ECMlink No Idle after switching to SD

[xn35]

Also tried with the biss opened up
Here is the log with the MAP sensor blocked off

 

[We're on Boost]

This is better with the MAP hose closed (post 26). It's running richer now and it even got into closed loop for the first time. O2 voltage stayed above 0.5 volts the whole time though. It should have started cycling at coolant temp of about 88 deg F, with closed loop happening also at 88 deg. But it was late going into closed loop (107 deg F) and even then the O2 didn't start to cycle.

In the logs before, the ones where it actually started, the O2 voltage would start high (richer than stoic) and after 5 seconds or so the O2 would drop to near 0 volts, and then another 5 seconds later usually it would die. A cold engine needs to stay richer than stoic when it's colder than 88 degrees. Yours was starting richer than stoic but then going leaner than stoic and then dying. I don't know why your plugs were wet, probably just from cold starts with never any warm running.

Now you have CrankingFuelAdjust turned on so we can see it, good. Looks like it's working but we can only see it for about 1 second, literally the first 1.8 seconds of the last log show a value for CrankingFuelAdjust. Anyway it is apparently working. That will throw in 60% more fuel during cranking at 70 degrees according to your table for it. Then if it starts, that will taper down to 0 extra fuel over the next 15 seconds or so. So it's very rich during that time but that's what you need to start a cold engine, and I suppose that is why your plugs were wet.

Now it's just staying rich. In closed loop the CombinedFT (Fuel Trim) is -10% which means it's trying to reduce fuel, and I think it's just too rich to go into closed loop. Or maybe there is some other reason why it's not going into closed loop, maybe the nonsense MAP value from having the hose closed.

Obviously you want the MAP hose open, but I think it needs to be richer then, which you'd get either by changing Global Fuel to something like -55% or maybe changing Global deadtime to a bigger number. I would just change Global Fuel to -55% for a first shot. VE numbers I'm sure will get changed at some point and I won't be much help for that.

In your first log (post 1) I don't give much credibility to your wideband numbers because sometimes the older model AEM wideband on the 1g EGR input is not very good, and in that log the wideband numbers look wrong. If the wideband says richer than stoic and the narrow band O2 says leaner than stoic, it's the wideband that's wrong usually. But usually the numbers on the AEM gauge are ok.
From post #5, hate to tell you this now, but instead of buying a new Bosch sensor for your old AEM wideband, you probably should have just thrown the whole thing away and bought a new latest model wideband. But we'll see I guess. Some guys have said that the "R1" scaling has given fair numbers with it (The R1 scaling is different than what AEM tells you in their own manual). But in my mind, you want reliably good numbers for AFR, not some kind of work-around.

Looks to me like your AIT doesn't have a pin assignment.
On that same page you should check the boxes for locking intake temp and EGR temp.

The last log shows ISCPosition, good. Hopefully we will see some number in there greater than 0 when you close the BISS screw back down a bit. Right now it is 0 because it is trying to slow the idle down to the target of 1000 rpm.

 

[xn35]

Thank you for the in-depth reply. It will now Idle. I unlocked and played with the global and set it at -44 on the second log to bring the AFR down from 16 @ -55 on the first log. It jumps around alot but 14.7 seems to be the middle number on the WB gauge at -44. I am going to track down a different gauge for the FPR as this one was off an older one, and I fear it may not be accurate and maybe the fuel pressure is set higher? Fuel per rev is still high. What else do you guys see? I also screwed the BISS back to where it was set before

 

[steve]

What else do you guys see?

Battery or alternator on the way out. It's not charging so it could be the fuse too.

 

[We're on Boost]

Thank you for the in-depth reply. It will now Idle. I unlocked and played with the global and set it at -44 on the second log to bring the AFR down from 16 @ -55 on the first log. It jumps around alot but 14.7 seems to be the middle number on the WB gauge at -44. I am going to track down a different gauge for the FPR as this one was off an older one, and I fear it may not be accurate and maybe the fuel pressure is set higher? Fuel per rev is still high. What else do you guys see? I also screwed the BISS back to where it was set before

In the -55 log it didn't go into closed loop until coolant reached 131 deg F so yeah there was still something there it doesn't like, but the O2 was cycling and we are getting non-zero ISC numbers now which is good to see.
In that log the Combined Fuel Trim is plus 10% so it is adding fuel because the basic settings are too lean still.

