EVO III GS-T Finally "done" and dyno'd... Nobody is knocking ! ?

[Jaraxle]

This is going to be a long story short....
Check the profile on the car, if you have questions on what this GS-T has...
Here are the pieces of data... I need someone who knows more than me, to figure this out.

1.) The knock wire to the ecu was CUT 3 inches from the ecu harness, and the re-twisted and soldered over a year ago. Don't bother to school me on this. I know. Move on. The AFC-II can't read 2g knock.

2.) Shortly after the stock #2 rod bearing gave out on the dyno. (with only a FMIC and 3" turbo-back). It is hard to tell if the noises coming from the engine before it gave was severe high rpm knock, or the rod bearing just giving out due to other reasons. My 7-bolt split-thrust (99) block sent to slowboy for steroid injections.

3.) 8 months later, my SBR built forged motor spins #2 rod bearing AGAIN. Stock boost, stock ecu, stock injectors, being babied during break-in. 1,300 miles on new engine.

4.) Present day. New SBR bottom end with new crank and #2 rod/bearing. All mods in profile. 19-20psi and 93 octane. DSMCIPS ecu trying for 16deg of timing. Seeing NO NOCK ever on the stock boost gauge. I don't see knock at 20psi in 100deg heatwave heat. Dyno chart from SBR shows shop temp at 107 deg F, car made 240hp and 270 tq to the wheels. Richest was 10.9:1 or so at 7k. Leaner for all rpm's lower than 7k.

First question: is this possible (to have no knock at ALL with 119-20psi in high heat) ?
Second question: if not, then is it the twisted/soldered knock wire degrading the signal to the ecu or something ? Could the engine be knocking, but the ecu doesn't know it ?

I've been tempted to run the tank low, put a few gallons of 87. Prove I can get knock, and then tank up on 100 octane. This seems like a dangerous game though. I need better ideas from smarter people.



-Jaraxle

 

[nanokpsi]

Yes the car can knock at those boost levels. Even at 20 psi I have to pull a degree or two in DSMlink to keep it form knocking. At higher boost I have pull 4 deg on pump gas. After using the link for a while I can't believe cars with AFC's live very long.

 

[Jaraxle]

Thanks for the response, but you didn't say what the timing was being pulled to. If your running a similar setup @ 20 psi and the timing is being pulled to 16deg, then that is another data-point that perhaps I really am safe, and not knocking at all as indicated by my boost gauge. If you read the post, I'm running dsmchips for 660 greddy injectors, the afc if only to fine tune it. So was your timing being pulled a deg or two, to 16deg or more ?

-Jaraxle

 

[andymoraitis]

I've been running a DSM Chips arrangement on my GSX for the last 6 months and Jeff and I have worked at using different timing and A/F maps. I used to run an 11:1 A/F at 16 degrees timing and it never knocked on pump and 22psi in 95 degree weather so I can cofirm your results. Since my setup is somewhat similar to yours, I can tell you for sure that you'd need a lot of boost to create knock at 16 degrees of timing. In the context of an adequate A/F ratio, timing will always make the largest difference in performance.

Let me put this as clearly as possible:

9.5:1 Chip with 16 degrees = 14.0 at 100.40 - 20psi, pump, zero knock
11:1 with 23 degrees = 12.82 at 106.10 - 22psi, pump/octane booster, 5 counts knock
11.7:1 with 23 degrees advance to be tested soon.

Have Jeff make you the same chip with a 21 degree map and see how much better it is. Remember you're looking for about 3 counts of knock maximum since that's the edge of the tuning envelope in terms of not pulling timing and being lean enough to make good power.

I felt a huge difference with a timing jump and I'm sure you will too. At the moment, you're not getting strong timing since you're not pulling airflow with the AFC as a result of having injector compensation. Have the timing map changed and you'll be able to make some power. Add race gas, up the boost and pull some more fuel with the AFC and you should be right in the tuning sweet spot. You'll also find that, within reason, increased timing will lower your EGT's and feel much better to the motor.

Hope that helps!

