Plot and analyze your .csv logs

[CasopoliS]

CSV LOG PLOTTER (http://www.waygate.net/plotter/)- online program

Airboy CSV ANALYZER (http://www.enginuity.org/viewtopic.php?t=957) - really robust Excel file


CSV Log Plotter (online):
This is a neat tool that allows you to analyze your .csv log and produce plots.

SAMPLE PLOTS (you can select which parameters to plot)

http://img444.imageshack.us/img444/5016/plot2zf.jpg

XY scatter - Boost vs. RPM

http://img207.imageshack.us/img207/706/plot25pe.jpg

[lookslikeanevo]

nice
i was thinking about this the other night

[2005garnetGT]

you should check out airboy's awesome log analyzer spreadsheet right here:
http://www.enginuity.org/viewtopic.php?t=957

yes, you need excel, but its awesome.

[mickeyd2005]

^^^ Thanks... that's a goody.

Except for the error in his VBA script. It appears he renamed the sheets but forgot to change the VBA script... oh well, easy fix.

[mickeyd2005]

What I really like is that it calculates load... or comes close to calculating the same load as the ECU. I wonder why the ECU calculation is somewhat different.

Anyway, that'll help 2006 owners. I don't think anyone every did post the location for the 2006 load. And, that's one less variable for me to log.

[2005garnetGT]

I usually just calculated load myself, honestly.
its a derived function, so why waste the cycle time.

[mickeyd2005]

As a noob, I do the opposite. I have always logged it (because that's what the instructions says) so this is the first time that I have bothered to look at a spreadsheet calculated value.

It varies and is usually 1 to 2% off but can be as high as 4.6% off. It should be identical except for round off error since it is a derived value and is derived from the same ECU recorded values (rpm and MAF).

I think the answer lies in the sampling rate/interval. Internally, the ECU has access to much more instantaneous values for RPM and MAF. Calculating the load from data sampled at 0.2 second intervals results in jerky load plots. See attached. I think this can be smoothed out by deriving interpolated values for MAF between samples and then adding a small delay correction. Of course, you can just look at the graph and see the jerkiness...

Just a guess. Please correct me if I'm wrong.

Anyway, this is a nice spreadsheet. I don't know anything about tuning and what a tuner wants to see and how he wants to see it, but I'm going to use it as a starting point.

[2005garnetGT]

are you having boost oscillations? mine don't look anything like that...

[CasopoliS]

or not taking readings quick enough?

[mickeyd2005]

Maybe.

Can you send me a sample log with MAF and boost?

[mickeyd2005]

This is not anything mechanical with the car because any anomalies in the car would still be reflected in the engine load calculation.

The calculated load is definitely off by a small amount (a few percent at most). Mine is exagerrated because I am looking at a very narrow band of load between 2.1 and 2.7 g.

I am pretty sure this is due to the sampling rate. Every parameter is read at approximately 19 ms intervals. So, if I am recording 10 parameters, the parameters can be anywhere between 19 and 190 ms out of phase. If the parameters are read in the order that ECU explorer records them, then I think it would be a simple matter to write an interpolator to correct for the phase difference.

I plotted edmundu's load. His is smoother because his points are so far apart due to his big power. However, even in his, I see about a 2% error.

If you scale his plot from 2.1 to 2.7 then you will see a bigger difference.

BTW, do you use this excel spreadsheet a lot? I'm finding a lot of bugs in the VBA code especially if you don't run it exactly like the way the author uses it such as grab a .csv file from a different subdirectory. I think it could use some better error trapping.

[2005garnetGT]

I mainly use it for timing analysis.

you should probably post something about it over at the thread I linked.

[MarcusDubya77]

Delta Dash has a similar program w/ freeware, you just can't use some of the nifty options..you all probably knew that, but I wanted to write something so that I could auto-subscribe to this thread!

[CasopoliS]

updated - put new link in post 1

[SubiGT]

Is there a description somewhere of what engine load is? These graphing aps are pretty cool:)

[2005garnetGT]

I see another sticky in the making...