In the -44 log it continues to warm up and it is in closed loop still the whole time, good.
In that log the Combined Fuel Trim is -10% so it is subtracting fuel because the basic settings are too rich now. Cool!
Went from +10 to -10 so the "right" setting would be half-way in between.

So the quickest way to "fix" that for now would be to set the Global Fuel to about -49%. Then the Combined Fuel Trim should come out close to 0%.

That won't be the end of changing Global Fuel though because this is only at idle. Global Fuel in the end needs to be set correct for WOT at high Load Factors. But it would be interesting to try -49% even if just to see if closed loop will start up sooner after the cold start. And then we can see how AFRs and Fuel Trims go while driving it in what should be closed loop territory, which is basically anything up to a Load Factor of about 0.9 (no boost).

 

[Stapl3]

That's the correct coolant temp for the switchover.

 

[We're on Boost]

It can go into closed loop at 87 deg but in some conditions it won't. Mine always goes in at 87 degrees after a cold start. If the O2 sensor isn't up over 0.5 volts by the time temp hits 87, then probably it wouldn't go in, or some other conditions, I don't know what they all are for 1g.

From the fueltrimupdatepoints wiki:

You must be logged in to view this image or video.

Steve had this to say in post #13 here:
"The 1G ECU doesn't try to enter closed loop until the coolant hits 87F and if you are already over that when you start there is a timer but I forget how long. This is based on reading the code and observation."

 

[xn35]

Thank you for all the info, guys it is greatly appreciated
I was playing with the global at idle to try and get it to average zero. Stayed in the positives after a drive and lowered it down a bit? What should my pulse width be set at?
Not sure why the battery reads so low I am getting 13.3v at the terminals.
What should the ICS be reading at 1000rpm idle? Should it still be in the 30.0/
I did a couple light revs since I put a new clutch in and wanted to see if it drags at all. Noticed it CEL/knock light come

 

[We're on Boost]

I was playing with the global at idle to try and get it to average zero. Stayed in the positives after a drive and lowered it down a bit?

Your Combined Fuel Trim is plenty close enough for idle, with global set to -49%. Now you need to see where Trim goes in closed loop driving. Cruise and light acceleration. I kind of prefer having the fuel trims slightly negative most of the time rather than slightly positive, but they are always bouncing around some. Lucas also told me he likes to have the trims slightly negative most of the time.

What should my pulse width be set at?

I don't know what setting you mean. Global Deadtime?

What should the ICS be reading at 1000rpm idle? Should it still be in the 30.

Your ISCPosition numbers are good enough for now. Later, after doing longer drives with a lot of minutes at full temp, you might end up adjusting it a little.
Are you still using the FIAV in your throttle body? Or is it bypassed?

Noticed it CEL/knock light come

Probably the knock. Per your "Dash" tab, your CEL comes on at 3 degrees of knock retard.

That knock you are getting in the 3rd log is happening after the throttle blip, after you take your foot off the gas. It's at very light load. So it's not actual knock from detonation. It's some kind of mechanical noise your engine is making, like from the lifters (hydraulic lash adjusters) or some other noise. Phantom knock in other words. It's probably something you will want to figure out at some point but it's not the fault of the tune. The best recent thread I know of for tracking down a phantom knock problem was by @dwb and it is here. In the end he replaced the HLAs and that fixed it. But along the way all kinds of ideas came up and he did a good job of checking a lot of them out.

I don't think you have to be scared to drive the car for a while with the phantom knock. See if the knock events taper off after running for a while fully warmed up.

Not sure why the battery reads so low I am getting 13.3v at the terminals.

