Running great save from some knock in 3rd. spark and knock threshold ?'s

[EfiniX]

So my car has been running terrific since moving it to SD a few months ago. I did bring the boost down a touch from 24psi to 22 to work some intermittent knock out of the system. Lots of fun!

Last week, I was pulling onto the freeway and had to push past a BMW. I got on the pedal a bit in 3rd gear and while I did pass him, the car felt a bit grumbly. So later in the day I did a WOT pull in 3rd and saw a ton of knock. Didn't have a logger on, but based on the dash, I hit 3 lights on the stock boost gauge. It's what, 7 counts per light, so that's a decent amount. I thought I might have some poopy gas, but it did the same thing on the next tank (just yesterday). 92oct from Chevron is usually fine, so I started considering other factors.

First, I checked my knock sensor threshold. It was set to 4500RPM's (which is right above my boost threshold on my big T3/04) and 70% throttle. That throttle value seems a bit high. Got me wondering if that grumble I was feeling was actually knock that my car wasn't detecting. I doubt I was more then 60% on the throttle when I felt it passing that BMW. (sorry car!). those of you with link and lumpy cams. What's your knock threshold set at?

For now, I'm going to replace the BPR7ES plugs that are installed with some BR7ES plugs. I actually had BR7's on for a while to deal with some spark blowout issues, but when I switched to speed density, I had the OmniBar sensor fail almost immediately, telling the engine it was getting a million pounds of boost and stalling the car on the freeway by going super rich. Got that sorted out, but English pulled my BR7's and installed BPR7's since that's what they had in the shop handy.

I figure I'll also do the plug wires since the ones that are installed, while NGK, are 7 years old. If that doesn't work, then I'll start searching for boost leaks.

Log coming tomorrow morning. For now, any thoughts?

(bonus question: can spark blowout cause knock?)

You guys are the best!

 

[GSTwithPSI]

It's hard to say without actually seeing a log. Where are you seeing the knock? Up high in the rev range or what?

 

[EfiniX]

Yeah, high revs and max to near-max boost. Right where I would expect little wrong things to cause problems.

 

[EfiniX]

Well I did two pulls this morning. I didn't see as much knock, but it was there. Some cursory observations...

I look lean at full boost. Certainly this could be the root of my knock. I would expect something around 11.5:1. This is a bit of a surprise to me, but it makes sense since my WB gauge seems to be about .5:1 lower then what I see in ECMLink. So when I see 11.5:1 on the gauge, ECMLink sees 12:1.

My LC1WB and my AFRatioEst are off by more then a little, meaning my WBFactor can reach double-digits at times.

hmmm

pull 1:
click
pull 2:
click

 

[biglady112]

10.5-11.0:1 is where you need to be on pump gas. I bet making that change alone will make it disappear.

 

[GSTwithPSI]

Neither of the two pulls you posted are at WOT. It's hard to make any real suggestions when the car is only at about 85% throttle.

I'd start by getting your AFRatioEst to line up with your WB. You need bump your global fuel a bit richer since your WB is reading about a point leaner than the AFRatioEst thinks the car should be.

I'd also pull 2 degrees of timing or so starting around 5,500 RPM. As soon as your ECU commands 10 degrees of advance, that seems to be where your knock pops up. I don't like to combat knock by adding fuel. I set my target AFR and boost, then advance or retard timing where necessary.

 

[prove_it]

Are you actually at WOT and your TPS is reading wrong, or are you not getting into it? This will change the load cells used and you could be in an area that is not tuned correctly.

Like GST said, Start with dialing in your WB. Which one are you using? I don't like AEM WB's as they seem to be inaccurate.

 

[EfiniX]

My bad for attempting to replicate the stumble at near-WOT. This is where I typically see the problem, so it wasn't actually a WOT pull. The damn car launches so hard, and the knock comes so early on that when I see (knock), I just come off the gas. The TPS is calibrated correct, but excellent question no the less.

I've actually just gotten an email back from Adam Davis at Innovate regarding the delta between my displayed AFR's on the gauge, and the slightly-leaner AFR's I see in link. He recommended adjusting the voltage offset and I asked him if the gauge is right, or if Link is right.

His answer:

"Yes, that would be correct as the MTX-L display does not use the DAC. It is serial data right off the controller. The analog output is a conversion of
The serial data to a linear voltage output via the DAC."

Before I do anything else, I'm going to see if I can't get the gauge and ECMLink to report the same AFR's. If my actual AFR's are richer then what link is seeing, it could throw off everything else (including my WBfactor).

So bonus question: Given the instruction to up my starting voltage by a few millivolts, where do I do this? I've got an LC1WB item in my log, but I have looked everywhere and cannot find a RAW WB voltage setting or log item.

 

[prove_it]

I wonder if He's referring the the WB switchpoint setting in the NB sim tab.

