throughfaith323
(39, 40, 41)
09/19/11 07:50 PM
Knock problem with ECMlink data! Please HELP!

Hey GVR4.org,

I have a knock problem and was hoping someone here could help.

MODS:
-EVO 3 16G (18psi)
-EVO intercooler with 2.5" piping
-BC 272 cams
-EVO/3g maf
-FIC 950cc injectors
-Base pressure: 44psi
-Global fuel: -56.6%
-Deadtime: 360usecs
-Walbro 255
-ECMlink

After installing the E3-16g, along with a bunch of other mods, I wanted to try and tune the car with a few pointers from my friend(first time ever trying to tune)....

Here is what I did:
1. Set the fuel adjustments for the 950cc injectors.
2. Brought the timing table to zero.
3. A few other things.
4. Went for a drive and logged a 3rd gear pull.

The AFRs seemed to be pretty good at WOT so my method from here was to keep on advancing timing until I noticed substantial knock popping up. I could not advance timing even 1 degree without making the knock increase. There is something wrong here correct? Timing advancement should increase with rpm's and I seem to have the opposite effect.

I thought my knock sensor was shot when I looked at it and the goo was completely missing. I switched it out with another knock sensor I had lying around and this did not change a thing. I thought for sure I had a bad knock sensor.



Below are screen-shots of the 3rd gear pull "LIVE datalog", timing table, and fuel adjustments respectively. Maybe something sticks out from this data that could point me in a certain direction?






Where should a go from here? Any suggestions would be very much appreciated!!

THANK YOU!!
Ed


spooling92vr4
(Yeah Toast! * Now with JELLY!!!)
09/19/11 07:56 PM
Re: Knock problem with ECMlink data! Please HELP!

can you post a log so we can see your settings?

vtecds1
(linkfail)
09/19/11 08:00 PM
Re: Knock problem with ECMlink data! Please HELP!

Not sure on a solution but just by looking at your log, the coolant temp seems to be pretty high.

throughfaith323
(39, 40, 41)
09/19/11 08:19 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting spooling92vr4:

can you post a log so we can see your settings?




how do I post a file? I can give you screen shots of anything that might help.

thanks!


throughfaith323
(39, 40, 41)
09/19/11 08:20 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting vtecds1:

Not sure on a solution but just by looking at your log, the coolant temp seems to be pretty high.




I thought our cars run about that hot? What is normal for them?


beaner
(Blew1 Guy Today )
09/19/11 08:52 PM
Re: Knock problem with ECMlink data! Please HELP!

Between 189 and 206 is full timing. From 206 to 224 the ecu pulls one degree. Over 224, two degrees.

throughfaith323
(39, 40, 41)
09/19/11 09:31 PM
Re: Knock problem with ECMlink data! Please HELP!

^^
Quote:

Between 189 and 206 is full timing. From 206 to 224 the ecu pulls one degree. Over 224, two degrees.




Hmm... that's definitely a problem for me then because that entire 2->4th gear it was 206degF then 210degF. THANKS for the info Beaner!!

That shouldn't be enough degF to cause detonation, then causing knock, at high RPM though would it?

EDIT:
Possibly enough to mess with the knock sensor?


beaner
(Blew1 Guy Today )
09/19/11 10:36 PM
Re: Knock problem with ECMlink data! Please HELP!

Just figure out a way to post the log.

WaRrIoRs16
(Member +)
09/20/11 12:05 AM
Re: Knock problem with ECMlink data! Please HELP!

Found this a while ago when I was trying to post a log. click

WaRrIoRs16
(Member +)
09/20/11 12:23 AM
Re: Knock problem with ECMlink data! Please HELP!

It looks like your a little lean for running gasoline. In your first pic your wideband shows you at 11.8 AFR, I believe you should be targeting 11.0. Also your AFRatioEst looks dreadful. You probably have some airflow reading problems going on because it is showing that your flowing much more air than you actually are. Is your Base MAF type set to evo8/3G? If not, that could cause some problems. Lastly, boost leaks perhaps?

throughfaith323
(39, 40, 41)
09/20/11 08:07 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting WaRrIoRs16:

Found this a while ago when I was trying to post a log. click




dang! thanks for the info.


throughfaith323
(39, 40, 41)
09/20/11 08:18 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting WaRrIoRs16:

It looks like your a little lean for running gasoline. In your first pic your wideband shows you at 11.8 AFR, I believe you should be targeting 11.0. Also your AFRatioEst looks dreadful. You probably have some airflow reading problems going on because it is showing that your flowing much more air than you actually are. Is your Base MAF type set to evo8/3G? If not, that could cause some problems. Lastly, boost leaks perhaps?