Yeah your battery in the log reads a little low but if you get 13.3 volts right on the battery terminals with a multimeter while it is sitting there idling at 1000 rpm, then you are not too bad off. That's about half a volt low. Your alternator must still be working, because if it wasn't working at all, your battery voltage would be less than 12 volts with the engine running. I think your voltage is probably due to old stock wiring and an old alternator. As you probably noticed, the voltage in a log is often times about half a volt lower than the voltage right on the battery.

My car's electrics are not great, but they work fine for good running. I have old stock wiring from the alternator like you probably do, and my alternator is a new NAPA reman 75 amp (like the auto-trans cars came with). I will typically see 13.5 volts while cruising and that will suck down to 12.5 volts during a WOT pull (the 2nd fuel pump kicks in, and fuel pumps draw more amps during boost than they do during cruise). I have 2 cooling fans that run all the time, even during cold start cranking !

If you were going to buy a new or reman alternator, you should probably wait until you have time to put in some kind of heat shield between the alternator and all the hot exhaust parts. And make sure that your PS pump isn't leaking. I went through all that stuff several years ago and it was a lot of work the way I did the heat shield. Most people would do it simpler.

_______________________________________________
LoadFactor is not working in your logs! We have to be able to see that. Major stuff in the tune keys off of load factor. I don't see it in your Available Values. You turned on "everything" a while back, didn't you?

 

[xn35]

Thanks for all the added info.

Good to know about the alternator too. I like the idea of having fans on, so I don't forget.

Still FIAV in the TB

I must not have logged the last one and only streamed it. ISC is a 60 on idle (see screenshot). One is a screenshot from a 5000rpm pull. WB gauge says 11.4/11.5 ECMLink say 11.9

Pulsewidth on fuel table screenshot

I definitely see the CEL while still on the throttle, maybe it is delayed, as that is why I lift off? I may have lifted off before I realized the CEL on the other throttle response
I will definitely look into the phantom knocks.

I really appreciate the time you are taking to help me out.

one the rain stops Ill try and get a few more logs

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[xn35]

I must have missed load factor as it would only allow me to log so many. I will be able to add it for the next logs

 

[We're on Boost]

ISC is a 60 on idle (see screenshot)

If that's what you are usually seeing now, then yeah you could open up the BISS screw (counter-clockwise) a little which would bring the ISCPosi down to around 30.

I asked about the FIAV because if you don't have one, you could let the ISC go all the way down to 0, or some other low number like 10, when the car is 1600 seconds into a log like in that shot. That would be to get more "+" range from the ISC, which helps to cover for not having an FIAV. My car is like that. But if you have the FIAV and it works, then it's better to stick with the regular "30 at hot idle".

Pulsewidth on fuel table screenshot

Oh the minimum pulse width, I think 1385 microseconds is fine for that. So, 1.385 ms would be the smallest InjOn time you would ever see I guess. That's fine. It is still going to 0 when it is supposed to (trailing throttle). Then when the injectors come back on, they come on at more than 1.385ms anyway, so you might never even hit that minimum number.

I definitely see the CEL while still on the throttle, maybe it is delayed, as that is why I lift off? I may have lifted off before I realized the CEL on the other throttle response
I will definitely look into the phantom knocks.

In the 2 logs from post 35 I don't see any real knock and I don't see any phantom knock either. There are little bits of 0.4 or 0.7 knock retard here and there, and most people just disregard those.

WB gauge says 11.4/11.5 ECMLink say 11.9

Well that's pretty close. Probably the gauge is correct. I've been looking at the wideband numbers in the logs and they seem sort-of plausible. So, I don't know. It's kind of impossible to look at your live log on the laptop and your gauge at the same time all while doing a WOT pull. I mean when you get your boost and everything else going good, a WOT pull is going to be scary enough even with no distractions.

