the so called useless Cocklink recorded.
http://www.sharebig.com/d/ssruvrqrc/...3/182knock.wav

 

Nice methods and results megrac. Power FC knock readout is similar.

Today I'm installing a TurboXS knocklite. Single tricolour LED in a more "attractive" casing. I'm tapping it to the original knock sensor and it takes an RPM input. Then you let it learn the background engine noise at three RPM levels and then it will flash for two severity levels of detonation thereafter with adjustable sensitivity if you need to. I think it will work quite well. Also doubles as a shift light, which the likes of pslewis doesn't need as his nice standard car wouldn't slam into the rev limiter in a hurry




http://www.turboxs.com/more_info.php?ID=212

No idea how well it will work yet. Sometimes through det cans we hear a little click at high revs, but where we mounted the det cans it could be releasing the throttle cable.

 

interesting reading guys,

its nice to see a genuine discussion about something that has been regarded as "chav crap" but can be used in such a number of ways to help protect the engine and your wallet.

 

So, can we settle at least one thing?

CockLinks are not at all useful for a standard Impreza? I think that is what I have worked out after all the years banging my Top Engineering Specialist head against a brick wall ........... The 'Tinkerers' will NEVER listen!

The Hobbyists, *****, Chavs and Rally Wannabbeees will ALWAYS maintain that their cars are so, so, so special that they need ever-so special kit to make them run at all - and that may be the case, they have ruined their engines by modding to such a point that they **** themselves whenever they turn the start key .........

I think the best thing for these would be a BIG RED LIGHT that lights up whenever they open the bonnet - it could flash too - this would tell them to CLOSE THE BONNET AND LEAVE THE THING ALONE!!!!!

Let the Hobbyists 'tinker' .... us true Engineers will continue to run our cars within their design limits - and to do that you don't need a poxy CockLink!

Pete

 

When you do your own home grown tuning you need this type of stuff more than anything.

John: i think the turboxs light should be better than the knocklink as the flashes will stand out more and shouldn't turn on with noise if it pulls off what it says on the box.
From looking at the spectrographs it still apears that there is det under the noise level and the spectrograph seems the best way of showing this up, my ears are crap as the cars to loud and driving it in the wet is not that fun.

If i was a mod on this forum i would have fun with the delete key

 

PsLewis's user account?DELETE

 

Excellent posts Mark/John Very informative responses.

 

John, it's the type 25 and the RCMS that I like too...
megrac...I like what you've got there...but also like the look of the Turbo XS Knocklite...I particularly like the idea of a unit which records a baseline noise level of individual setups...seems much more likely to spot det early..John let us know how you get on...

 

Nice idea, but doesn't work. After teaching it the noise levels either free revving, at cruise or with low or high boost, it couldn't be desensitised enoughto not light up when there was no det, so I'm selling it. I'll be getting a knocklink and Bosch sensor again - it is just as good as det cans properly interpreted.

 

Thanks for that plot megrac, can see why it is a bit hit and miss with thresholds.

I've put a wanted for a knocklink in the wanted section. If you sell your knocklink to me, pslewis will become your lifelong friend and I will give you money for your trouble