-I thought the optimal stoichiometric mixture for gasoline and turbo was 11.6-11.8 at WOT?

-I was told that in ECMlink the "AFRatioEst" would not be near accurate. Is this not true?

-</font><blockquote><font class="small">Quote:</font><hr />
You probably have some airflow reading problems going on because it is showing that your flowing much more air than you actually are.



What should the airflow amount roughly be at Xrpm with a E3big16g?

-I do have the base MAF set at evo8/3G

-I looked for boost leaks some time ago. its been a while so I will definitely check them when i can get the tool from a friend or make one myself.


gramkrakr89
(Now with Jelly next week is roasted marshmallows and hershey bars)
09/20/11 08:33 PM
Re: Knock problem with ECMlink data! Please HELP!

I think the stock ecu tries to hit 10.something AFR at WOT... i think... right?

(BTW, the estimated AFR in ecmlink is way off base because of the calculation settings in the logger... it's not actually that lean, the pc just calculates that value based on data that you enter in the "fuel" settings somewhere...)

EDIT: actually, it's in display preferences for AFRatioEst on the log graph itself, for older firmware versions


beaner
(Blew1 Guy Today )
09/20/11 08:38 PM
Re: Knock problem with ECMlink data! Please HELP!

11:1 is pretty common I think. Afratioest will be off if your global fuel, deadtime, and/or maf settings are wrong. If your new knock sensor is tightened too much into the block, it can give a false reading. I tightened mine with just the weight of the ratchet and a dot of thread locker for peace of mind.

throughfaith323
(39, 40, 41)
09/20/11 08:41 PM
Re: Knock problem with ECMlink data! Please HELP!

ok cool. thanks for confirming the AFRest thing. lets just hope my wideband is reading right and nothing is screwing that up.

throughfaith323
(39, 40, 41)
09/20/11 08:48 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting Beaner:

11:1 is pretty common I think. Afratioest will be off if your global fuel, deadtime, and/or maf settings are wrong. If your new knock sensor is tightened too much into the block, it can give a false reading. I tightened mine with just the weight of the ratchet and a dot of thread locker for peace of mind.




- Hmmmm...can anybody confirm that the deadtime I chose for the 950s is correct please?

-I read about the knock sensor issue somewhere. I thought mine was bad and I replaced it with one from another motor I have. The sensor seemed to be pretty new. I thought for sure it was phantom knock and this new sensor would fix it but i was wrong. I installed the "like new" sensor and applied the right torque (It was an odd number if i remember correctly. something really low. I will double check this though and make sure its good. maybe i'll try your method .


throughfaith323
(39, 40, 41)
09/20/11 09:22 PM
Re: Knock problem with ECMlink data! Please HELP!

So today I checked my spark plugs to see if any of them looked out of the ordinary. They all looked the same besides cylinder 4 which was wet. I think its just wasted gas from leaving it in the head while checking the others for spark. All sparked fine from what I could see/hear during the day. Does anybody here have any experience/know how to "read spark plugs?" These pictures were taken after cruising for 4 minutes on the way home from some pulls:



Now as far as the pulls. The outside air temps were probably about 10degF cooler than yesterdays logs. This allowed my coolant temps to stay much lower. After doing a pull with the same settings as yesterday(shown in pictures from my first post), the knock was a bit lower (shown below). I then decided to see what would happen if I added a degree of timing across the board, and then another degree of timing across the board. The knock got worse each time(as expected). During all this the coolant temp remand the same as the first pull. I am still going to address my abnormal heat conditions, and I know it doesn't help my problem, but at least we can rule out the coolant temps causing my knock. I thought that would be kind of pushing it, but I did not want to rule it out.


There is one thing I've noticed in mostly all of my logs from the very first time I started dealing with this issue. There is a point where my knock hits a maximum and then starts to go down. This should happen as the car is pulling timing to prevent the knock, correct?. However, for the last 700 or so rpms of my pulls, the timing is increasing and the knock continues to go down. I'm not sure if this helps get down to the problem, but figured I would point it out.

