ECMlink LinkV3: BoostEst and boost gauge not equal {attached log}

[razrman8]

I have been doing some tuning to my GSX the past couple weeks and I have encountered something. I have noticed that anywhere in my logs, my BoostEst is not matching what my boost gauge reads.

On the log that I have attached, DSMlink is estimating about 13.6psi. My boost gauge reads about 15 psi, and then creeps beyond that. All of this is not showing in DSMlink. It is not showing that my boost is going above 14psi. Is there some adjustment that I am unaware about that needs to be made in order to get them to somewhat match?

I appreciate any information that is provided to this subject.

 

[GSXTasy005]

Your MAF Comp tab appears to be wrong. How did you get these values? This tab should appear as more of a curve, with little or no spikes anywhere.

 

[razrman8]

I was merely doing what DSMlink was suggesting. I thought they were rather sharp myself. That's how I acquired these values.

 

[Archer Fabrications]

I was merely doing what DSMlink was suggesting. I thought they were rather sharp myself. That's how I acquired these values.

i think i read somewhere boost EST is only really accurate from 5000-5500 RPM your log doesnt work btw.. and i have the most updated software on my computer.

 

[razrman8]

I have the most updated version as well.

 

[razrman8]

Either way, the boost isn't reading the same.

 

[Calan]

Since you aren't logging actual boost, BoostEst is completely useless. Use your wideband and fuel trims to dial in your airflow calibration, and don't worry about BoostEst.

 

[Archer Fabrications]

Since you aren't logging actual boost, BoostEst is completely useless. Use your wideband and fuel trims to dial in your airflow calibration, and don't worry about BoostEst.

i feel like i read about you saying this in another thread

 

[Calan]

Yeah...just a few hours ago. Weird how that works huh?

EDIT:

It was actually yesterday afternoon:

http://www.dsmtuners.com/forums/tun...why-am-i-knocking-link-log.html#post152984268

 

[razrman8]

I get what you're saying.

 

[cjridert1]

Here is a great writeup on MAF calibration: JeffGST
You can do it without a MAP sensor if your memory, multitasking and misc variables are all healthy, but a logged sensor makes it incomparably easier.
So you can get your BoostEst = Gauge # Boost by just changing your MAFComp.

Your other option is to compare your logged A/F ratio to your Direct Access # (11.1 after 3500rpm on the log for you - just look at the logged AFRatioEst). That 4400 lean area is more than likely your MAF sensor not being calibrated. Since your Duty #s are way less then 80% + if your base Fuel Pressure, FPR and pump are all working appropriately this should be correct. My only other concern that this wouldnt work real well is your wideband - AEMs logged vs gauge values are notoriously innacurate, but if it seems about right, great. This is dirty tuning, but should steer you close enough, providing all variables mentioned are working appropriate. Again, a logged MAP sensor is correct, controlled and way easier to work with, plus your can use it for MAFClamp later if you should start running more boost and max out your MAF.

 

[razrman8]

After reading that several times through, I am not quite understanding what is being said.

 

[Calan]

You can do it without a MAP sensor if your memory, multitasking and misc variables are all healthy, but a logged sensor makes it incomparably easier

It's pretty much impossible to do without having a log to study, because even if you could somehow memorize several snapshots of both RPM and boost during a pull, you won't know what load the engine is at for any given RPM.

AEMs logged vs gauge values are notoriously innacurate, but if it seems about right, great.

There are ways to get them pretty close, using a custom linear sensor definition, and comparing the gauge to ECMLink while the sensor is wrapped in a rag soaked in fuel or brake cleaner, etc.. You can vary the amount of fuel in the rag to get different values on the gauge, and then set up the voltage and AFR values in ECMLink so it matches at a few different points. Search the Link forums for more info on custom linear sensor definitions.

 

[razrman8]

That's a little better. Is it worth it trying to get them to match?

 

[Calan]

That's a little better. Is it worth it trying to get them to match?

The wideband gauge and the wideband value shown in ECMLink?

Absolutely. If ECMLink isn't showing the correct measured AFR, you have no point of reference for tuning while reviewing logs. It's also a good idea to compare your wideband gauge and logged value to a known calibrated wideband at a dyno or speed shop.

Calibration is everything...whether you are talking about wideband sensors, boost gauges/sensors, expected airflow to actual airflow, amount of fuel delivered (measure how much fuel the pump can flow, get the injectors tested, etc). Anything that isn't a known value adds some amount of guesswork into the tune, and getting a solid repeatable tune is difficult enough even when you are working with known values for everything possible.

 

[razrman8]

So if what I am interpreting is correct, my BoostEst and actual readings to the boost gauge are all having to deal with MAFcomp? That is the value I am trying to get correct. Not that the AEM wideband needs calibrated as well.

EDIT: Nevermind, I just looked at the posted link above as to adjusting to dial in MAFcomp.

