EFI Technology's Data Acquisition - Bits Byte Back

Squiggly lines, distance plots, time plots, frequency plots, histograms, X-Y plots, 3-D plots, constants, math, and many other tools are the way we make sense out of data collected. In the last article we discussed the tools for collecting data, sensors, wiring looms, loggers, and dash displays. In this installment we will discuss what we can do with this data using EFI Technology's analysis software.

One of the most important steps in analyzing data is to be organized. You can have all the data in the world, but if you don't know what track you will be competing on with a specific driver in a certain car, in the end that precious data is worth nothing.

Depending on the how critical your data is, sorting the data into an easy-to-find manner is a good place to start. Typically it's grouped a few different ways (event, track, day, session, data set). For example, when you're looking for the fastest lap from Practice 1 at Sears Point in 2004 during the sixth hour into the race it's not going to take you days to track down the information.

Most data systems will have a setup sheet, outing info, or some type of sheet that will ask for a variety of information. The sheet may contain simple things like track, engine, weather, vehicle, or chassis setup (see the image of the setup sheet).

It's also highly recommended to collaborate with your engineer to do a session log. A session log will typically highlight the standard setup variables of the vehicle and the goals of the session. By keeping a session log you will be able to see that when the car came in on lap 48, and the front camber increased by negative 0.5 degrees, and the rear wing gurney went from half an inch to five-eighths of an inch,so there was a definite change. When analyzing your data later, you will know when the key changes occurred and will be able to quickly get to the data to figure out why the car went faster.

So now we have collected some data, organized it, and we are ready to start going faster, right? Not quite yet. Let's first cover the simple basics of what we are going to be doing with the data. Most data systems will have the sensor information sorted into channels, these channels will then be sorted into groups such as analog, digital, or onboard math. So what is a channel? Simply stated, a channel is the data that was collected on an assigned input. For example Analog 1 is a channel, but what channel? Typically the channel properties will be modified to give the input some meaning and to apply a calibration to the raw bit value that we talked about in the past article. Channel properties can be color, scaling, units, group, resolutions, and applied filters. A channel properties window and channel data window is displayed in image 2.

OK, we have channels, we have set their scaling, colors, applied filters as necessary, and made the calibrations to convert bits into engineering values, so now what? We look at it, but how? There are many different ways to look at and interpret the data so we are going to start off with the two most basics: plot time and distance.

Even the most basic analysis packages will include a simple time plot. You will have time as your X axis and your channel values as the Y. Typically there will be a "tiled mode," that will automatically evenly display the channels, or you may be able to set your own scaling. There will also be the standard overlay mode which will have all of the channels on one plot. Most plots should include basic information about the channel selected-minimum values, maximum values, average, standard deviation, filter, and the rate the channel was logged at. This will save a large amount of time looking for the basic questions. An example would be if you're looking at shock travel and wanted to see if the shock is bottoming. Depending on how you've zeroed your sensors, you can simply look at the minimum value and you know that below X value the shock has bottomed, or hit the packers.