Using Datazap, Excel, and V-Dyno to Evaluate Before/After Data
Grab a snack, this will take a bit. For those that are really interested in learning an overall process, youíll want to open the links, download the logs, and try to follow along. This is all based on Cobb information, and I'm breezing over a lot of info simply due to brevity. As stated multiple times throughout, this is only an example of the overall process to show how you can use the different programs in conjunction with each other when datamining.
Before we get started, here is the link to Cobbís monitor list (as of this writing)- COBBS MONITOR LIST. Become familiar with your monitors and what is/isnít applicable to our platform, and try to think about how each monitor is used in and relates to the overall system as a whole. With enough datalog evaluation youíll start to be able to have a deeper understanding of not only what is logged but how that particular monitor is related to other monitors and the engine as an entire system. Take note of the fact that some of the parameters are inferred and therefor cannot be fully relied on as absolute data, but their presence can still be used and referred to when making general conclusions.
Thereís no way that I can walk you through every-single possible comparison scenario in a reasonable amount of time. For this thread Iím going to use one single comparison that Iíve gathered from some of the current personal testing that Iím doing, but the overall process is the same for pretty much all others-what changes is the data logged, how you log it, and how you choose to view it. You can (and will) compare two logs a multitude of ways depending on what it is youíre looking for or evaluating.
For the sake of this thread weíre assuming that you already know how to get consistent datalog information (DATALOGGING 101) and already know the quality of your fuel (YOU'RE USING INFERIOR FUEL). You can pull up the logs Iím using here (open them up in separate tabs on your Ďputer)-
87 Octane Testing, 87E10
87 Octane Testing, 93E0
I think you can download the actual Excel files if you scroll down below the chart and spreadsheet, and if not feel free to PM me if you want to use them to practice pruning or whatever. I use Excel for the majority of my actual data comparisons and I only use Datazap as a means to quickly view an overall relationship or to chart a pretty line for posting on here. Logs 2 and 4 are both heading NW and 3 is SE on 87 octane, then 5 and 7 are heading NW with 6 being SE on 93 octane. All six runs were done on the Cobb 93 octane Stage 3 OTS tune.
The first step in any comparison that I do is upload it to Datazap for an overview. I do this to make sure that my data is clean and get an overall picture of whatever data Iím going to be comparing, which is going to be ignition corrections and timing for this example. I will verify that no related parameters are too far from each other (coolant/oil temps, etc) and generally make sure that they are all Ďsimilarí within the before and also within the after, and then as a whole. All this is is a check on overall consistency.
If everything looks close, the next step that I do in most performance evaluations is prune the log to only show me what is happening at WOT. To do that, go up to the Trim button (towards the upper right above the chart), select Accelerator Pedal Position % and move the left slider all the way to the right (100%). This is going to prune the chart to only show what was logged when my foot was on the floor. I also remove the Boost Pressure monitor because itís not something weíll be looking at in this instance.
Now that that is done, open up all four ignition corrections in both charts. What you should notice once you look at it long enough is that even though both show mostly positive corrections with some reductions, chart 5 is generally more consistent, has fewer reductions, and shows mostly higher corrections. This is to be expected because the effective octane is higher, but here you can visually see it.
Next weíll add the Ign Timing Cy 1 monitor to the chart. That monitor is showing the timing for cylinder 1 after corrections, and if you go to 5992 rpm on chart 2 youíll see that that cylinder in my car had 10.3 degrees of total timing advance at that RPM. Moving over to chart 5 and 5997 rpm youíll see that that cylinder is now achieving 15.13 degrees of advance. Again, not anything that we didnít expect, but you can now visually confirm the difference.
Now for manually gauging performance with Excel and V-Dyno. If you're not familiar/comfortable with Excel then you will want to open it up and follow each step to see what's being done.
In the most concise way of describing it, you’re going to remove any data not applicable to what you want to evaluate, simply for making it easier to look at. I always keep a ‘master’ datalog with all of the logged information and save pruned logs if I’m that interested. Either way, here is what we’ll do with the Datazap logs for this, and this is also where we’ll introduce V-Dyno (though you can technically use it at any time). We’re going to swap to logs 3 and 6 (both SE runs) for this because I literally just lost my log 5 while trying to speed through this write-up. Oops.