So check DSMlink forums for information on getting AFREst and wideband to match? Could it be that it is in the rear o2 location?

 

[Calan]

Let's get back to basics for a minute...

In stock form, your engine pulled in a known amount of air at every engine operating point (load + RPM), and the ECU was programmed to inject the correct amount of fuel at all of these points. As soon as you change anything that alters the amount of air entering the engine (turbo, bigger intake/exhaust, cams, different air filter, different piping, etc etc etc), the ECU becomes out of calibration, and no longer knows exactly how much air is entering the engine at any given time...so it injects too much or too little fuel and the car runs rich or lean (since it's still working off of what it was originally told, and it doesn't know anything about modifications you've done after the MAF).

So to compensate for this and get the airflow the ECU sees matched back up with what's really coming into the engine, we have MAFComp sliders (MAF Compensation).

Let's say you know the AFR should be 12.0:1 at some operating point, and the ECU knows how much fuel you are really flowing (your injectors are really flowing what they should, the pump voltage is good, base pressure is known, etc). If the measured wideband doesn't show 12.0:1, then it's because the amount of air being reported to the ECU at that one operating point isn't what is really coming in, and the ECU is injecting too much or too little fuel.

So you adjust MAFComp at that airflow point until the AFR matches, which means the ECU now knows what airflow it has to work with (again, at that one point) and is once again injecting the correct amount of fuel. You then apply this same thought process for every point in the airflow range covered by the MAFComp sliders, and you'll end up with a mostly calibrated airflow signal across the entire range of operating conditions, and the measured AFR will always match what the ECU is aiming for.

So, the goal is to get the reported airflow to the ECU to match what is really flowing into the engine, so that it injects the proper amount of fuel. You can either use AFRatioEst and an accurate wideband to get the AFRs to match, or you can use BoostEst and logged boost until those values match. The wideband method is more accurate, and works over the entire operating range. The BoostEst method is only accurate in the range of peak engine VE (5000 to 6000 RPM) at different load (boost) levels, and is a bit more tricky to get dialed in.

In closed loop (idle and cruising around), the ECU is always shooting for an AFR of 14.7:1, and is using the front O2 sensor to constantly measure and correct the AFR error. The amount of error it sees is stored in a "fuel trim" parameter, and this equates directly to how far off in one direction or the other the measured airflow is, assuming the fuel flow is accurate. There are different fuel trims used at different times, but they all represent the same thing; an error in reported airflow vs. actual airflow, as a result of too much or too little fuel being injected while trying to maintain an AFR of 14.7:1. So, you can use the fuel trims to dial in MAFComp while the car is operating in closed loop, since they are giving you the same info as a wideband is...in a round about way.

If you read that a few times, you'll realize how important calibration is. Measured AFR is a result of airflow and fuel delivery; you need one of those to be known, in order to compensate for the other. Fuel flow is much easier to get true measurements for, so we use that as our "known" values, and compensate for airflow. In order to say "my airflow is off, and I need to adjust MAFComp to correct for it", you are making the assumption that your injectors are flowing exactly as much fuel as the ECU is telling them to, and that your wideband is reporting the true AFR. If either one of those are inaccurate, then you have no reference point for adjusting the airflow compensation.

If fuel flow or the wideband isn't accurate and airflow is unknown, it's easy to get into a vicious circle of blindly tweaking things and hoping you hit on a magic combination...although that seems to be a valid tuning method for some people.


Clear as mud?

 

[Archer Fabrications]

well said. now if we can get one of these on timing.............

 

[Calan]

well said. now if we can get one of these on timing.............

Timing is easy; just set it to ANYTHING other than -8* at WOT.

Bet you've heard that before huh?

 

[Archer Fabrications]

.............. NO BUT REALLY!

even dipping down to -3 i dont like but i peak to 2.2*! yeah i really needa figure out why i knock with the evo mod 1 anyway. that will be for my thread when my trans fluid comes in tomorrow! gotta do the mantinance

 

[Calan]

Since I'm in a rambling mood while waiting for the ramen to cook...

I've said this before, but this seems like a good place to say it again; there is a big difference between "tuning" and "calibration"...but most people combine the two and end up going around in that big circle I mentioned up there ^^^^.

"Tuning" is easy and quick. "Instead of an AFR of 11.5:1 and 10* of timing, let's try 12* of timing and an AFR of 12.0:1 and see if power goes up or down." That's it. Plug in some different AFR, timing, and boost numbers, and see if you get positive results. Piece of cake.

So why is it so damn difficult? It's because most people spend very little time on calibration, and end up blindly tweaking things without having a solid reference to compare to. Even if they get lucky and stumble onto a combination that runs great, repeating it under different conditions is almost impossible.

