ECMlink Help with tuning SD. What next?

[Justin DuBois]

Not by any means...

, ok you win

 

[We're on Boost]

This gauge is supposed to have a range of 10-17.

I did change the 4v Lambda value a tiny bit lower than the factory says it should be, to help get my AFR readings in ECM link better aligned with what the engine was doing with fuel trums under closed loop conditions. I found with the "normal" settings, my AFR in ECMLink, were always reading a little leaner than it actually was. The log data also better matches the gauge itself this way. I believe this could be caused by a small discrepency in the voltage supplied, or ground point/wiring etc. I normally use fuel trims to adjust VE rather than wideband log data anyway so I figured this should be OK.

I would be more concerned about changing the lower range setting, because I have nothing else to compare to at WOT, and the guage obviously has no problem providing 0 volts.

I am not sure I can adjust this one (autometer phantom) but it might be worth checking to see if it can be adjusted in the meter, or perhaps even even better idea would be to try and get it calibrated against another sensor.

Ahh, this is something I was thinking about asking you - if the Autometer gauge agrees with the LinWideBand numbers you see in ECMlink. Sounds like you are saying that they were not agreeing when you had the Graph Item Preferences set the way they are "supposed" to be (4 volts = 1.16 lambda = AFR of 17). So you changed Graph Item Preferences to 4 volts = 1.10 lambda and that made the agreement better.

The reason I was going to ask if your gauge agrees with your logs is because the EGRtemp input on our 1g ecus is, according to ECMtuning, not one of our better inputs. It's what we have available when using the IAT input for speed density, but it's, well, here are the 2 pages I know of in the wikis that I can think of with complaints about it. You've probably seen one of these already.

https://www.ecmtuning.com/wiki/externalsensorinput

https://www.ecmtuning.com/wiki/ecuinputdata

So maybe fudging it a little bit in Graph Item Preferences is a fair thing to do, to compensate.
I agree that the other end of the voltage scale (0 volts) is more important.

I see that your Phantom 10 - 17 AFR is model 5779 and it has different user instructions than the one I was looking at. In your instructions it does say very clearly "The signal provided is a linear 0-4 volts output. 0 volts out equals 10.0 Air/Fuel Ratio, 4 volts out equals 17.0 Air/Fuel Ratio." They don't say anything about being able to change that, and there aren't any silly buttons on it.

 

[justacanuck]

Ahh, this is something I was thinking about asking you - if the Autometer gauge agrees with the LinWideBand numbers you see in ECMlink. Sounds like you are saying that they were not agreeing when you had the Graph Item Preferences set the way they are "supposed" to be (4 volts = 1.16 lambda = AFR of 17). So you changed Graph Item Preferences to 4 volts = 1.10 lambda and that made the agreement better.

Exactly. I typically use fuel trims to tune VE in open loop closed loop anyway, but when they didn't match it was confusing me.

I just looked again at one of my recent logs, and I think 1.05 Lamba would help make the AFR agree more closely with ECU fuel trims around the 14.7 AFR range.

The reason I was going to ask if your gauge agrees with your logs is because the EGRtemp input on our 1g ecus is, according to ECMtuning, not one of our better inputs. It's what we have available when using the IAT input for speed density, but it's, well, here are the 2 pages I know of in the wikis that I can think of with complaints about it. You've probably seen one of these already.

https://www.ecmtuning.com/wiki/externalsensorinput

https://www.ecmtuning.com/wiki/ecuinputdata

Yes, I read those, but since they didn't complain about my gauge, I figured it was worth trying.

So maybe fudging it a little bit in Graph Item Preferences is a fair thing to do, to compensate.
I agree that the other end of the voltage scale (0 volts) is more important.

I see that your Phantom 10 - 17 AFR is model 5779 and it has different user instructions than the one I was looking at. In your instructions it does say very clearly "The signal provided is a linear 0-4 volts output. 0 volts out equals 10.0 Air/Fuel Ratio, 4 volts out equals 17.0 Air/Fuel Ratio." They don't say anything about being able to change that, and there aren't any silly buttons on it.