Thanks for the support guys!
Ed


beaner
(Blew1 Guy Today )
09/20/11 10:31 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting throughfaith323:

can anybody confirm that the deadtime I chose for the 950s is correct please?



Not without a log. That deadtime seems common for those injectors so it's probably close. The ecu uses voltage and many other values to determine deadtime. Therefore, each car is different.

I noticed you're at 44psi base fuel. It should be set at 37psi, as that is what the ecu is expecting.


vtecds1
(linkfail)
09/21/11 09:07 AM
Re: Knock problem with ECMlink data! Please HELP!

^^^I agree with the dead time being different for every car. The ones you see posted on the Link board are just starting points. You have to fiddle around with the numbers. I thought I had mine all set, but it was over 100 off.

Also, my tuner was getting small counts of knock at over 6500rpm. He messed around with the timing and it was still the same. Check the plugs and there were all good. Phantom knock seems to be the culprit on my car.


**DONOTDELETE**
(Unregistered)
09/21/11 12:20 PM
Re: Knock problem with ECMlink data! Please HELP!

For what it's worth, when i changed my thermostat to a 160 from the stocker and added 2 12" fans in front of my radiator i was able to pick up about 7-10* of timing just by lowering my coolant temps.

Dark_Horse
(Member +)
09/21/11 12:44 PM
Re: Knock problem with ECMlink data! Please HELP!

IMHO, you should drop your fuel pressure down to 37psi and drop your injector global to -53%. Then tune the global till your gram/rev is hovering around 25-26 and your combined FT at idle is +/- 2% or so. That will give you a better starting point.

Also, for pump gas, 11.8:1 is pretty lean.


throughfaith323
(39, 40, 41)
09/23/11 07:27 PM
Re: Knock problem with ECMlink data! Please HELP!

Thanks so much for the input guys! I'm learning a lot through this process. This knock will be a huge benefit in the long run.

So here is what I've done. I lowered the fuel pressure to 37psi using the ECMlink diagnostic mode. I then set my global to -53 and deadtime to 330. Started the car up and fuel pressure was at roughly 39psi. Left it at that for now (getting eatin alive by mosquitos and was running out of time ) until I really learn how to tune....when I do, I will then start from scratch. I then played with global and deadtime a little, since FP was not exactly 37, and got the idle A/F's and part throttle pretty decent Averaging 14.9ish over a period of time.

From here I tweaked MAFcomp a little to bring down my A/F ratio to low 11's. I know I am missing some steps in my tuning process, but I am in the learning process and have huge overload from all the reading I've been doing. I have questions about tuning with ECMlink and will get to that in my next post. After adjusting MAFcomp to enrichin the A/F ratio, this is what I ended up with:



From this log you can see that the knock dropped to almost zero counts across the board!! The 3 logs right before this one (as I was adjusting the MAFcomp) showed similar signs of knock to the knock I was getting before these fuel adjustment (richer), but was improving after each adjustment towards low 11 A/F ratios.

ALSO I MUST NOTE: All of these logs were done with 92octane gas from WAWA(cheap gas). I will be filling up with Sunoco 93 octane fuel only from this point on and now that I'm actually learning to tune and do things right.

Tuning/coolant questions to come......


throughfaith323
(39, 40, 41)
09/23/11 07:59 PM
Re: Knock problem with ECMlink data! Please HELP!

Now for the engine cooling questions.....

Engine Cooling (thermostat):

I've read a TON of threads on people going back and forth about running a colder thermostat. My question is, what do you guys recommend? Keep in mind I do have ECMlink, so I WILL be able to force my car into closed loop IF coolant temperatures are below 180degF (stock thermostat temp).

My driving conditions: I do not daily drive my car nor do I often get stuck in traffic when I am driving my car. I live in New Jersey and do not drive in the coldest times of the year (when the roads have been salted). When I am driving my car I usually "get on it" a lot (not every chance I get, but enough).

Some of the pros and cons I've either thought about or read about when using a lower temp thermostat (correct me on anything I may be wrong about):

Some PROS could be:
-Safety margin with headgasket concerns
-Lower chance for detonation/knock to occur
-Lower under-hood temperatures
-Lower intake temp in some cases due to lower under-hood temperatures (haven't gotten to optimal intake pipe routing yet )

Some CONS could be:
-Mainly lower engine efficiency
-Decreasing fuel atomization (20 degrees shouldn't affect it too much right?)
-Winter coolant temps being too low when cruising (could be adjusted with some airflow blockage to the radiator like trucks/buses do)
-Others, but I am drawing a blank right now.


