Hey all, I posted this on norcal evo, my local board.

It was a very interesting experience and finding that I want to share with everyone in the EcuFlash forum.

I will copy-paste the the post here without edits, so the screenames won't make much sense but it doesn't really matter.

Here is the norcal evo thread if you would like to view it there as well:

http://www.norcalevo.net/forum/index...html#msg195933

------------------------------------snip-----------------------------------------

Alrighty... Sit down this is going to be a doozy!

mygsx came over last night for a tune. This car has a bunch of bolt-ons:

1. 272/272 cams
2. 720cc injectors
3. FMIC
4. IC pipes
5. Ported 02 housing
6. Fuel Pump
7. TBE exhaust, no cat (lots of resonators, quietest non-catted evo I have heard yet)
8. Cam Gears set at -2/-3 (if I remember correctly)
9. EcuTek Flash
10. 10.5 hotside
11. Works Throttle body
12. AEM open air filter intake
13. Maxford header
14. Sun Hyper voltage system
15. Match ported intake manifold
16. Hallman MBC

mygsx said he was happy with the current power but wanted to see what I could do with it. So I did two baseline pull logs with Evoscan and I used Datalog lab to visualize it.

Here are the datalogs from those two baseline pulls visualized. Keep in mind these two pulls felt pretty smooth and strong.





Yes, that's 20 counts of knock right off the bat, notice it shoots up and flatlines at 20 on the second graph. Even more interesting is the timing curve which stays flat at 5 degrees and then slightly starts to increase at 5500 rpm to about 12 degrees up top. More about this in a bit.

Now, here is the real shocker, here is the DLL power curves from those pulls:



Now you are probably asking yourself, how can these curves be so strong and so smooth with this much knock? It took me awhile to come up with a theory about this. More about this in a bit.

At this point I was very interested in seeing the rom this car had in it. So I open it up. This was the first time I have been truely in shock and awe by opening a rom. This Rom had 12 degrees as target from 3000 rpm load column 180 to 260 to 5500 rpm load column 200 to 260. TWELVE DEGREES This is more than twice the target ignition timing I like to set on 91 octane gas.

I open up the low octane ignition map. Carbon copy. Same exact map as high octane. I couldn't believe it. Now, back to the power curves above. After thinking about it for a bit and then seeing the ignition maps. The car was knocking 20 knock sums. Timing was being pulled 7 degrees to 5 degrees from the target 12. The octane rating was obviously low from the 20 knock counts, but being that the low octane ignition map was set the same as the high octane map, the ECU could only pull timing down to 5, it could not start interploating between the maps because the maps where the same.

So the timing stablized at 5, no big dips in timing and it actually slowly increases past 5500 rpm when the knock sums slowing started going down. Smooth flat timing curve with 5 degrees and still increasing a little up top = the car is still making power and the graphs are smooth because there is no "sawing" of the timing curve up and down. Amazing eh?

This is a perfect example showing that just looking at power curves sometimes won't be telling you the whole story about how the car is running.

I take a look at the rest of the maps. The car is scaled totally incorrectly for the 720cc injectors. The fuel tables are of course set very high to compensate for what is really no scaling of the 720cc injectors. Fuel trims totally off.

So I go about scaling the injectors correctly. I load a base flash I have been working on for cammed cars with larger injectors and flash the car. We go drive around a little to make sure the injectors are scaled well and the drivablity is way up.

Then we do a couple logs at WOT:



Knock sums are almost non-existent. timing curve is a little odd with a little dip. More about this after the real shocker:



Hey look at that, almost match the power curves from the flash that was in the car before and with almost non-existent knock (highest sum was 3 and that was even before full boost). You mean you can have a powerful car and be safe too? Oh the horror!

We do a couple more pulls and the car starts having about 3-5 knock sums right at peak boost. I check the boost. The car is running 24 psi. On 91 octane. I ask if he upped the boost since the original tune. Nope. That is what it was set at.

I drop the boost to 22psi peak and 18 or so up top. Do a couple more tweaks on the car and log and reflash it half a dozen more times.

The final outcome is this:





Couple sporadic 2 knock sums but overall a clean tune with good timing and advance. The night was getting late so I told the owner the tune is safe but I would like to take some more time next week to completely clean the tune up. He agrees. All is good in the world.