Andy

 

[nanokpsi]

Thanks for the response, but you didn't say what the timing was being pulled to. If your running a similar setup @ 20 psi and the timing is being pulled to 16deg, then that is another data-point that perhaps I really am safe, and not knocking at all as indicated by my boost gauge. If you read the post, I'm running dsmchips for 660 greddy injectors, the afc if only to fine tune it. So was your timing being pulled a deg or two, to 16deg or more ?

-Jaraxle

Once you go over 2.1 g/ rev the max amount of timing the stock ecu will give is 16deg. Timing was being pulled below that.

As far as 20+ deg maps with dsm chips, in no way does that jive with high boost and pump gas if you have any type of load (airflow) going through the engine.

 

[andymoraitis]

As far as 20+ deg maps with dsm chips, in no way does that jive with high boost and pump gas if you have any type of load (airflow) going through the engine.

I beg to differ based on firsthand experience. If you read my post carefully, it says "pump gas/octane booster". The octane booster is Torco Unleaded Concentrate which runs at an aggregate of 104-107 on straight 93 depending on concentration. If I ran it on pump gas only, I'd see 12-13 counts of knock at the same boost levels, A/F and timing, which meant that timing was being pulled back to 21 based on knock count.

Granted, the timing curve in my chip is made more for race gas, but I'd bet I could get away with it on pump and minimal knock if I ran an FMIC or water injection.

Andy

 

[nanokpsi]

I beg to differ based on firsthand experience. If you read my post carefully, it says "pump gas/octane booster". The octane booster is Torco Unleaded Concentrate which runs at an aggregate of 104-107 on straight 93 depending on concentration. If I ran it on pump gas only, I'd see 12-13 counts of knock at the same boost levels, A/F and timing, which meant that timing was being pulled back to 21 based on knock count.

Granted, the timing curve in my chip is made more for race gas, but I'd bet I could get away with it on pump and minimal knock if I ran an FMIC or water injection.

Andy

Well, I am speaking form personal experience as well, all of which being datalogged from DSMlink. The fact that your car is seing vert little load may help you, but in my experience the cars simply don't like timing and high boost on pump gas. I can leave the timing curve alone and it will probably have a few degrees pulled from knock at 20 psi. On a max timing scale of 16psi that would probably leve me with 12-13 deg timing at peak. This is regardless of AF ratio as well. In my experience dumping fuel in the combustion chaper does very little to help surpress this type of knock on pump gas. It gets worse trying to run 22-23 psi on pump. I have dyno tuned several cars with DSMlink and street tuned each of these cars countless times. Each car is different though and YMMV.

 

[andymoraitis]

Well, I am speaking form personal experience as well, all of which being datalogged from DSMlink. The fact that your car is seing vert little load may help you, but in my experience the cars simply don't like timing and high boost on pump gas. I can leave the timing curve alone and it will probably have a few degrees pulled from knock at 20 psi. On a max timing scale of 16psi that would probably leve me with 12-13 deg timing at peak. This is regardless of AF ratio as well. In my experience dumping fuel in the combustion chaper does very little to help surpress this type of knock on pump gas. It gets worse trying to run 22-23 psi on pump. I have dyno tuned several cars with DSMlink and street tuned each of these cars countless times. Each car is different though and YMMV.

I would have to agree that it depends on the state of tune and slightly by car. I've noticed that A/F affected my performance much less than timing did, but I too feel that tuning for big timing and big boost on pump gas is a royal pain. I've never been a believer in drowing the car in fuel either so I strike a middle ground by slowing or controlling the burn rate with higher octane levels. In my case, running the equivalent of 104-107 unleaded has helped, but at a great expense. Another limiting factor is the simple truth that the B16G is a horrible pump gas turbo since it generates so much heat at high boost.

Cheers!

Andy

 

[CanadianTSi]

I also agree that high boost, high timing, and pump gas don't mix.

Running 23psi on pump I get around 19 deg's but my friend running 25psi gets about 16, even with the lower timing the extra couple pounds give him 25AWHP on me...

 

[Jaraxle]

I've been running a DSM Chips arrangement on my GSX for the last 6 months and Jeff and I have worked at using different timing and A/F maps. I used to run an 11:1 A/F at 16 degrees timing and it never knocked on pump and 22psi in 95 degree weather so I can cofirm your results. Since my setup is somewhat similar to yours, I can tell you for sure that you'd need a lot of boost to create knock at 16 degrees of timing. In the context of an adequate A/F ratio, timing will always make the largest difference in performance.