[Jon [in CT]]

Is there a description somewhere of what engine load is?In this case it's the amount of air (in grams) consumed by the engine during each revolution.

[2005garnetGT]

I wrote it up once over at scoobypedia...

its something like maf/(2*rpm/60)

[SubiGT]

thx:)

[mickeyd2005]

FYI. Airboy updated his spreadsheet. It's really good and I recommend everyone use it to overlay their datalogs on their rom data.

Also, I modified Airboy's spreadsheet to scale the load. For example, the stock timing tables only go to 2.5 grams load, but it would be nice to scale it to 3.0 grams load for stage 2. I used to modify it manually but a smooth scale would be nice for surface plots and for using ecuflash linear interpolation. If you guys use my function, please let me know if there are any bugs. I've already found one.

http://www.enginuity.org/viewtopic.php?p=16328#16328

[CasopoliS]

FYI. Airboy updated his spreadsheet. It's really good and I recommend everyone use it to overlay their datalogs on their rom data.

Also, I modified Airboy's spreadsheet to scale the load. For example, the stock timing tables only go to 2.5 grams load, but it would be nice to scale it to 3.0 grams load for stage 2. I used to modify it manually but a smooth scale would be nice for surface plots and for using ecuflash linear interpolation. If you guys use my function, please let me know if there are any bugs. I've already found one.

http://www.enginuity.org/viewtopic.php?p=16328#16328

thanks!

[hogmeat]

Holy Crap This Is Awesome. I Can't Believe This Only Has 21 Reples

[mickeyd2005]

Hogmeat, have you tried datalogging with enginuity? It has more defs than ecuexplorer so you can datalog turbo dynamics.

Also, if you are running a Cobb OTS map, ask someone with street tuner to give you the tables so that you can paste it into Airboy's spreadsheet. The spreadsheet is a lot more powerful if you have the map data.

[hogmeat]

no - i'm running (currently) a PDX map.
What is turbo dynamics? Never heard of it...
explain?

[mickeyd2005]

It measures how far off target boost is. IIRC, your log showed 90.2 WGDC and about 10.5 psi at redline. That boost seems low to me considering you have no cats.

[hogmeat]

well, the 90.2 was a PDX thing - it jumps to 90.2 and stays there, no matter what. And my redline boost is about 11.5 - but recall it's only a stage 1 map...

[mickeyd2005]

This is not a stage 1 map:

http://putstuff.putfile.com/63915/4983490/3

Stage 1 maps all have a peak boost target of 15-16 psi at about 3500 rpm and none of them have as high a wgdc of 90.2 That's too dangerous with the 2nd cat in place. Stage 1 maps assume 2nd cat in place.

Your map has the hallmark of a PDX stage 2 map.

[MarcusDubya77]

can somebody take a look at these, and tell me whatcha think??

[CasopoliS]

can somebody take a look at these, and tell me whatcha think??

first off, you can get rid of logging fuel pump duty and throttle voltages

some timing is getting pulled at 3200 rpm but it recovers quickly and KC gets back to 6. Not sure what would cause this... maybe boost building quickly or something, or trying to hit targets.

good redline MAF and MPH

[MarcusDubya77]

well, Casapolis, it looks like I might not be able to get an Actuator & injectors installed..I am gonna check out a local tuner and see if he can tweek my setup...these numbers just seem so low to be running VF40/18G , pump, and other goodies

[mickeyd2005]

What are the mods? And what is the turbo? The spool is late compared to a VF40 so I'm assuming you have a larger than stock turbo.

There's something really odd about this log. At 6000 rpm, the turbo is blowing roughly 14 psi but your MAF is only showing 212-220 g/s. Usually the amount of air consumed by the engine at WOT is a function of volumetric efficiency, pressure, and displacement. For example, at 12 psi, I push 240 g/s at 6000 rpm so I would expect your engine to push even more.

[CasopoliS]

oh if there are substantial mods then yea that log looks way off....