For a homemade way to check it better, if you could rig up a cam to shoot video of the wideband gauge, that might be one way to see it safely and correctly. What you could do is, you know how a clapper board is used in film making so the studio can sync the sound track to the video track later ? (because they are recorded by separate devices). Well here the sound and video would be recorded on the same device and the logger is the separate device. You could start the log, then start the camera, then when the log gets to some even seconds like 100 seconds or whatever, you would clap. One loud clap. So when you are looking at the video later you would know that when you hear the clap you are at exactly 100 seconds in the log. Then you would be able to sync the log with the video. You would know that XXXX seconds in the log is YYYY seconds in the video. Something like that.

 

[Stapl3]

What should the ICS be reading at 1000rpm idle? Should it still be in the 30.0

The ISC has an operating range from 0-120. The idea is to aim it at being 30 during idle, as that leaves 75% of it's operation an ability to add airflow. To set it, you want the car to be fully warmed up, headlights off, radio off, etc etc all off. Then target 30 for the ISC by adjusting the BISS.

Say you got it at 30 and you want to put the headlights on, that'll ask the alternator for some more load, the engine will "bog" BUT the ISC will step to compensate. Smooth transaction. Say you turn the AC on while moving the windows up and down and the radio is blaring - the alternator will provide the power for that - but no free lunch, the engine will "bog" and the ISC will compensate. Another smooth transaction if everything works good.

For the most part, the ISC is only going to open more to compensate and that is way more common than the other way around. It is rare that the car will need less airflow so we aim to have the ISC at about 30, so there is 25% working range to go 'backwards' and 75% range to offset loads. 30 is just what the community settled on, do not pull your hair out trying to target it.

 

[We're on Boost]

One is a screenshot from a 5000rpm pull. WB gauge says 11.4/11.5 ECMLink say 11.9

I looked at this shot full size and managed to get a few infos from it
Mostly I noticed there is some actual knock there, pretty much all through the pull with 2.5 deg retard where the cursor line is and probably ~5 deg retard earlier.
There is only 2.8 psi of boost there so it's pretty lightly loaded.
But still, even at that light load, 24.7 degrees of timing is probably too much on gasoline.
I looked for that cell in my timing tables and 25 degrees is what I would be using on E70! (flex fuel setup)
On E10 gasoline, I'd be in the TimingMinOctane table and my timing would be only 20 degrees in that cell.
Pretty sure you need less timing on your map just about everywhere from load factor 1.0 and above.

I'm kind of blank at the moment about what to give you for a gasoline timing map that is "safe" at high boost, except you could copy and paste my TmngMinOct map really easy if I post a log here. Pasting in an entire map is real easy. Then if you don't like it you can revert I think or just paste in a map from one of your previous logs or somebody else's log.
But I am saying, you would paste it in as your MaxOct map and your MinOct map. So it would be your timing, period.
It has pretty low numbers. Take a look, shot below is it, the grayed cell I think is the cell where you had 24.7 deg timing and 2.8 psi of boost.
When ER did this map, the priority was safety for whenever I have to run E10 gasoline. It was used in the "base tune" which they run before adding any ethanol to the tank.

BTW we have similar turbo and cams. It's the fuel that's really different.
Looks like you have some plans possibly to throw methanol on it to kill the knock, and that would let you use more timing, but that would be another whole thing!

You must be logged in to view this image or video.

 

[We're on Boost]

Here's a recent log that you can copy the MinOct table from if you want.
You'll notice I'm driving it at only 25 - 26 psi of boost and with 55% ethanol in the tank.
The ethanol content is shown by the item called FlexEthMix.
This tune has been my tune for 9 years now. It's a little odd in that I don't have an IAT.

 

[nice97civic]

I'm going to take your maxoct timing table...or at least see how it compares to mine (also flex fuel).

 

[xn35]

Here are two logs. Now with loadfactor, one with a small pull. The other is with some general full throttle and driving. Once it started building boost, AFR is down in the low tens.

I had a few litres of old gas from last year and put in a boostane mix of what would be 102 octane.

The meth was in the car when I got it, so we will see how it shakes out before we go down that path. Sadly, fuel options aren't great around here. The most we can get at the pump is 91 or to run race gas. Race gas just doesn't seem reasonable at this point, so I am trying out Boostane.