Let me put this as clearly as possible:

9.5:1 Chip with 16 degrees = 14.0 at 100.40 - 20psi, pump, zero knock
11:1 with 23 degrees = 12.82 at 106.10 - 22psi, pump/octane booster, 5 counts knock
11.7:1 with 23 degrees advance to be tested soon.

Have Jeff make you the same chip with a 21 degree map and see how much better it is. Remember you're looking for about 3 counts of knock maximum since that's the edge of the tuning envelope in terms of not pulling timing and being lean enough to make good power.

I felt a huge difference with a timing jump and I'm sure you will too. At the moment, you're not getting strong timing since you're not pulling airflow with the AFC as a result of having injector compensation. Have the timing map changed and you'll be able to make some power. Add race gas, up the boost and pull some more fuel with the AFC and you should be right in the tuning sweet spot. You'll also find that, within reason, increased timing will lower your EGT's and feel much better to the motor.

Hope that helps!

Andy

Andy - I really appreciate the response. I was starting to loose faith in these forums all together. Too many people bench racing and asking the same lame questions, so I haven't been on in ages. Your data makes me feel better about my knock sensor status. My factory boost gauge (knock) NEVER moves, but it sounds like that is expected given my conservative fuel and timing tune. What is the best way to get a hold of Jeff for another eeprom ? I e-mailed him once about this, but didn't get a response. I'll probably have him burn a 19 or 20 deg advance map, since I NEVER run race gas... Only 93 octane. Do I need to re-tune the afc high-throttle fuel maps with a wideband after putting in the more advanced timing map ? Also, I'm running rich on low-throttle, there is soot on my bumper/tailpipe. I was planning on leaning out the low-throttle settings using the short-term fuel by logging and driving at a particular rpm. Any tips there would be great. Besides a litle low-throttle rich, the car is running awesome with dsmchips. 270 tq in 107deg heat is no joke. Thank god I have a limited slip , wish I had awd.

Thanks,
Jaraxle

 

[andymoraitis]

Jaraxle,

It's no trouble and I'm glad we could all help out in some way. I'm kind of a strange bird in that all of my A/F maps and timing are controlled by the chip. I don't have an AFC and will likely never add one. If you have plans to always run pump gas, then you won't need one either, but you can still hang on to it for fine tuning. If you have Jeff run the A/F maps for you, like I recommend below, I'd zero out the AFC and see how she reacts. On a 10.5:1 map, I had soot everywhere as well. Going leaner on the maps certainly helped with this and I don't see why it wouldn't help you as well.

At the moment, the temps are warm here so I mix a 60/40 ratio of pump and 110. At 22 psi and 23 degrees of advance, I only see 3 counts of knock at the top of 3rd, which is right where I want to be. I just tried this yesterday and it works much better than the octane booster I was using since the leaded gas creates a slower burning flame front.

Just to get myself up to date, I looked for injectors in your profile, but didn't see any. If they're larger than stock, I would have Jeff compensate for those, change the A/F maps to 11.5:1 and take your timing to 19 degrees. In the presence of adequate advance, boost will make more horsepower than timing. In my case, where I mix pump and 110, I can have strong boost and timing and get the best of both worlds. If owned your GST and was sending Jeff specs, this is what I'd e-mail to him:

1999 GST M/T
2G MAF
950 RPM Idle
No Fuel Cut
Octane Reset
11.5:1 A/F
19 Degrees Advance
Knock Sum Boost Gauge
Injector Compensation (based on size and brand, but you need 650's on an EVO III)

With your current parts, assuming those injectors are in there, there's no reason you shouldn't be able to handle 20-22 psi on pump, but tread carefully and log it just to be sure. One thing I would definitely be cautious of is that ECU wire. I would really have that checked out by a reputable tuning shop to make sure that the splice will allow the ECU to see knock. I don't think I need to tell you the ramifications if it isn't registerting.

As far as getting in touch with Jeff, he was on vacation for a few days. Drop him a line at [email protected] and he'll be sure to get you going fairly quickly with an upgrade.

