Evaluating Modifications Using Logged Data, One Nerd’s Perspective
Thread Starter
Senior Member
Joined: Feb 2009
Posts: 329
Likes: 21
12C P85DL
Evaluating Modifications Using Logged Data, One Nerd’s Perspective
Let’s face it, I’m a nerd. I like Italian cars for the sound and German cars for the technology. With respect to the nerd content in German cars, I’m quickly learning that Mercedes have the thickest glasses and the biggest pocket protectors. They have done a very nice job with the ME97 fuel control which is shipped in most new Mercedes.
Back to the point of this post-
I am not a tuner but I have done tuning and I build tools to aid the tuning process. One of the most valuable tools for my past efforts has been data logging. In the past, this involved grafting data acquisition to a car to evaluate its performance and any performance gains from modifications. These days there is so much available directly from the ecu that you can learn a lot from simply recording values the ecu is already looking at. This approach is very helpful from the tuning standpoint in that you base your changes on the exact same data the ecu is using to make it decisions (as you got the data from the ecu in the first place).
I would evaluate a modification like an ecu reflash by loading the car with and without that modification while recording data. You can load the car on a dyno or simply rip down the street. The factory ecu can provide a fantastic amount of data (the factory tools have access to all these values as well as some tuner tools). Doing back to back runs lets you look at things like throttle opening, calculated torque, engine load, ignition advance, knock retard and the like. You can see straight away if a modification managed more calculated torque or perceived load. You can see if the tuner added too much timing causing undo and dangerous knock. You can check mixture to make sure the car is not running too rich or lean and injector duty cycle to see if you are running out of the ability to provide fuel to the car. The list goes on and on..
I also find it very helpful to look at data from track events and ¼ mile runs.
I did this post as I have seen a lot of posts speculating on the value of this modification or that modification. Perhaps this approach will aid in evaluating changes and bring some of the process out into the light of day.
Back to the point of this post-
I am not a tuner but I have done tuning and I build tools to aid the tuning process. One of the most valuable tools for my past efforts has been data logging. In the past, this involved grafting data acquisition to a car to evaluate its performance and any performance gains from modifications. These days there is so much available directly from the ecu that you can learn a lot from simply recording values the ecu is already looking at. This approach is very helpful from the tuning standpoint in that you base your changes on the exact same data the ecu is using to make it decisions (as you got the data from the ecu in the first place).
I would evaluate a modification like an ecu reflash by loading the car with and without that modification while recording data. You can load the car on a dyno or simply rip down the street. The factory ecu can provide a fantastic amount of data (the factory tools have access to all these values as well as some tuner tools). Doing back to back runs lets you look at things like throttle opening, calculated torque, engine load, ignition advance, knock retard and the like. You can see straight away if a modification managed more calculated torque or perceived load. You can see if the tuner added too much timing causing undo and dangerous knock. You can check mixture to make sure the car is not running too rich or lean and injector duty cycle to see if you are running out of the ability to provide fuel to the car. The list goes on and on..
I also find it very helpful to look at data from track events and ¼ mile runs.
I did this post as I have seen a lot of posts speculating on the value of this modification or that modification. Perhaps this approach will aid in evaluating changes and bring some of the process out into the light of day.
Last edited by lolachampcar; Feb 12, 2009 at 08:24 AM. Reason: my grammer sucks
Thread Starter
Senior Member
Joined: Feb 2009
Posts: 329
Likes: 21
12C P85DL
Also, I got a request back from a tuner that we take the ECU (ME97) reported torque and calculate a HP number (Torque X RPM/5252 or something like that). I've asked the application programmer to add that as a calculated field for viewing. It will be interesting to see how it stacks up against estimated crankshaft horsepower from dyno runs.
Former Vendor of MBWorld
Joined: Jan 2009
Posts: 600
Likes: 0
From: CA working like a slave
C63 AMG Arctic w/ Stage 1 Tune, DoubleXX pipe, MCD-631X coming soon
I have had a datlogging solution for the COMMAND system since last year (I have my vids posted on youtube as well), no one is really to interested for some reason. I have found some VERY interesting things evaluating hardware and software changes.