[MarcusDubya77]

lets see...running VF40/18G, Perrin TMIC, Catless UP, TBE w/ HFC, AVO Panel Filter, and a high flow fuel pump... now am I low, and how bad...also I am a 5EAT

[2005garnetGT]

5EAT won't affect the air flowed, though

[mickeyd2005]

FYI. I added an acceleration plotter to Airboy's spreadsheet. Check it out and let me know what you think. I can calibrate it to show horsepower and torque if I can get some data logs and matching dyno charts.

http://www.enginuity.org/viewtopic.php?p=17607#17607

[mickeyd2005]

I added some more spreadsheets.

Here is one to help you fine tune your MAF table. Be careful. You need to match your upper range MAF table (greater than 60 g/s) to your lower range MAF table (less than 60 g/s).

http://www.enginuity.org/viewtopic.php?t=1871

Here is another one to help monitor your KC across the entire map. You need to datalog for a long time and drive like an @sshat to get all the numbers. However, try to minimize the amount of time where you completely let go of the gas pedal. Coasting in gear changes the ecu settings. The spreadsheet is still in work. Hopefully, some people will try it out and we can adjust the logic as needed.

It's in post #13 of the thread below.

http://www.enginuity.org/viewtopic.php?t=2029

[MarcusDubya77]

Somebody please check these out...and help me understand why my car is not producing the power at the lower range..ergo allowing it to make top end power??? Could the ignition timing issue be the cause, because it only experiences it when I get above 18.427 PSI..

Could my 50/50 AVO BOV cause this?? I wouldn't think so, but if it would do i need to loosen or tighten the screw...

[mickeyd2005]

Don't do anymore WOT logs past 4000 rpm unless you make a change that you believe will fix this.

Other people have had a hard time interpreting me so I'm going to be blunt. YOU HAVE A VERY BAD TUNE.

It has nothing to do with your BOV. Although your BOV may cause other problems. They are not related to your logs.

I'm pretty sure I know what the problem is but you should really consult a professional tuner.

In the meantime, delete the following parameters from your logs (logging too many parameters reduces resolution):

Intake Air Temperature. Write down the ambient air temperature. We can figure out the IAT from that.
Fuel Injector Duty Cycle. We've determined that your injectors are okay. If your MAF starts going above 260 g/s, then you will need to start logging it again.
Manifold Relative Pressure. You're already logging the corrected MRP.
Fuel Pump Duty.
Knock Signal. You'll never see anything here because the resolution isn't fine enough. You'll almost always miss the knock signal.
Main Throttle Sensor.
Sub Throttle Sensor.

Add the following to your datalog:

Boost Error.
Fine Correction or Fine Learning (degrees).
Feedback Correction (degrees).
AFR Learning #1.

And check your IAM before and after your log.

The Perrin TMIC actually does not cool as well as the stocker (measured outlet temperature using thermocouple). This will contribute to your tendency to knock especially considering your ambient temperature appears to be in the mid to high 90's. The Perrin TMIC's benefit is that it flows better. My friend determined that it flows approximately 10 g/s better than the stocker and has about 1 psi less pressure drop. It also doesn't have end tabs to blow off. However, it does not cool as well.

It's also probably time that you got a wideband O2 sensor and put it in your DP bellmouth. You might be running too lean for your temperatures / TMIC combination.

Remember, do not datalog past 4000 rpm. I don't want to hear about your engine rebuild.

[MarcusDubya77]

no DP bell mouth here..also what is IAM?..but thanks for all of the input..i am hoping the tuner gets this fixed on the 11th...

[MarcusDubya77]

I'm pretty sure I know what the problem is but you should really consult a professional tuner. Please enlighten me, so I might be able to assist the tuner with what the possible cause could be...I do know he has never done any tuning on a VF40/18G, so he has started from scratch with one of his straight up 18G maps

[MarcusDubya77]

Ok...logged data with what I think are the parameters you stated, some are nonexistent on ECU explorer...I did notice, throughout this ordeal that at above 19 PSI is when knock correction goes bad..

[mickeyd2005]

You didn't log boost error, fine learning, or feedback correction. Those can be helpful for diagnosing problems.