Final tune power curves vs original tune power curves:



Matched the torque and actually get it to hold a little longer up top, 30whp increase up top, and a safe almost completely knock sum free tune

It was an interesting night.

 

Older motors like the 6 bolt in DSM’s would make more power by shoving timing down their throat. They detonated before MBT (mean brake torque). What that basically means is that even on race gas you could see gains in torque by increasing timing and making the motor knock.

Newer motors like the evo start to knock at about the time you are reaching MBT. On race gas you can actually add enough timing that you’ll loose torque and the motor still won’t knock.

Even a 6 bolt on methanol will not knock before reaching MBT. A motor like this will go straight into preignition. Basically you can add timing to a meth motor thinking your adding power (but your not) and you go past MBT and enter preignition. If its just too much timing you may just bend rods. If you enter true hot spot related preignition you won’t knock, you won’t hear anything, then you’ll see your rods on the ground behind you.

This is what I like about the evo 9 head. You can run good a/f ratios on pump gas, decent boost, and the motor just loves to make nice smooth power. This is a catback and MBC evo 9.

 

Thanks for the info Dan. I usually don't push leaded race gas tunes so that is good information.

 

Thanks for the nice write up Bryan.

What I've noticed with large knock sums is that they decay slowly even when the triggering event has abated. They seem to be on a decay timer like the octane number (although obviously a shorter period). This is probably what keeps this particular example smooth as it takes a long time to decay from a knock count of 20.

With the resultant timing of 5 degrees (12-7 correction from the 20 knock count) I think it could have pulled a further 5 degrees off because the knock sum can go to 36. However, with bad gas and a big overboost, maybe it would have knocked even with zero timing, so the safety margins are a bit tight.

 

Their is absoutley a knock decay timer in thea code. I've actually messed with it in my 1g and even turned it off. I'm smooth like that. Note that I turned it back on, its actually nice. On built motors you can reduce the decay proportional to how much you hate your motor.

You can also knock at 0 degrees timing. You can knock at less than 0 degrees timing. Actually by definition detonation (knock) is uncontrolled burn rate (IE rapid flame proprigation) after 0 degrees crank. While timing can induce what we call knock just by running a low or sub 0 value of timing does not make it so that uncontrolled burn rate will not happen after top dead center. Just crank the boost high enough and she'll knock especially at a low rpm.

Also note that low rpm knock pounds rod bearings a little harder than high rpm knock. THe piston isn't running away from the uncontrolled burn rate as fast as it is at higher rpm. Also low rpm knock makes the timing curve suffer for the rest of the rpm band as the decay timer is f u c k i n g with you.

 

Well put

 

First of all, awesome post, Bryan!

I highlighted these two parts of your post because the first quote seems to correlate with what I thought all along. The Evo ECU pulls timing the same exact way the DSM ECU does, roughly 1* per 3 counts. Thanks for the data.

I quoted the second part because I agree 100% with it and I argue this point over and over again with some of the 'old school' tuners that just use dyno's #'s and curves to tune their car. They are so certain that they can tell if a car is knocking just from the torque curve....well this is proof that is not true.

Very good post, Bryan. I love it when real information is posted and not mythical crap.


Thanks,
Eric

 

Wow that is very nice man. I am glad you helped him out and dam your even doing it for free.

 

I agree, I have seen quite a few EVO's come through with 22+ degrees on 91octane...(well it was trying to hit it...by the time the retard hit, it got about 12). The car basically has so much timing it rides the knock sensor all the time. Pretty crazy.

 

nice to see you help other members out like that...

 

Makes sense: car makes less calculated HP with too much timing. using 91oct

That car would have screamed running that ignition on alc inj at 25 psi. But I get your main point - person got car tuned by someone who managed to get smooth torque under knock using same LO/HI fuel maps.

Do you have the before/after excel scans?

 

You're looking at them visualized above. I loaded them into DataLogLab.

 

If your running a wussy shot of methanol you could probably force that much timing down the motor and it would run acceptable.

If your running a real methanol shot like you should if your using that stuff then you can break bottom end rotating assembly parts easilly given that you are moving enough load to warrant not using that amount of timing. Burn rate increases with cylinder pressure and less timing is then needed at that given rpm. Increasing burn rate is an easy way to reduce negative mechanical work and maximize BMET.