- I hope you like Excel!


How and Why you should data log.

3y ago

This is probably the 4th time I've sat down to write this article and it's always been all over the place. Not dense enough, yet too much information. There is simply so much to data logging, I can't cover it all in one article. Don't be intimidated if you're new to this, I'm going to keep things super basic to start! Also, to keep content flowing (and articles a reasonable length), I'll be releasing "Datalogging" as a series of articles, rather than one crazy long one. Today is the first step: convincing you that yes -- data logging is worth it, and, showing you how to actually do data log. For most of you, this article is probably not needed, but I wanted to write it anyway for the sake of completion. Part 2 will be about a few basic safety parameters while part 3 will start to look more into power related things. After that, I hope to cover some tools you can use to analyze logs, other tricks, and maybe even discuss custom tuning a bit. Hope you're excited, because I am!

Why Datalog?

First off, I'm a huge advocate of spreading knowledge about cars (hell it's why I write these blog posts). In simplest terms, more knowledge is better than less knowledge. It makes you feel good, gives you a since of pride about knowing your car better, and perhaps most importantly, can save you a lot of money. On that note, let me convince you with my own personal experience.

When I moved to Salt Lake City from Oregon, I went up about 4,000 feet in altitude, and naturally, my car felt slower. Even after going stage 2 (downpipe + tune), the car still didn't feel as good as stage 1 (tune only) at sea level. It didn't make sense to me. I data logged, but I didn't really understand logs at the time. I just kept seeing that the car wasn't making boost targets, but wasn't sure why. I kept logging and hoping for some kind of "light bulb to go off". Well, that light bulb went off a little too late, and I found a cute little turbine wheel in my resonator, most ironically, while I was data logging! All of which could have been avoided if I knew how to log properly, and how to analyze the data. So it's not just logging, it's logging well, and knowing how to read the logs!

What You'll Need

The focus of today is simply how to capture data, nothing more nothing less. So let's start with what you'll need.

1. A data logging device. Most of my articles will focus on using Eurodyne flash, because that's what I use, not because it's the best. If you have a JB4 or Cobb, life should be easier thanks to a great user interface. VCDS on the hand, is probably the most difficult to log with, and I may write a separate article on that later. Most of what we talk about (especially today) will be independent of a logging device, so don't panic.

Have you found it yet?

Have you found it yet?

2. Your own Mexico. Seriously though, find a nice open road, preferably flat where you can do at least 3rd gear pulls, or even better, 4th. The longer (in seconds) we are at WOT without shifting, the more data points we get. So theoretically, logging in 5th would be even better. The road being level isn't imperative, but it will help later if you want to analyze power. Also, be sure it's on dry pavement. Spinning will result in "bad" logs, and I can't recommend doing 100+ on wet pavement, even if it is Mexico. Remember, we're essentially trying to simulate a dyno pull.

3. I highly recommend having a friend with you, especially if you're using a computer-based logger like Eurodyne's or VCDS. Reaching over to the passengers seat to start and stop your data logger while driving isn't ideal. Alternatively, if you keep it running for minutes at a time you get a really big file that's hard to work with. Again, you JB4/Cobb users have it easy here!

Selecting Variables

You've got your logger hooked up, your friend with you. Now it's time to set up the logger. On a JB4 or Cobb, the variables should be preselected and the whole process is pretty straight forward. However, if you're using Eurodyne it's a bit less straight forward. You'll need to open the app, select "diagnostics and data logging". Then select "get controller info", this basically pairs the device with your car. Finally, you'll open up the logger, and for starters, select all variables and begin the process. For VCDS people -- you'll have to wait for another article, sorry!

Your First Log

Why are you data logging? Does the car feel good? Are you confident ripping it anytime, or are you trying to diagonse an issue. Data logging isn't only useful at WOT. In fact, sometimes we even need to data log just steady cruising. If it's your first time logging, I recommend doing a mild pull to start.

Get the car to about 2000 RPM in your logging gear (3rd or 4th) and maintain speed for a moment. Start the data log. Ease into the throttle, and do a half throttle pull to about 5,000 RPM. Lift or shift, and stop the log. Congrats, your first data log!