Don't log AFR Correction #1. It's useless during WOT.
Don't log AFR (mA). You're already logging lambda.
Don't log AFR Correction #3. That's the rear O2 sensor.
Don't log AFR #1 resistance.

It's hard to diagnose over the internet and also because you didn't log some of the important variables.

Based upon what you've posted, my guess is that you are running about 1 to 1.5 psi boost too much for your setup.

I also believe that you are running slightly too lean. I don't know what's causing it. It could be the tuner set up your fuel table too lean or your MAF scaling is slightly off. Keep in mind that the stock MAF scale varies slightly from car to car. That's one of the reasons why the stock fuel table is conservative. Your AFR learning #1 is -3.13 %. This AFR learning is based upon the front O2 sensor when it reads between 50 and 60 g/s. It applies this for all MAF > 60 g/s even if it is not needed. So, let's say your MAF table is perfect at 100 g/s but off at 50 g/s then your AFR learning will skew your readings at 100 g/s. In your case, it is subtracting 3.13 % fuel thus making it that much leaner.

In any case, I think you're running too lean. At 3000 rpm and 2.0 grams load, I would normally run AFR 12.0 and I have a stock TMIC (cools better than Perrin) and ambient temperature is < 90 F here.

At 3000 rpm and 2.0 grams load, your front O2 sensor is transitioning between lambda 0.79 to 0.75. The front O2 sensor is skewed because of the high exhaust pressure in the exhaust manifold. In my limited experience, I've noticed that when the front O2 sensor transitions to lambda 0.75, this indicates a true AFR of roughly 12.5. Anything that reads lambda 0.75 on the front O2 sensor from then on is less than AFR 12.5.

This is why I think you need a wide band O2 in your DP.

I really want to see your boost error because your WGDC values are set up funny.

So in sumary, my belief is that your lean AFR is causing the slight knocking between 2800 rpm and 4000 rpm. Your high boost is causing the heavy knocking above 4000 rpm.

[MarcusDubya77]

alolt of these that you are asking me to log must be listed as something else on ECU explorer..I searched through the listing of variables, and the only ones that were remotely close to the names were the ones I selected..are they called something different on ECU explorer?

[mickeyd2005]

Use enginuity to log and make sure you have the latest definitions.

http://www.enginuity.org/viewtopic.php?t=1642

IAM = ignition advance multiplier. This is used for rough correction.

KC = advance x IAM + feedback correction + fine learning

feedback correction and fine learning are always negative.

[MarcusDubya77]

So I downloaded and ran the enginuity..Seems like a great program, but at the bottom it flashes every have second that its unable to connect,check serial port...blah..blah, but I can look and tell that its reading the ECU is this common??

[MarcusDubya77]

Do I need to set up a different engine load with Enginuity like you do on ECU explorer FF0B0?? Also, what is the open image deal on the main screen of Enginuity

[mickeyd2005]

No. Enginuity has a look up table (defs). It will automatically find the correct IAM location in ram for you.

Make sure that you have the latest defs so that you can log boost error, fine correction, and feedback correction.

[MarcusDubya77]

Ok, I plan on datalogging the improvements of the dyno tune

[MarcusDubya77]

what are these "images" deals on enginuity??

[praedet]

Bumping this thread becuase I plan on using it soon...

[Carter]

Bumping this thread becuase I plan on using it soon...
Dont get too comfortable...Waygate is offline now. It no longer works.

[praedet]

So, where can I get this stuff?

[Carter]

Ive been looking for a while and havent found anything.

[mickeyd2005]

I've been working with Airboy on his spreadsheet. It's awesome. It's really taken a life of its own.

A road dyno has been added. It has the ability to scale load quickly and easily. 07 STi tuning tips have been added. Ability to save curves from previous runs so that you can overlay them.

The only problem is that the documentation is set up for someone who understands EXCEL and tuning. Someone really needs to setup a simple step by step procedure for beginners.

Also, the tables have been set up based upon enginuity defs. For example, the throttle columns in the boost tables represent the actual throttle angle and not the ecu direct values. I believe ST uses mostly ecu direct values.