Autometer also says, "Our advanced auto calibration system automatically calibrates your gauge on every startup to assure you of the highest degree of accuracy over the entire lifespan of your sensor (no free air calibration is ever necessary)". So, I assume the meter itself reads correctly, just that the output in ECMLink didn't match exactly.

There may be 2 other issues that could be contributing to the disagreement: Autometer says, "Turbocharged applications should have the sensor installed 4-5" after turbo on the down pipe, however ours is located ~2 FEET downstream. I believe I read that locating the AFR sensor further downstream could make the AFR read a little leaner than the actual AFR. Is this true? But I don't know how I could, or should, better calibrate that sensor output, or if the discrepancy is worth worrying about.

Second, I grounded one of the black wires from the AFR gauge on the inside of the dash area, instead of directly to the ECU ground. According to the instructions, there are 2 black wires, and when we were connecting everything, I wasn't positive which was what, so I thought it safer to connect to chassis ground instead of ECU ground, at least until I could confirm it works, but I never went back to it. Could this be the cause of the discrepancy?

I do wonder if I am running a little richer than the logs indicate. I am having trouble keeping HPbyAir above 9. If the WB is misleading me, possibly by as much as 5%, I could be leaving a fair amount of power on the table. I was doing some more reading on turbos, and I'm inclined to believe that I might be able to get closer to 400hp with ~40lbs/min of airflow. (assuming the airflow numbers from ECMlink are accurate). Is that realistic on 91-93 Octane Fuel?

I should really find a way to log fuel pressure as well although I'm not really seeing any symptoms that would lead me to think I am running low on fuel pressure. Aside from the basic static tests, I really have no idea if the fuel pump is keeping up at higher flow rates, so I could be potentially be tuning my VE table at a lower fuel pressure than desired.

Thoughts? Advice?

 

[steve]

Exactly. I typically use fuel trims to tune VE in open loop anyway, but when they didn't match it was confusing me.

Did you mean closed loop? The trims aren't updated when your in open loop and I forget if they are even factored in the fueling math for open loop.

 

[1990TSIAWDTALON]

Second, I grounded one of the black wires from the AFR gauge on the inside of the dash area, instead of directly to the ECU ground. According to the instructions, there are 2 black wires, and when we were connecting everything, I wasn't positive which was what, so I thought it safer to connect to chassis ground instead of ECU ground, at least until I could confirm it works, but I never went back to it. Could this be the cause of the discrepancy?

Thoughts? Advice?

My tuner always has me ground my AFR gauge to the same point that the ECU grounds at, for the gauge and ECMLink to read the same. I moved my gauge ground, and it did, in fact, start matching up to my ECU readings. I run a LC2 from Innovate, BTW.

 

[justacanuck]

Did you mean closed loop? The trims aren't updated when your in open loop and I forget if they are even factored in the fueling math for open loop.

Right, I meant to say closed loop.

 

[justacanuck]

My tuner always has me ground my AFR gauge to the same point that the ECU grounds at, for the gauge and ECMLink to read the same. I moved my gauge ground, and it did, in fact, start matching up to my ECU readings. I run a LC2 from Innovate, BTW.

Thank you. I will certainly make this change... I was scared of grounding anything directly into ECU harness, just in case the ECU ground was not great, I didn't want to put any extra load on that tiny ECU ground wire. Grounding to the same location makes sense though.

Where is the ECU normally grounded?

...the car is down right now due to a failed alternator. I should have a replacement alternator in a couple of days.

 

[steve]

Where is the ECU normally grounded?

#7 is used for both the power and sensor ground on a 1G.

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

 

[We're on Boost]

Yeah that picture #7 that Steve showed. My AEM wideband and the other 2 AEM gauges ground to a 6mm bolt on the same piece of sheet metal, and it seems to work fine. It's at the orange arrow in this pic:

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

As for your wideband sensor location, 4 to 5 inches after the turbo is closer than the usual recommendation. Makes me wonder what sensor they are using. You may have seen O2 housings made with 2 sensor bungs with one of them tall and finned. That finned one is for if you want to put a wideband sensor that close. It keeps it a little cooler, and usually stepped out a little farther too - I mean out of the pipe a little instead of protruding into the pipe.
Here's where my sensor is, and it works fine.
Notice it is in front of the flex section. I don't trust those things. Seems like they could be an air leak. Any leak is bad of course.