Alright, let's walk the "why" behind this procedure. Firstly, we want to start at a low RPM for a few reasons. One, more data. But also, we want to give the turbo time to spool. Factory IS12, IS20, and IS38 turbos are great that you can get near full boost by ~3000 RPM, but, not instantly. Even though the part of the log we're most interested in 3500+ RPM, we want it to be a full boost to by 3500. To do that, we need to get on it well before then. As for half throttle, it puts less strain on components (in case there is something up), pretty self explanatory there. Lastly, capping the RPM also helps safety, because if we have a tune that's too aggressive, up top is where we're most likely to overspin the turbo *cough* Eurodyne at altitude *cough*.

Before you go full send...

Once you have this log, you're going to want to check out two things before you do a "full send". Somehow, your log should end up as a ".csv" file (comma separated value). For Eurodyne users this ends up at "C:/Windows/Eurodyne" by default. While there's lots of awesome viewers out there, I think the best way to learn is to dive in head first, so open it up with Excel. If you don't have Excel, you can also open it up using Google Sheets. Either way, you should end up with some monstrosity that looks something like the cover photo.

For the scope of this article, I'm only going to explain the bare minimum details. The first column is the header. It tells us what data is each column. Each row is a data point, collected at a particular time stamp.

Example log.

Example log.

We're going to look at timing correction and boost for starters. Here's the variables you'll need:

DriveTribe likes to cut pictures off, so I had to add extra whitespace, sorry.

DriveTribe likes to cut pictures off, so I had to add extra whitespace, sorry.

TIMING CORRECTION - Here we want to see values an absolute value less than 4 before doing a full pull. (Absolute value meaning no less than -4, or no more than positive 4. If you're using Eurodyne, all values should be negative, but, JB4 logs display correction as a positive number). Also, a single cell of "-4" isn't something to freak over, it's when all cylinders are 3-4+ degrees of correction where you should start to get concerned. This choice of "4" is somewhat arbitrary, but its a more detailed conversation I'm leaving that for the next article. The point is, the closer to 0, the safer, but don't freak out over non-zero values either: even factory tunes will often have timing correction.

BOOST - For starters, we want to make sure that we're getting close to hitting boost targets. If were significantly off, we know something is up, and we probably shouldn't push the car any harder. Essentially we want to compare "target" boost to "actual" boost. In the first few data points we expect boost to be below targets -- it takes time to build boost. But after 1000 RPM of acceleration, we should be close. So if you started accelerating at 2500, you should look between 3500-5000.

Essentially, we want to find out how far off the target boost is from the actual boost. Feel free to just eyeball this, as its easy to see. If boost is off by more than 2 psi in the midrange or top end, something's probably up. However, if you'd like to be more precise about it, here's how:

Simplified example.

Simplified example.

To calculate boost actual vs. target more precisely, we'll use percentage difference: (Target - Actual)/Target x 100%. (Eurodyne users, you'll want to convert both units to PSI for this). I've done this in the figure above, and also used a heat map to make it easier to read. As you can see, in the first few frames, boost is ramping up, so the difference is huge (86-30%). But once were in a steady state, the boost remains very close (within 5%) from targets. This is a good sign that the car is healthy enough to do a full WOT pull.

Full Pull

Depending on your comfort and confidence, feel free to iterate from doing partial loads, analyzing and working up to doing a full pull. But if you're pretty confident in everything, do a full send. For data logging, I highly recommend revving it all the way to at least 6000 RPM on an IS12 or IS20, and 6500+ RPM on an IS38 (or larger). We want to paint a full picture of what the car is doing. Now you've got some solid data, so what does it mean? Stay tuned!

Join In

Comments (3)

  • Excellent info. Thank you! I've been looking everywhere for more info regarding eurodyne datalogging and tuning.

      9 months ago
  • Thank you for writing this up! I have been trying to learn and have not found a ton of info on it. How do you convert to psi from a Eurodyne file?

      3 years ago
    • Eurodyne plots pressure in absolute mbar.

      Raw Pressure (mbar) * 0.0145038 = raw pressure (psi)

      Then to correct it to "boost": psi - 14.7 = boost

        3 years ago