I thought that was the point. SMART pulls smalls amount of timing to keep noise at a reasonable level so that a knock event isn't seen by the ECU. Thus the large timing adjustment is avoided.

Oh and I wasn't really referring to the magnitude of the event in freedoms post, I was just commenting on how distinct and unambiguous the knock event was.

d

 

I get you. With a knock event such as this I do not know if there would be a prelude to the knock itself for the SMART to pick up on and thus pull timing. To be honest, I do not know how the SMART can work any diferently then the stock ECU exept that when it see's a knock event it may not just flat out cut timing like the ecu but just try to trim the timing based on the sevarity of the knock. Maybe it works in a proportional manner rather then a bianary. I have seen a map that seemed to have some smaller knock events, but I do not know if knock is knock, or not. In other words, is there such a thing as mild or small knock that the SMART may be able to pick up and trim before the ECU even registers knock, or is all knock as distinct as the prior map shows?

 

Here are my crazy *** graphs

I have yet to get rid of this knock. I dont have a single datalog without it.

What do i need to do? you can see my AFRs arent crazy, even on the rich side. I've tried pulling out some timing, maybe just not enough. The knock moves around RPM ranges too. Maybe i have too much boost?

 

That's my understanding of it. The knock buffer provides increased resolution of the knock signal, so the noise level can actually be controlled/mapped rather than just set at a threshold. The stock ECU would never be able to do good short term timing advance because it can't really tell the difference between very low noise and just about to knock.

d

 

How much boost? What gas?

l8r)

 

Can you post your exact set up (including mods, base map, and octane)?

d

 

91 octane, pretty much always 76.

As for maps, I ran the 03-04 basemaps converted to smart for awhile, but tuned the fuel maps myself. This run should have been on a 05 map from the boards here, the only one that vishnu had given out before they updated their website. that map was rich on my car too and i leaned it out bit by bit, but still not enough. I have been tuning my own maps. mostly just leaning out afrs bit by bit, and removing timing bit by bit in the affected RPM ranges. I think i've just been making too small of changes.

As for boost, this is an 05 MR without the passive emulator. It hits 1.5 bar on the stock mitsu gauge. 100% on the upper rpms in smart boost.

as for mods, this run may have been with an AEM intake and a Type-RS BOV. Hold off on giving me crap for that, they will probably be coming off soon. AND YES, datalogs from before the intake/bov install are very similar, including the knock events. Only difference i can tell is that since the intake install i have lost a bit of boost according to the mitsu gauge (1.4 bar now or so) but the gauge also reads just below 0 when off, so the gauge may be off a bit. new Defi gauges coming soon.

so, aside from removing the intake and bov, anything else you would recommend? in two days i'll have a full turboback and go on to the stage 1 v360 map. hopefully that is closer to what i need.

 

Not to make you feel bad, but I can run peak boost levels of 1.64 kg/cm2 with zero issues even after multiple back to back runs. And when I cite peak boost I don't mean a spike up to 1.6, but sustained boost in mid-range with a high gear.

I think some engines are knock prone, but are you sure about the fuel octane? Have you changed stations / brands? For reference I'm running BP 93 octane.

Edit: I'm running more retard (peak of 3 @ 6500 AND I'm on 93 octane.

 

While I don't have any practical experience with 91 octane fuel, you can see from Shiv's maps that he runs way more conservative map numbers.

A less restrictive exhaust will help to prevent reversion which in turn will help your situation. So perhaps you should revisit this with the new exhaust.

 

More logs from today, used strategy that Ludi mentions and took out timing in the knock prone areas.

first graph the 2 logs where done on the same stretch of road in the AM today.
On second graph is comparing second AM run with one late afternoon today after driving in rush hour traffic, using same map.


I took out more timing but the engine felt flat in the upper range

I'm at -1@5k to -2.7@7k rpm

 

Especially you current maps for fuel, timing, smart timing, smart fuel. You should not be getting that kind of knock activity at what looks to be about a 10.0-10.3 AFR unless the timing and boost are out of line. That being said, there do seem to be some rather knock prone Evo's out there. My theory has to do with the head casting, I have seen a picture that Propellerhead put on EvoM a while back that had HORRIBLE casting flash in the number four cylinder, but when I took my own head off it was very good all of the way across the board. The head that propellerhead aquired came off of nastea's car after he wrecked it (karma is a b!tch) and he had always had problems with knock activity. Just a theory, but your maps and boost will tell a lot...

 

Yeah, thats pretty much my plan for now.

 

See my edit above about timing.

 

I just saw the data you put up while I was typing, it looks like you have a lot of timing, you may try to pull some. I say this after looking at some of the base maps that Vishnu has. I am by no means a "tuner" though, so this is purely my take on things..

 

The new 91 oct v360 maps i will be loading have even more than that, up to -4.1@7000. so my timing may just be that far off. we'll see...