 

[GSTwithPSI]

According to your log, ECMlink is reading exactly as it should according to the preconfigured setting for the LC-1 WB:



As you can see from your log, the WB is reading 14.7 and the raw voltage output for the sensor is 2.45v, which is right where it should be. Unless you have reason to believe that 2.44 is not stoich, it's reading right. Since neither of us know what the "delta" output of your gauge is, e.g. what voltage correlates to AFR, it's a moot point. You know ECMlink is reading correctly according to the switch point of the LC-1 sensor. Who cares what the gauge says?

Currently, your AFRs are leaner than what link is seeing. This is verified by your AFRatioEst being significantly richer than your actual AFRs. You could modify the switch point of the sensor, but for what? You already know it's set to where it should be, and unless you have another WB to confirm it's off, IMO it would be silly to make adjustments just so your gauge and ECMlink read the same. Just make a mental note of the difference between the gauge reading and ECMlink, and leave it be. If you screw with the WB switch point, you're asking for trouble.

Make your priority getting your AFRatioEst to line up with your WB by adjusting global fuel. That should be your first and foremost concern.

Quoting EfiniX:
Given the instruction to up my starting voltage by a few millivolts, where do I do this? I've got an LC1WB item in my log, but I have looked everywhere and cannot find a RAW WB voltage setting or log item.




I've answered this as simply as I can in your other thread. Let me know if something below doesn't make sense:

ECMlink is configured to read AFR based on your wideband's switch point. As long as you haven't modified it, the LC-1's switch point according to ECMlink is preconfigured by the software to 2.44v at stoichiometric (14.7:1). Simply, this means when the sensor sees 2.44v, ECMlink will show 14.7:1 AFR in your datalog. In order to see this, press F9 and look in the "to be displayed column". You should see whatever you named your WB as. Highlight your WB by clicking on it, and then click the Raw value -> button. At this point, you'll see another list of raw values to add. Click on the one that says in parenthesis Raw(whatever you WB is called). So if your WB is called LC1WB, it should say RawLC1WB. Click OK and you should now see the raw voltage value of your WB in the displayed valves of your datalog below the graph.

Once you have the WB value and raw values displayed, capture a log and review it. Find a point in the log where your WB reads 14.7:1, and you should see the raw value reading very close to 2.44v. If it's not, you have a problem. If it is, then the LC-1 and ECM link are working as they should. It may just be that the gauge is off. The only real way to tell if the 2.44v is actually stoich is to verify your AFR with a second WB. Otherwise, you'll just have to assume that when the LC-1 sees 2.44v, it is indeed stoich. I suppose you could also sudo-double check using your factory narrow band if you still had it installed and knew what it typically cycles at.

You can test the DB gauge in real time. Try to run the car and get the LC-1 voltage to stabilize at 2.44v. At that point, look at your gauge reading, which should right around 14.7:1. If it's not, then it's wrong.

As far as the calibration goes, you are supposed to calibrate it in open air. If the sensor is easy to remove, pull it out of the exhaust and calibrate it. If not, let the car sit overnight (Innovate says at least 8 hours) and then calibrate it. I've done it in the car before and it worked just fine.

Here are a few related helpful links on the topic:
WB switch points
LC-1 Install manual

 

[prove_it]

FWIW I don't use a WB gauge to tell me. I don't believe it's safe to be driving and looking at the damn thing anyway. Color changing leds are much safer, IMO. I use an old school Split second ARM1. It's based off of the WB narrowband output and I use it solely for knowing if something goes wrong. I can see the color out of the corner of my eye and it's hidden. If your out wanting to know your ratio, log it and review it while parked.

Otherwise don't fret it.

/rant.

I would focus on better fuel tuning as your using 109 in the high VE cells. It works, but if your global and base fuel is right, then it shouldn't go over 100, from what I've been told and read off of the ECMlink forums.

Be aware that too much fuel can knock too. Rare, but true.

 

[GSTwithPSI]

^+1. I personally don't have a WB display/gauge. I use ECMlink to tune the car, and leave it. I have the check engine light set to 2 cnts of knock to tell me if sh*t has gone awry.

Your WB/ECMlink is reading true, so go by what it says. Get your AFRatioEst and WB to line up, pull some timing, and I think that will be a good start.

 

[prove_it]

Oh yea, and check your knock sensor for leaking goo. It could read inaccurate if it's failing. Common enough on 20+ yr old sensors. Mine was leaking and I changed it and it cleared up a little phantom knock I had.