I put a new FPR gauge in, and it was showing 39psi. So, hoping that this new one is correct, I bumped it back up to 43.5 psi. I played with the global and BISS

Wastegate spring pressures are 15psi and I have a brand new Ceramic FF from TMZ not sure how much boost I should put through it yet. Apparently, they say little to no break in. I am not planning on launching for a while anyway.

Also on the prior two logs, I have your minoctane as my max. I am playing with the WB to get it closer to gauge and AFRest.

In the middle of the log, it is pulling timing. Seems like some pulls won't, and others CEL comes on right away, even before any real boost. AFR is usually 12.0 or richer when this timing is being pulled.

Looks like my rad started leaking on my last logs. explains the low temperature of 180°F. Might explain the knock? Hopefully nothing got messed up from the cooling system being partially empty

 

[We're on Boost]

I'm going to take your maxoct timing table...or at least see how it compares to mine (also flex fuel).

Ok

In the middle of the log, it is pulling timing. Seems like some pulls won't, and others CEL comes on right away, even before any real boost. AFR is usually 12.0 or richer when this timing is being pulled.

I've only looked at the 3rd log so far, the one with knock, and I think it must be phantom knock you are having, rather than actual. Because the timing is pretty low now, the fuel seems reasonable, and it's still at very low boost.
Every knock event in the 3rd log starts at around 3100 rpm, except for the one at 31 seconds, and that is a relatively mild one. Around 3000 rpm is a typical point for phantom knock to start from what I've seen in other people's logs.

If you could see the actual voltage signal from the knock sensor, it wouldn't be spikes all the way through the whole region where we see knock retard in ecmlink. There would be a voltage spike at the front end of it where the actual knock happened, then if there are no more voltage spikes, the knock retard decrements gradually for the next little while like we see in this 3rd log. That much of it works the same whether it is real knock or phantom knock.

Like you said, it's a little hit and miss. In the pull that starts at 37 seconds there is no knock. None. Maybe that's because you started it from about 3500 rpm and maybe that's high enough rpm to avoid annoying that thing that gets jangly at 3100 rpm. So that again sounds like phantom knock.

So, I doubt that reducing timing anymore will help anything. I went and shot some screens anyway of the worst of the knock areas in the log, around 80 seconds. For one thing I wanted to make sure it is using only the MaxOct timing and looks like it is. Although there is an easier way to do it, if you can find a loggable item called "Octane" and capture it, that is supposed to tell us which timing map it is using or the interpolation between.
So in these screen shots, I right-clicked the 2 timing maps, and picked "detach table" from the right click menu, then right clicked again and picked "track datalog". Then you can poke through the log and it will highlight whatever cells are being used at the time. It's a cool thing. If this was real knock, you could do this to see what cells need to get reduced timing. In this case, I doubt that it would help to reduce timing in these areas, but since it is so easy it's not a total waste of time to try it if you wanted just to see what happens.

The coolant temps actually look ideal to me. They got up to around 190 deg F. That's perfect. I suppose it's possible that whatever is doing the phantom knocking might be temp sensitive and that might become apparent at some point.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

I think you could try this knock sensor control function that is on the ALS/Knck tab.
I put the help about it in the shot so you can see what it says about it.
If you agree that it seems like the knock events are almost always starting from about 3100 rpm, then I would try setting the rpm activation point to like 3300 or 3400 rpm. Leave the TPS one at 0 for now.
You'll want to set it back to 2200 rpm occasionally to see if the phantom is getting worse or better.
Eventually the phantom should get fixed I suppose.
People have had regrets about setting this activation point too high in rpm. I don't remember where Marty had his set for a while that turned out badly but I think it was around 5000 rpm. He missed some real knock which wrecked the engine as near as he can tell.

You must be logged in to view this image or video.