Let me know if that makes sense,

Andy

 

[Maglin]

So you spun another #2 rod bearing on your build motor? If would assume since it's the same hole that you might have one bad injector that is going lean maybe. That is if you put the injectors back in the same cylinders. You can just tap on the knock sensor with a metal object and have someone monitor your knock/boost gauge to see if the knock sensor is working. If it never moves you have either a bad connecton or a bad sensor.

 

[Jaraxle]

Maglin,
Both rod bearings were #2 with the same stock injectors. You could be correct, that the #2 cyl was lean. Now I am running 660 Greddy injectors (--Andy). Hopefully that was the problem, but I'd really like to verify the knock system. As for your test.... I thought the ECU only used the knock sensor under certain conditions (ie: over 3k rpm, high throttle, loaded for example). Can anyone verify that I can test the knock sensor in neutral with a hammer as described ? Does anyone know the actual conditions that are required for the knock system to be active ? Also, I believe you just tap the block not the actual sensor which is crammed below the intake manifold. It should pick up any sharp harmonic traveking through the block, if it is working AND listening.


Side note off topic - My 16g EVO III falls to 15-16psi by redline. This isn't normal is it ?

-Jaraxle

 

[andymoraitis]

Jaraxle,

I'm glad to hear that you have a set of 660's installed. These will typically flow more than their rating and Jeff can compensate perfectly for these with deadtime and global adjustments. I can't answer with respect to testing the knock sensor in this manner, but another DSM'er or a reputable tuner shop may be able to do so.

With respect to the boost falloff, that's not something I'd consider normal. You may want to hunt around for boost leaks to ensure that the system is tight. Typically, the EVO III has a tendency to creep and not loose boost. I even saw this on my Big 16G when I tried running boost in the 15-16psi range. Setting the MBC to the point of the creep brought it under control, but I would expect that this isn't the issue. I'd make sure the wastegate actuator isn't blowing open and also check for leaks as mentioned before.

Other than that, the parameters I gave you a few posts up should give you solid performance on pump gas and plenty of guts for the street while still keeping things in a safe range.

Let us know what you find,

Andy

 

[red91gst]

Maglin,
Both rod bearings were #2 with the same stock injectors. You could be correct, that the #2 cyl was lean. Now I am running 660 Greddy injectors (--Andy). Hopefully that was the problem, but I'd really like to verify the knock system. As for your test.... I thought the ECU only used the knock sensor under certain conditions (ie: over 3k rpm, high throttle, loaded for example). Can anyone verify that I can test the knock sensor in neutral with a hammer as described ? Does anyone know the actual conditions that are required for the knock system to be active ? Also, I believe you just tap the block not the actual sensor which is crammed below the intake manifold. It should pick up any sharp harmonic traveking through the block, if it is working AND listening.


Side note off topic - My 16g EVO III falls to 15-16psi by redline. This isn't normal is it ?

-Jaraxle

You CAN test the knock sensor with a hammer on the back of the block. Just hookup the logger, turn the key on, and give it a whack or two on the back of the block, or anywhere on the block for the matter. Its just a piezo sensor/microphone, and it should pick that up!

Nate
SBR

 

[jmakado]

On DSMLink the knock sensor is activated above about 2200rpms as a default. Not sure if it's the same running a stock ECU. You could always free rev to 2500 and then whack it just to be sure.

 

[Jaraxle]

Guys,
This is where I could really use some help. I had the car in the garage at 3k rpm. I had my bro (a mechanic) hit the block with a hammer. He wacked the metal brace by the #1 exhaust manifold runner. Pretty hard too. Over and over. The knock gauge (stock boost) stayed on the bottom. No movement. A logger showed 33 deg advance, and did not move at all. I need someone to verify that this is an accurate test. Unless I know the known expected results, I cannot be sure I have a problem. In short - no knock gauge response or timing retarding at 3k rpm in neutral with the block being wacked. 33 deg advance observed.

Help me out guys. Please.