Calibration is very time-consuming, and about as much fun as watching paint dry... unless you are a perfectionist and fascinated by what lies in the details. But it pays off big time once you start to really tune for more power or better mileage. Once your fuel flow is known, and airflow is dialed in... then you can simply tell the ECU "lets try this" and then see what happens.

For example, let's say your airflow and fuel flow are nicely calibrated, and you change some AFR targets from 11.5:1 to 12.0:1 in the AFR DA table, and increase the ECU boost control to 25psi. Your boost gauge then shows 25psi, AFRatioEst is 12.0:1, your wideband gauge reads 12.0:1, the wideband in Link shows 12.0:1, and you see what happens to the horsepower and torque curve. If it's all good, you save the log so you can load it up at a later date and get similar predictable results.

Or you can go this route...

"My AFRatioEst at WOT is 12.0:1, but my wideband is showing 12.5:1 in ECMLink at 5000 rpms and the gauge is showing about 11.5:1. (Combined fuel trim is usually around -8%). My boost gauge is bouncing around, but seems to be at about 22psi. I think my injectors are 680s, but I'm not really sure. I'm getting a lot of knock, but not sure why. I moved the fuel sliders and it seemed to help, but should I add more fuel?"

That ^ is the tuning equivalent of the proverbial Chinese fire drill, and could have been pulled from almost any thread on "tuning" issues that pops up (it wasn't; any resemblance to actual persons is purely coincidental . The only problem this poor make-believe guy has is that nothing is calibrated, so nobody can (or ever will be able to) figure out what's going on. He's so far away from being ready to truly tune his engine that it isn't even funny.

Spend the time (hour after hour if needed) on calibration and getting accurate numbers to plug in wherever possible, and learn to think of it as a separate process from tuning. Do the calibration once and do it correctly...and then you can easily "tune" to your heart's content.


Ok.. ramen is done.

 

[razrman8]

Okay, all of this helped out a very great amount. I see what I need to get done now. It's just figuring out how to get it done that is the tricky part. Calan, thanks for your ramble.

My goals today are to try and get my BoostEst and boost gauge to be the same. Also my wideband and AFRest to be closer.

Though the main thing that I am trying to figure out, is trying to get the wideband and AFRest close, which I'm still not getting.

 

[Calan]

My goals today are to try and get my BoostEst and boost gauge to be the same. Also my wideband and AFRest to be closer.

To keep it simple, pick one method or the other. You'll pull your hair out trying to get your wideband to match AFRatioEst, while also trying to get BoostEst to match actual boost. Just take my word on this.

My suggestion - Ignore BoostEst and focus on getting your fuel trims (CombinedFT) to be as close to 0 as possible during closed loop, using ECMLink's MAFComp adjust tool. Then work on getting your wideband to match AFRatioEst in the open loop areas. This page will help, if you haven't seen it: MAFComp Adjust (CombinedFT)

BTW - When studying your logs and comparing wideband to AFRatioEst, ignore the areas where big changes are happening, such as between shifts and when you first get on or off the throttle. AFR errors at these points are due to factors not related to static airflow adjustments (MAFComp) and can be ignored during this process.

Also...

This should have been mentioned at the beginning, but...

It is ABSOLUTELY VITAL that you have all boost and exhaust leaks fixed before doing anything with the ECU. Any leaks in the intake or exhaust (even small ones) will throw off the measured airflow and/or wideband, and will make all of your calibration efforts useless.

 

[razrman8]

When I was dialing in my CombinedFT trying to get it close to 0 as I could at idle, I was adjusting the injector deadtime. I was following a DSMlink video.

So are you telling me that by adjusting the MAFcomp, that will be changing my CombinedFT? Was I doing it the complete wrong way?

By using the DSMlink's MAFcomp tool, you can see in the log where the first adjusted point is drastic (50hz). I don't get why the adjustment is so high, even if I were to try and smooth it.

 

[Calan]

Remember that I said that the actual AFR is a result of airflow and fuel delivery. When you adjust injector deadtime, you are changing fuel delivery; when you adjust MAFComp, you are changing the ECU's idea of airflow.

So both of these will affect CombinedFT, which is just the AFR error as measured by the front O2 sensor.

WBFactor = AFR error, obtained by comparing a logged wideband to AFRatioEst.
CombinedFT = AFR error, obtained by the ECU monitoring the front O2 sensor and keeping track of corrections it has to make while in closed loop.

It's a balancing act between adjusting injector deadtime and the MAFComp sliders. Deadtime has much more effect at short pulse widths (low fuel flow, idle, light cruising). The ECU is only briefly firing the injector, so the mechanical delay of the injector has more of an effect. When the injectors are open for long periods of time (WOT, high fuel flow), the deadtime isn't as much of a factor. So you have to find a balance between injector deadtime and airflow calibration at different operating points. It's a PITA, but manageable after you get the hang of it.