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

I don't remember if we have shown you the wiring diagrams on the hot rod coffee shop web page yet. Engine/ecu/sensors/ignition, on one diagram. They have one for 1990 and one for 1991-94. Sometimes it's real handy to know the differences between 90 and 91-94.
Also note that these diagrams when you open them up full size are 2507 x 1901 pixels which is good and it's too big to post here (will get resized to 2000 pixels on the longest edge). So it's a good idea to save these 2 jpg pictures to your computer to use for ever and ever ho-ho!

 

[We're on Boost]

..the car is down right now due to a failed alternator. I should have a replacement alternator in a couple of days.

You might someday want to make a heat shield for the alternator.
Here's my alternator heat shield thread.
It was kind of a high effort way of doing it, but it is still on the car and so is the same alternator I had then.

https://www.dsmtuners.com/threads/pics-of-my-alternator-heat-shield.521305/

 

[justacanuck]

You might someday want to make a heat shield for the alternator.
Here's my alternator heat shield thread.
It was kind of a high effort way of doing it, but it is still on the car and so is the same alternator I had then.

https://www.dsmtuners.com/threads/pics-of-my-alternator-heat-shield.521305/

That looks like a really good solution. Thanks for sharing that!

 

[We're on Boost]

really have no idea if the fuel pump is keeping up at higher flow rates, so I could be potentially be tuning my VE table at a lower fuel pressure than desired.

Do you know what model of fuel pump is in there? If we knew what pump it was, we could use the flow chart for it to estimate when it will be going over the edge.

 

[justacanuck]

Do you know what model of fuel pump is in there? If we knew what pump it was, we could use the flow chart for it to estimate when it will be going over the edge.

not sure yet... when we tried to remove it, the fittings on those lines were all seized. We did blast them with anti-seize but it's still on the "to-do" list.

In other news, we got the alternator replaced and I continue to refine the SD/VE table. It's running better and better.

Another thing: My understanding was that 1.0 load factor was supposed to represent going into boost, however, our BoostEst and the map sensor do not agree at that load range. So, what I did was reduce the open loop thresholds so it goes into open loop as soon as it hits boost, at .85 load factor and I modified the AFR table to start getting richer than stoic at .85/.9 load factor. This should help prevent it from running so lean when first going into boost and I am hoping this reduces the incidence of knock we used to encounter at relatively low throttle inputs.

We also ordered some "7" heat range spark plugs. Once we have those installed, I am going to try advancing the timing again.

 

[We're on Boost]

I modified the AFR table to start getting richer than stoic at .85/.9 load factor. This should help prevent it from running so lean when first going into boost and I am hoping this reduces the incidence of knock we used to encounter at relatively low throttle inputs.

Yes, I think if you have it adjusted so that the closed loop trims are a little negative ("pulling") in this region just before and as you go into open loop, then, as soon as it goes into open loop it should immediately switch from trying to be stoic to running whatever your table says which will be richer.

when we tried to remove it, the fittings on those lines were all seized.

Those fittings are always seized. One way to deal with it is - if you can cut through the steel crimp collar with a dremel wheel or a die grinder, without ruining the barb that is inside the rubber hose, then you have it like this:

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

That barb is perfect fit for 5/16" fuel hose.

Then you can get it put together like this:

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

The fitting underneath the car is harder to work on and I was a little worried about fire hazard and all, but it's disconnected from the tank when you are doing this and you'll have the tank cover closed of course. Anyway I did it like this in 2009 and sometimes I wish it was still this way, it was so clean, simple, and cheap. People get excited about the stock steel pipe under the car being possibly too small to flow "enough", but it just makes your fuel pump run at a few psi higher pressure than it would with a -6 or -8 hose running that whole length. We have pumps now that run to about 20 psi higher pressures than what we had 15 years ago (Walbro GSS351 for near stock fit). So they can do more pressure and not hit that "knee" in their flowrate vs pressure curve where they fall off so fast. So I think you could go quite a ways on gasoline with that stock steel fuel pipe underneath before it becomes a problem.
Or if the steel fuel pipe is way too rusty, or too mangled for some reason, then you might want to get rid of it.
Or just do whatever seems simpler!