 

[EfiniX]

Quoting GSTwithPSI:
According to your log, ECMlink is reading exactly as it should according to the preconfigured setting for the LC-1 WB:



As you can see from your log, the WB is reading 14.7 and the raw voltage output for the sensor is 2.45v, which is right where it should be. Unless you have reason to believe that 2.44 is not stoich, it's reading right. Since neither of us know what the "delta" output of your gauge is, e.g. what voltage correlates to AFR, it's a moot point. You know ECMlink is reading correctly according to the switch point of the LC-1 sensor. Who cares what the gauge says?


Here are a few related helpful links on the topic:
WB switch points
LC-1 Install manual



I feel ya, but here's the challenging bit for me. I've talked with Innovate and they are telling me that the value being displayed on the gauge is real AFR. The value on the gauge represents the actual serial data from the WB sensor.

"Yes, that would be correct as the MTX-L display does not use the DAC. It is serial data right off the controller. The analog output is a conversion of
The serial data to a linear voltage output via the DAC."

There are a number of people telling me to ignore the gauge entirely, that's it just a reference point; that I should be trusting the input to ECMLink. Given what Innovate is telling me, that wouldn't seem to be true. They are telling me that getting the numbers to match (gauge and ECMLink) is needed to ensure that the correct AFR is being reported to Link.

I do get that ECMLink is showing 2.45v at 14.7:1, but isn't that 14.7:1 just a representation of the voltage being received? So if there is something off with the WB itself, it could potentially send a 2.45v analog signal to ECMLink at 14.2:1 instead of 14.7:1, right? That's the calibration that Innovate is telling me to do. To use a LinearWB configuration and adjust the low voltage threshold up a couple milliamps to compensate. That somewhere between the raw serial data the controller uses to display the value on the gauge and the analog value the DAC outputs to ECMLink, something needs to be tweaked.

From a consumer standpoint, this makes sense to me, that the gauge would be the true AFR. Innovate is selling a product that reports AFR and when they include a gauge, I'm sure they want that gauge to be the 100% end-of-story AFR being reported so that they don't have to rely on 3rd-party products such as ECMLink to report true AFR.

Is my thinking correct on all of this? Your advice is absolutely valuable (seriously, if this were Reddit, I'd give you a year of gold) and you certainly have more experience in this area than I do, so thank you for your time.

 

[EfiniX]

oh, and I've had a hell of a time finding that voltage setting in link. Just doesn't seem to be there, but the fact that you can display it leads me to believe I'm nuts?

 

[prove_it]

Did you buy the MXT-L brand new? If not the previous owner may have changed the internal settings of the MTXL.

 

[EfiniX]

brand-spanking new

 

[prove_it]

Any exhaust leaks pre sensor?

 

[EfiniX]

Funny you should ask. My FP racing manifold has warped a bit and did start to leak on the passenger side. I shored it up with EHMotorsports to a point where it chatters when cold, but seals up well when hot. I'm hoping to put off pulling the manifold and sending it of to be resurfaced until I need a new turbo (just purchased a house, so if I can save the money, great).

That being said, I could say the leak is minor, but when troubleshooting, it is low-hanging fruit. However, I'm not clear on why this might cause what I'm seeing. If the gauge is showing actual values, and those values are 0.5:1 richer then what I see in link, how would an exhaust leak cause that delta? I might expect to see the reported AFR actually be leaner then what the actual AFR is because of additional air being introduced into the system, but even then, the value on the gauge and the value in link should be the same, right? I mean the ECU receives and analog voltage signal from the WB controller and then reports that voltage as an AFR based on the voltage settings and the defined lambda. So if the DAC in the WB controller is doing it's job and ECMLink has the correct voltage range and lambda configured, the AFR reported by the gauge using the raw serial input from the WB, and the AFR in ECMLink should be the same, even if they are lean because of a leak.

Is my reasoning correct? I mean my values at idle might seem lean, but once I'm cruising, I still see this delta. it's uniform across all rpm/load ranges.

here's a log from my drive to the office today

click

Obviously it doesn't show you what I'm seeing on the gauge, but before I start tackling my AFRatioEst/WBFactor variance, I want to make sure I'm working on the right problem.

Thank you once again for responding. I really appreciate it.

 

[strokin4dr]

Quote:

There are a number of people telling me to ignore the gauge entirely, that's it just a reference point; that I should be trusting the input to ECMLink. Given what Innovate is telling me, that wouldn't seem to be true. They are telling me that getting the numbers to match (gauge and ECMLink) is needed to ensure that the correct AFR is being reported to Link.



You (your gut) and Innovate are correct. The gauge is the correct reading, not the value displayed in dsmlink.

Don't listen to these guys telling you to go with what you log, that could be very incorrect. Ask Tom Dorris....

To be accurate, you need to do the "linear wideband" setup. CLICK

There are also some threads on the dsmlink board about how inaccurate the preset wideband settings can be. Set it up as a linear wideband, get ecmlink to read EXACTLY what the accurate gauge is displaying and you will be all set.
Good luck!