-Jaraxle

 

[red91gst]

Guys,
This is where I could really use some help. I had the car in the garage at 3k rpm. I had my bro (a mechanic) hit the block with a hammer. He wacked the metal brace by the #1 exhaust manifold runner. Pretty hard too. Over and over. The knock gauge (stock boost) stayed on the bottom. No movement. A logger showed 33 deg advance, and did not move at all. I need someone to verify that this is an accurate test. Unless I know the known expected results, I cannot be sure I have a problem. In short - no knock gauge response or timing retarding at 3k rpm in neutral with the block being wacked. 33 deg advance observed.

Help me out guys. Please.

-Jaraxle

Try removing the knock sensor from the block and tapping it on the block, that should give you a reading. If not, try another knock sensor from another car to eliminate the sensor being a problem.

Nate
SBR

 

[Spoolin98]

knock gauge (stock boost) stayed on the bottom. No movement.
-Jaraxle

Why do you refer to the stock boost gauge as a knock gauge, i never heard of this before?

 

[red91gst]

Why do you refer to the stock boost gauge as a knock gauge, i never heard of this before?

With a keydiver chip, or dsmlink, you can set the stock boost gauge to read other things, usually it is set as a knock gauge.

Nate
SBR

 

[Jaraxle]

Nate,
I'll try that asap. I'm still curious if anyone has FIRST hand experience with knock sensor testing in this manner. If so, that would make me feel better. I'm worried that the ECU doesn't monitor knock until a certain engine load, which can't be recreated in neutral. Anyone done this first hand ?

With 93 octane, 16 deg advance, and running slightly rich... I've probably been safe (including SBR dyno time). I don't want to advance the timing as described above, until I'm sure the knock system is a-ok...

-Jaraxle

 

[andymoraitis]

Jaraxle,

Nate's a real trustworthy guy, but I figured I'd try this for myself. I jacked up my car today, started it and had my wife sit inside and monitor the boost gauge for knock while I whacked the block on the back with a large wrench. I gave it three distinct hits and each time, the knock gauge twitched.

I'm telling you for sure that you'd need hellatious boost on your setup to make it knock at 16 degrees, but I hope the test makes you feel more comfortable. If you had no luck when doing the same thing, the sensor may be faulty.

Ease your mind and go have some fun,

Andy

 

[Jaraxle]

Andy,
You are THE MAN ! Thanks a ton for the data. Was the car just idling ? And where exactly on the block did you strike? I'm gonna give this a shot... maybe even the wife can assist too. Nice car by the way.

-Jaraxle

 

[andymoraitis]

I tried this at idle and then at 2000 and 3000 RPM respectively. I whacked on the back of the block (I was underneath it with the front on jackstands) below the head as close to the sensor as possible. It's tight back there so find a spot where you can give it a few. If yours doesn't register anything then I would have to say that Nate's recommendation of trying this with another knock sensor should help. You may also want to make sure that yours isn't loose and not registering.

Let me know what you find and hang in there pal!

Andy

P.S. Thanks for the compliments. I'll have you know that when I was selecting wheels, yours was one of the cars I admired the most. Kinda makes me wish mine was silver.

 

[Jaraxle]

Yesterday I put the car on my pop's lift. He hit the block on the flat spot below the intake manifold brace. Idle and 3k produced NO gauge movement. Then I tried, by taking a steel pipe, bracing it on the intake manifold bracket bolt, and wacking the heck out of the end of it with a hammer.. At idle and 3k.. Nothing.

So before leaving work today, I disconnected the knock sensor at the engine connector.
It didn't seem to be doing anything for me anyhow, and I wanted another datapoint.

Not 1/4mile down the road, I see my knock gauge doing the hoo-ha. Finally some movement! ECU says no codes and no pending codes. (which is ironic because I thought the ecu would KNOW that there was no knock sensor attached, there is a code listed in the book for that). ? - any thoughts or data on that guys ? Anyone brave enough to recreate that one for me ?

Not sure whether to conclude a deaf knock sensor... ???
There is always that solder joint haunting me... Which was done quite well really.

Can I conclude that the open connector introduced enough noise and garbage that the ecu "heard" a problem and declared knock, thus verifying a deaf dead knock sensor ?


Today was a great day otherwise.

1.) Found out there is a dyno 10 min from my job. That beats 3 hours if my math is right.
2.) Got a call that I may get into a porsche club racing event on Oct 11th at Watkins Glen.


Thanks for all the help guys.

-Jaraxle