Ted, PM me your email address. I'll send you a copy.

[CasopoliS]

Ted, PM me your email address. I'll send you a copy.

Need me to host it?

[LittleBlueGT]

I've been working with Airboy on his spreadsheet. It's awesome. It's really taken a life of its own.

A road dyno has been added. It has the ability to scale load quickly and easily. 07 STi tuning tips have been added. Ability to save curves from previous runs so that you can overlay them.

The only problem is that the documentation is set up for someone who understands EXCEL and tuning. Someone really needs to setup a simple step by step procedure for beginners.

Also, the tables have been set up based upon enginuity defs. For example, the throttle columns in the boost tables represent the actual throttle angle and not the ecu direct values. I believe ST uses mostly ecu direct values.

Ted, PM me your email address. I'll send you a copy.

Could you send me the latest airboy ss? I have the one with the accel plotter, but I don't think it saves curves.

Thanks.

[amason]

Me too! Also with an offer to host it. I'd give you an ftp account so you could upload updates...

(all I'd want in return would be a few tuning hints!)

Andy

[mickeyd2005]

I've asked Airboy to send it to ride5000 so that he can add it to the temporary repository of enginuity information.

http://www.ken-gilbert.com/wrx/enginuity/

[ejoy1220]

Could anyone give me a list of say 10 parameters that I want to log?

(I'm using EcuExplorer and Digital Dash, which is awesome btw)

[NSFW]

Air/Fuel Correction #1
Manifold Rel. Pressure (Corrected)
Boost Error
Mass Air Flow
Engine Load
IAM
Ignition Timing
Ignition Timing Advance
Feedback Knock Correction
Engine Speed
Throttle Opening Angle

[mickeyd2005]

Airboy has posted the latest version here

http://www.ken-gilbert.com/wrx/enginuity/

It is called Log file Interpolation ver3p6h public.zip. This is newer than the one that I emailed out so download the newer version.

The AFR interpolate sheet now includes MAFv. This makes it just that much easier to scale your MAF. Keep in mind that when d(MAF)/d(time) is large, you should be aware that the data may be inaccurate.

For scaling MAF below 60 g/s, I use steady state conditions in closed loop. I then use a statistical spreadsheet to filter the data.
For scaling MAF above 175 g/s, I use a 3rd gear WOT log starting at 2000 rpm. The data above 175 g/s is relatively steady.
For scaling MAF between 60 and 175 g/s, I use a 3rd gear partial throttle log starting at 2000 rpm.

[LittleBlueGT]

Airboy has posted the latest version here

http://www.ken-gilbert.com/wrx/enginuity/

It is called Log file Interpolation ver3p6h public.zip. This is newer than the one that I emailed out so download the newer version.

The AFR interpolate sheet now includes MAFv. This makes it just that much easier to scale your MAF. Keep in mind that when d(MAF)/d(time) is large, you should be aware that the data may be inaccurate.

For scaling MAF below 60 g/s, I use steady state conditions in closed loop. I then use a statistical spreadsheet to filter the data.
For scaling MAF above 175 g/s, I use a 3rd gear WOT log starting at 2000 rpm. The data above 175 g/s is relatively steady.
For scaling MAF between 60 and 175 g/s, I use a 3rd gear partial throttle log starting at 2000 rpm.

Thanks tuning guru!

FWIW I found that tuning MAF above 60 g/s can be more accurately and easily be done using 5th gear pulls if you live in the middle of nowhere.

[mickeyd2005]

Airboy has published a new spreadsheet.

http://www.romraider.com/forum/topic957.html

It's got a lot of new features like total timing and load rescaler.

I haven't tuned a CANBUS car yet but looking at the new roms, the boost control and throttle mapping go hand in hand now so it makes it more difficult to tune partial throttle. Airboy added some mapping routines to make it easier to visualize.

Unfortunately, for the LGT 5EAT guys, the road dyno for autos hasn't been incorporated. There's still a beta version towards the end of that thread that includes it so you'll have to use that one.

[Infamous1]

Thanks for the update.