Here's how mine looked underneath, with the new 5/16" hose and double fuel injection type hose clamps:

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

 

[justacanuck]

Installed brand new NGK BPR7ES plugs gapped at 0.027.

We did one WOT pull and spent a bit of time idling, while we were checking timing and BISS.

Any plug reading gurus out there? (Photo attached)

The old plugs are NGK BR7ES which were gapped at ~0.029. I have attached of picture of it as well.

Feedback welcome

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.

 

[Justin DuBois]

P is for Projected:

Found this on the EVO forums.
"The B(P)R7ES will foul less... they are a warmer plug. The P = Projected plug which is kinda like making it even a little warmer... it should technically foul less with the P but it may knock slightly more."

 

[justacanuck]

P is for Projected:

Found this on the EVO forums.
"The B(P)R7ES will foul less... they are a warmer plug. The P = Projected plug which is kinda like making it even a little warmer... it should technically foul less with the P but it may knock slightly more."

We installed the BPR8ES plugs yesterday. Car runs fine, but still seeing excessive knock at part throttle. Picture of plug and log attached.

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

 

[Justin DuBois]

We installed the BPR8ES plugs yesterday. Car runs fine, but still seeing excessive knock at part throttle. Picture of plug and log attached.

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

Knock is what the EVO forum warned about with a P - Projected plug.
I'll take a look at the logs

 

[Justin DuBois]

Knock is what the EVO forum warned about with a P - Projected plug.
I'll take a look at the logs

Yeah, 18 deg and part throttle 22%TPS and you got knock retard.
I have my ALS/Knock set to ignore knock below 30% TPS, but I'm also running E25 to help manage the terrible octane of 91 pump gas.

Remind me what Fuel you are tuning for?

 

[justacanuck]

Yeah, 18 deg and part throttle 22%TPS and you got knock retard.
I have my ALS/Knock set to ignore knock below 30% TPS, but I'm also running E25 to help manage the terrible octane of 91 pump gas.

Remind me what Fuel you are tuning for?

91 :-(

We did run a full tank of 93 in it, but it really didn't seem to make any significant difference.

Worked on it for a few more hours today:
- moved variable intake cam setting to 3degrees advance. (was 6)
- triple check base timing = 5degrees
- double check TPS sensor (got 5V with full pedal, but still goes well below 0.62V with 0% throttle)
- as a result, still have to use TPS adjust to simulate idle switch otherwise it doesn't know we get into light throttle

Couldn't keep the car idling when trying to set base timing, unless we give it some small amount of throttle input, otherwise it surges and dies. Idle quality is even worse than before I started. Still surging slightly at idle, still wants to die at stop signs. Something must be wrong with this throttle body and/or we have developed a boost leak again.

Remove knock sensor. Most of the goop is still covering the internal mechanisms so it isn't in as rough shape as I thought it might be. I even tried putting some thread tape on the threads before re-installing it. Still knocks. This knock sensor is really frustrating. I'm really starting to wish we had just purchased a Haltec ECU, and started replacing crappy factory sensors with better quality aftermarket units sensors. I just can't seem to get the knock to stop unless we really turn down large swaths of the timing table which, of course, gives it really weak response.

 

[We're on Boost]

Couldn't keep the car idling when trying to set base timing, unless we give it some small amount of throttle input, otherwise it surges and dies. Idle quality is even worse than before I started. Still surging slightly at idle, still wants to die at stop signs. Something must be wrong with this throttle body and/or we have developed a boost leak again.