This all being said, what temperature thermostat do you guys suggest I use? If you are very knowledgeable and have experience with this question, please chime in! Please point out any CONS that I may have left out and would be of some concern.


beaner
(Blew1 Guy Today )
09/23/11 09:28 PM
Re: Knock problem with ECMlink data! Please HELP!

Just keep it between 190 and 205. I think it comes down to how good the radiator is. I have a civic half radiator in mine and experimented with a few thermostats to keep it where I wanted. It hovers in the low 200s but it for sure doesn't have a 200+ thermostat in it.

MuffinMan7580
(Member)
09/24/11 09:35 PM
Re: Knock problem with ECMlink data! Please HELP!

Quoting throughfaith323:

I lowered the fuel pressure to 37psi using the ECMlink diagnostic mode. I then set my global to -53 and deadtime to 330. Started the car up and fuel pressure was at roughly 39psi.





Fuel pressure should be set at 0 vacuum/Boost (@ atmospheric pressure). Set it with engine off, only running the fuel pump, or you can start your car, disconnect the vacuum line on your AFPR, set fuel pressure, then re-attach.

Coldest thermostat I would run would be 180*. If you run anything much colder than that, oil temperature could possibly not come up to operating temperature. If you're worried about underhood temps, you can raise the rear of the hood a bit (washers underneath the hinge), and if there's weatherstripping there, you can take that out (I haven't looked on mine, although I know 1g DSM's have it).


throughfaith323
(39, 40, 41)
09/24/11 10:04 PM
Re: Knock problem with ECMlink data! Please HELP!

Quote:

Fuel pressure should be set at 0 vacuum/Boost (@ atmospheric pressure). Set it with engine off, only running the fuel pump.




That is what I did. I turned the pump on using ECMlink and set the pressure to 37. I then set global/deadtime and then turned the car on. I checked the base fuel pressure by taking off the vacuum line to the FPR. the FP gauge was then reading 39 . I will change it sometime soon before I get into too much tuning. I have a lot to learn first.

Quote:

Coldest thermostat I would run would be 180*. If you run anything much colder than that, oil temperature could possibly not come up to operating temperature.




This is why I am unsure what to do. I have read about many people suggesting running a thermostat 10* (170*) or even 20* (160*) cooler. The other suggest what you have suggested here....stick to stock and run the 180*. I do not know what to believe because both stock t-stat and cooler t-stat have have their advantages and disadvantages.


MuffinMan7580
(Member)
09/24/11 10:11 PM
Re: Knock problem with ECMlink data! Please HELP!

Stock t-stat is 190*(ish). The 180* would give you a small advantage, but, IMO, it would be negligible.

Dark_Horse
(Member +)
09/25/11 02:34 AM
Re: Knock problem with ECMlink data! Please HELP!

There's no reason to run a colder thermostat unless you're having problems with your car running hot. And even then, there are other things you should do before jumping into a colder thermostat. Did you do a radiator flush? Is your rad cap holding pressure? What coolant mix are you running? Did you bleed the system properly? Does your rad core have a bunch of fins that are all bent up? Is your AC condenser still in the vehicle?

Plenty of stuff that can be addressed before running a really cold thermostat.


thecman02
(too lazy to look)
09/25/11 12:58 PM
Re: Knock problem with ECMlink data! Please HELP!

I'd recommend sticking with stock thermostat. When your car is running too hot and the thermostat is for sure opening the problem is a lack of heat exchange between the radiator. Make sure the radiator is clear of debris, make sure the coolant is flushed out well. Also make sure it is a 50/50 mix and that it isn't straight coolant. Check to make sure the radiator cap is holding correct pressure.

As far as tuning you should really be at 11:1. You need to post logs too, that is 1000x more useful then a screen shot. We can go into the log and see all the settings you have your car set to.



Contact Us | Privacy statement GalantVR-4.org

Generated in 0.007 seconds in which 0.004 seconds were spent on a total of 4 queries. Turbo powered.


Hertz's Galant VR-4 Page