1. The first thing I will do in this situation is prune out all row data that is not 100 on the Accelerator Pedal Position column. While this is not necessary for V-Dyno, we’re doing it now to show the manual process for Excel. This means that on datalog 3 we’re fully deleting rows 2-5 first, then Rows 60 and up, leaving you with 59 rows of data. On datalog 6 we’re going to delete rows 2-6, then rows 59 and up.
2. Now that that is done, we’ll prune out all columns that are not Time, Accel %, or RPM. Again, this will change based on what you’re evaluating, this is strictly for this exercise. If you were comparing against another platform you would want to log and keep Vehicle Speed. For now, we’re going to delete columns C through I, and then from D on up in both datalogs.
3. Now that we have three similar columns of data, we’re going to find the lowest RPM that is relatively comparable (not exact) between the two, and prune out all data below that number. Because datalog 3 doesn’t have anything in the 2100 rpm range, the only thing we’re deleting in this step is row 2 from datalog 6.
4. Now that the low end of the RPM range is basically equalized, we’ll go look at the top end and see what we have. In this case both logs end at basically the same RPM so there’s nothing to do, but if one log continued further on then we would prune out all of the rows that were higher than the highest of the low log.
5. In A60 of log 3, you’re going to type in the formula ‘=A59-A2’ (without the apostrophes) and hit enter. That should give you the solution of 8.554 seconds. In C60 you’ll enter the formula ‘=C59-C2’, which should give you the solution of 3909 RPM. In short, it took 8.554 seconds to make the 3909 RPM sweep. Dividing the number of RPM by the time it took to sweep that range gives us 456.98, which is the average RPM increase per second (acceleration). This is only one formula, and if you're familiar with Excel then you understand that you can create many and use them accordingly-depending on what data you're evaluating.
6. Do step 5 to datalog 6. The formula will be slightly different due to having two less rows of data, but you’re subtracting your lowest data (Row 2) from your highest data (Row 57 in this case) and the formula must be in the same format of ‘=(highest Row)-(lowest row)’. If you’ve done it correctly then you should know that it took 8.26 seconds to make a 3899 RPM sweep, with an average of 472.03 RPM increase per second.
Ok, we’ve done the manual conversion with Excel and can now verify a performance increase based on the RPM per second (RPM/s). Lets take this to V-Dyno and compare data from 3 and 6. If you go plug those two datalogs into V-Dyno you’ll notice that the curve is slightly different, and outside of the last 1k rpm you’ll see that the 87 octane log was producing more power. What gives if our Excel math just said that there’s an increase?
This is where V-Dyno can come in handy. First off, a 15 RPM/s over roughly 8 seconds is not a lot. Secondly, there is additional information that must be taken into account in order to understand how charts are comparable and what previous or additional variables exist. In this instance, the only thing that changed between the charts was fuel, and that’s a pretty key factor because of a reason different than you’ll think. Simply put, my car weighed more on the second half of the test than it did on the first half.
Gasoline weighs approximately 6 pounds per gallon, and I added 10.5 gallons prior to the second test. Because V-Dyno has separate areas to add/subtract weight, we’ll do this now. I use 2740 as the base curb weight, which would include fuel. Because the first half of the runs were done near empty, we can now remove 63 lbs from that curb weight. Add whatever you would like to the occupant weight so long as both stay the same and you’ll see that along the entire curve the vehicle was making basically the same power except for in the higher RPM range. If you look at the ignition corrections and timing for those two charts in Datazap, you’ll see a very distinct correlation between the moment that the 87 octane can no longer effectively add timing, right around 4600 RPM.
Hopefully this gives a general idea of how you can utilize Datazap, Excel, and V-Dyno in conjunction with each other to be able to record and analyze/compare data. It’s time consuming, but a fun exercise for those that like data-mining and effective (not exact) for doing before/after comparisons-whether that’s before/after part installs, fuel testing, or whatever else your heart desires to know. While it is a bit intimidating initially, hopefully it helps ease you into the overall process and the more you use it the more you’ll be start identifying relationships between different parameters. Remember to try and be as consistent as possible when logging and you should have no issue pulling out the necessary information to answer the questions that you may have. All of these programs have their own place individually, but using all three together will give you a much more complete idea of what is going on.
|All times are GMT -5. The time now is 02:07 PM.|
Powered by vBulletin® Version 3.8.8
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.1
vBulletin Security provided by vBSecurity v2.2.2 (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.