You might have a vacuum leak or you might have a bad ISC. I don't think you've replaced yours have you?
The ISC has a test or 2 but it's almost a waste of time because a bad ISC tests good usually from what I've seen.
Anyway, in this last log, the 15-18 log, it looks pretty likely that the ISC is bad. When mine was bad I'd get intermittent high idle speeds like that and then maybe it would die. When it wants to idle high, the rpm then goes even higher because it hits the high timing that the stock maps have at low loads. So even after I put in a new ISC, I also drastically flattened out my timing at low loads up to about 3500 rpm:

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

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

A new ISC is about $60 from Rock Auto and they are good aftermarket brands.
The one on my car is the Standard Motor Products AC146. The WVE 2H1098 looks good too, I have one of those for a spare.
ISC for 1g DSM

I just can't seem to get the knock to stop unless we really turn down large swaths of the timing table which, of course, gives it really weak response

Your knock at 12 sec in the last log, 15-18, doesn't look so far off because the timing is 25 degrees.
But in the log before that, 14-08, the knock at 31 seconds seems weird, because the timing is only 15 degrees there.
Have to wonder if that is really detonation knock or some other noise.

It would be interesting to throw some ethanol or methanol at it somehow to see if that type of knock goes away.
I noticed there is a VP Fuels dealer in Calgary. Looks like about 180 miles away from you. Hmm.
A 5 gallon bucket of X98 (their fancy E98) costs $71 USD.
A 5 gallon bucket of M1 Methanol costs $61 USD.
Here's the Facebook page for that dealer. They look pretty real deal from what I can see.
VP dealer in Calgary
If you let your tank get real low sometime and then threw about 3 gallons of E98 into it, and then retuned for it at your usual Triple Fast Forward work speed, you could probably get some interesting logs from it.
If you actually did this test you'd have to be careful about running out of injector, which would happen pretty early in the rev range!
VP X98

 

[justacanuck]

You might have a vacuum leak or you might have a bad ISC. I don't think you've replaced yours have you?
The ISC has a test or 2 but it's almost a waste of time because a bad ISC tests good usually from what I've seen.
Anyway, in this last log, the 15-18 log, it looks pretty likely that the ISC is bad. When mine was bad I'd get intermittent high idle speeds like that and then maybe it would die.
A new ISC is about $60 from Rock Auto and they are good aftermarket brands.
The one on my car is the Standard Motor Products AC146. The WVE 2H1098 looks good too, I have one of those for a spare.
ISC for 1g DSM

Yes, we already replaced the ISC with a RockAuto inventory item. Seems to be responding as expected.

When it wants to idle high, the rpm then goes even higher because it hits the high timing that the stock maps have at low loads. So even after I put in a new ISC, I also drastically flattened out my timing at low loads up to about 3500 rpm:

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

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

Thanks for this. I'll try that.

Your knock at 12 sec in the last log, 15-18, doesn't look so far off because the timing is 25 degrees.
But in the log before that, 14-08, the knock at 31 seconds seems weird, because the timing is only 15 degrees there.
Have to wonder if that is really detonation knock or some other noise.

Exactly!

It would be interesting to throw some ethanol or methanol at it somehow to see if that type of knock goes away.
I noticed there is a VP Fuels dealer in Calgary. Looks like about 180 miles away from you. Hmm.
A 5 gallon bucket of X98 (their fancy E98) costs $71 USD.
A 5 gallon bucket of M1 Methanol costs $61 USD.
Here's the Facebook page for that dealer. They look pretty real deal from what I can see.
VP dealer in Calgary
If you let your tank get real low sometime and then threw about 3 gallons of E98 into it, and then retuned for it at your usual Triple Fast Forward work speed, you could probably get some interesting logs from it.
If you actually did this test you'd have to be careful about running out of injector, which would happen pretty early in the rev range!
VP X98

Yes, I believe they are the real deal. We will probably take it to the 1/4 mile drag strip sometime this summer and try this. Some of the other racers are usually pretty willing to sell a few gallons, so I'm hoping to give this a try on our new stroker as well. It also has 1000cc injectors so we should be able to push it quite a bit further.

 

[justacanuck]

You might have a vacuum leak or you might have a bad ISC. I don't think you've replaced yours have you?

I removed and disassembled the throttle body again last night. I messaged the brass plate with a file to remove some high spots to help give it a better seal, and smashed the throttle plate screws with a punch to ensure they can never fall out. We also replaced the BISS o-ring with an official original replacement.