 

[Imback]

These are some of the noisiest engines then add to the fact that our knock sensors are old as dirt you almost always have issues with phantom knock. You could always inspect the physical sensor by seeing if the backside is gooey. Since you are on pump i would set the rpm to 3500 and tps at 50% since anything below is just scoting around town or merging onto a highway. Everyone has their own approach on how to tune cars so do what feels safe to you. I agree with it not hurting to see if reducing the timing in that area around 3400. I normally get rid of little spikes of knock like the one in your log by smoothing out the timing and not making the jump from cell to cell so big.

 

[We're on Boost]

I am playing with the WB to get it closer to gauge and AFRest.

Tweaking your wideband scaling to be closer to the gauge should be a good thing to do.
But don't tweak the scaling to make it closer to AFRatioEst.
Wideband and AFRatioEst numbers should stand as seperately derived numbers, because then you can see if they are different and by how much. When you are in open loop you'd like to see them be close to the same, but don't get that by scaling the wb gauge.
As far as I know, when you are in open loop and fully warmed up, the AFRatioEst is basically just tracking what you have in your AFR target tables that are in Direct Access (OpenLoopMaxOct and OpenLoopMinOct). Your actual AFR could be different.

 

[Stapl3]

AFRatioEST has global injector size, deadtime, battery voltage, airflow adjustments, and fuel trims baked in. Probably some other stuff too. It's nearly perfectly exact as for what the ECU is targeting and thinks is happening. It's the best tool of 'link I can think of.

 

[We're on Boost]

AFRatioEST has global injector size, deadtime, battery voltage, airflow adjustments, and fuel trims baked in. Probably some other stuff too. It's nearly perfectly exact as for what the ECU is targeting and thinks is happening. It's the best tool of 'link I can think of.

Right, but one of the things it has baked in is VE, and the VE numbers are numbers that you tell it. While you are working on an unfinished tune, you at first don't have the most accurate VE numbers in there, probably. So some of the info the ecu is using is a little off at first. That's what you work on to finish the tune. During that time the wideband AFR and the AFRatioEst won't always agree everywhere. That's what I meant. Isn't that right? That's why I said, don't scale the wideband to make it agree with the AFRatioEst. You want to see whatever differences exist between those 2 things while you are working on the tune.

 

[Stapl3]

Right, but one of the things it has baked in is VE, and the VE numbers are numbers that you tell it. While you are working on an unfinished tune, you at first don't have the most accurate VE numbers in there, probably. So some of the info the ecu is using is a little off at first. That's what you work on to finish the tune. During that time the wideband AFR and the AFRatioEst won't always agree everywhere. That's what I meant. Isn't that right? That's why I said, don't scale the wideband to make it agree with the AFRatioEst. You want to see whatever differences exist between those 2 things while you are working on the tune.

That is why I only use AFRatioEst at first during peak VE. I'll put the value used at peak VE to 100, then adjust my fuel until WB=AFRatioEst at peak VE, then adjust the table from there.

Can't speak on scaling a wideband as I've never done it.

 

[xn35]

Ended up getting some help remotely. The fuel system seems to be leaking in the tank and wouldnt hold pressure very long after shutting the car down, which would explain some of the global inj setting being off. The car wouldn't deliver more fuel at 35psi because of the leak. The Wideband is scaled closish to the gauge and placed in the front o2 pin now. and then 02 simulated. Ive set the knock up at 3400rpm and that seems to have done the trick so thank you for that. I seem to be getting a lot of pressure loss through the oil drain tube when leak testing. I am not sure how much is acceptable. It holds a little bit better when the engine is hot, but still not great.. It still has the 6165 knock-off turbo that I am looking to replace. Does anyone have experience with the BWS300 sx3 60mm? There is a new unused one nearish to me. I am also looking to go S363.

 

[We're on Boost]

The fuel system seems to be leaking in the tank and wouldnt hold pressure very long after shutting the car down, which would explain some of the global inj setting being off. The car wouldn't deliver more fuel at 35psi because of the leak.

Wow. It wouldn't go over 35 psi of fuel pressure? What did you do to fix that?
Here in this video is one type of fuel pressure test you can do with the car parked and engine not even running, with the fuel pump turned on in ecmlink. I put more info in the "more" field (Description field).