We are re-installing it right now. I'll do another leak test and see if the idle problems are fixed.

 

[justacanuck]

I removed and disassembled the throttle body again last night. I messaged the brass plate with a file to remove some high spots to help give it a better seal, and smashed the throttle plate screws with a punch to ensure they can never fall out. We also replaced the BISS o-ring with an official original replacement.

We are re-installing it right now. I'll do another leak test and see if the idle problems are fixed.

Idle problems not fixed.

I can confirm that the ISC is the same model, "STANDARD MOTOR PRODUCTS AC146," mentioned above, but it does not seem to be working properly IMO.

We have the BISS set to provide a baseline ISC position of ~30 at warm idle. When we rev it up, the ISC position rises, then starts closing as the RPM drops and doesn't reverse course in time to catch the idle before the RPM drops too low. Maybe I don't understand how it is supposed to work, but it almost seems like it is backward.

The timing is also jumping around way more than I would expect. How does it go from 20 degrees to ~4 degrees in a couple of seconds? There must be another idle control timing table built into the ECU that we simply cannot see.

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

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

Do we have a new boost leak somewhere?

We also tried hooking the PCV valve on the valve cover to the intake manifold. This made the car idle too high, and the ISC position went to ZERO. It held a really strong idle and didn't surge and didn't want to die out. But it was sucking air through from the other side of the valve cover. We have reverted to letting the valve cover vent to the atmosphere.

I was enjoying learning how to tune with the ECMLinkV3, but I'm starting to wish we had just spent the money on a Haltech Elite. There are just too many weird and unexplained behaviours for my liking:
- TPS Delta doesn't seem to work
- idle timing jumps around like crazy
- ISC doesn't want to keep the engine running properly
- knock sensor doesn't seem to be a very reliable method of detecting knock
- documentation isn't great
- community support is excellent, but having proper documentation and vendor support would be nice when the documentation doesn't make things clear.

To be clear, I'm not really disappointed with ECMLink. I they likely did a great job, considering the ECU is just a "black box." However, additional functionality must be burned into the factory ECU that either can't be accurately reverse-engineered or perhaps it has simply just not been exposed to us in the software.

 

[enriquez2000]

Idle problems not fixed.

I can confirm that the ISC is the same model, "STANDARD MOTOR PRODUCTS AC146," mentioned above, but it does not seem to be working properly IMO.

We have the BISS set to provide a baseline ISC position of ~30 at warm idle. When we rev it up, the ISC position rises, then starts closing as the RPM drops and doesn't reverse course in time to catch the idle before the RPM drops too low. Maybe I don't understand how it is supposed to work, but it almost seems like it is backward.

The timing is also jumping around way more than I would expect. How does it go from 20 degrees to ~4 degrees in a couple of seconds? There must be another idle control timing table built into the ECU that we simply cannot see.

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

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

Do we have a new boost leak somewhere?

We also tried hooking the PCV valve on the valve cover to the intake manifold. This made the car idle too high, and the ISC position went to ZERO. It held a really strong idle and didn't surge and didn't want to die out. But it was sucking air through from the other side of the valve cover. We have reverted to letting the valve cover vent to the atmosphere.

I was enjoying learning how to tune with the ECMLinkV3, but I'm starting to wish we had just spent the money on a Haltech Elite. There are just too many weird and unexplained behaviours for my liking:
- TPS Delta doesn't seem to work
- idle timing jumps around like crazy
- ISC doesn't want to keep the engine running properly
- knock sensor doesn't seem to be a very reliable method of detecting knock
- documentation isn't great
- community support is excellent, but having proper documentation and vendor support would be nice when the documentation doesn't make things clear.

To be clear, I'm not really disappointed with ECMLink. I they likely did a great job, considering the ECU is just a "black box." However, additional functionality must be burned into the factory ECU that either can't be accurately reverse-engineered or perhaps it has simply just not been exposed to us in the software.

Ecmlink does a great job for most of us. If I was going to ever leave it behind it would be for fueltech. Just so many features and pretty easy to configure. I believe fueltech would require us with 1g to add a crank trigger... but thats not a horrible thing to do