Trending Topics
MB World Stories
The Best of Mercedes & AMG
Manual Mercedes? 6 Times Sindelfingen Let Drivers Have All The Fun
Verdad Gallardo
Mercedes SLR McLaren 722 S Is Extremely Rare Example Modified by McLaren
Verdad Gallardo
8 Classic Boxy Mercedes Designs That Have Aged Like Fine Wine
Verdad Gallardo
Flawlessly Restored Mercedes 190E Evo II Heads to Auction
Verdad Gallardo
Electric Mercedes C-Class Unveiled: 11 Things You Need to Know
Verdad Gallardo
Mercedes EQS Gets A Major Update: Everything You Need to Know
Verdad Gallardo
5 Underrated Mercedes-Benz Models That Don't Get the Love They Deserve
Verdad Gallardo
Mercedes 300D Has Pushed Well Past 1 Million Miles and It Ain't Stopping
Verdad Gallardo
10 Most Reliable Mercedes-Benz Models You Can Buy Used
Verdad Gallardo
Thread Starter
Senior Member
Joined: Feb 2009
Posts: 329
Likes: 21
12C P85DL
I do not sell products and am not a registered vendor on this site. I post to share technical information and occasionally look for help on projects. I recently migrated my ME97 data logger (and programmer) from using a proprietary USB driver to working simply as a Mass Storage Class device or, better known as, a memory stick. I have been looking for a local car to do some testing. If you are interested in buying a logger, I will contact a reseller registered as a vendor on this forum and ask them to PM you.
As for the data rates, the ME97 does ten samples a second which I have found useful. I did an ME2.7/2.8 logger and that was way too slow to be of much value and the project was scrapped.
The data loggers have not been terribly popular with tuners. They are helpful when doing development work but no tuner seems interested in providing this capability to the end customer. Given the insight loggers provide, I guess I can understand this position.
For my part, I would love a situation where several people get together to evaluate the performance of different calibrations (tunes). It would be interesting to get three or four different calibrations from different tuners (all paid for and the legal property of a car owner attending) and running them side by side on one or more cars. The ME97 programmer takes about thirty seconds to change calibration on a car so it would be a simple matter to switch, allow the adaptations a little time to recover then log the result. The ME97 provides torque as one of its outputs which, when combined with mass airflow, load and some other values, will provided a good yard stick. You do not really even need a dyno for the process although it is always neat to get HP numbers from something other than the ECU.
Just musing….
MBWorld Fanatic!
Joined: Jun 2007
Posts: 2,949
Likes: 0
From: Philadelphia, PA
2008 A8L, 2002 996TT X50, 2009 X5
For my part, I would love a situation where several people get together to evaluate the performance of different calibrations (tunes). It would be interesting to get three or four different calibrations from different tuners (all paid for and the legal property of a car owner attending) and running them side by side on one or more cars. The ME97 programmer takes about thirty seconds to change calibration on a car so it would be a simple matter to switch, allow the adaptations a little time to recover then log the result. The ME97 provides torque as one of its outputs which, when combined with mass airflow, load and some other values, will provided a good yard stick. You do not really even need a dyno for the process although it is always neat to get HP numbers from something other than the ECU.
Just musing….
Just musing….
I find this stuff extremely interesting, but honestly programming is completely foreign to me. I wanted to comment on your quote above - "all paid for and the legal property of the car owner attending." Here's what I don't know - when you purchase a tune, are you purchasing the basic utility and efficacy of the tune but the tuner reserves the legal rights to the proprietary code, OR, are you purchasing the utility of the tune and the rights to the code?
Datalogging is something that I recommend anyone do if they're modding, but datalogging a few parameters like AFR, timing, load, airflow, etc, is one thing, having full read/write access to files and maps on the ECU is a step beyond that. IMHO.
Think of it this way - you go to the store, and buy a copy of Microsoft Office. Do you (legally) have the rights to the code, or just the use of the program? Same can be said for many industries - for example, if you purchase a cake, should you also have access to the recipe? I don't know the answers to these, but it does make for interesting discussion.
Thread Starter
Senior Member
Joined: Feb 2009
Posts: 329
Likes: 21
12C P85DL
I've been told by those in this industry far longer than me that calibration data is not protected IP. I believe the major OEMs have tried to shut down tuning and were unable to do so for the above reason.
As for what you own and do not own, I tend to believe that a buyer purchases the ability to use a particular calibration data set on a car. I would not feel comfortable duplicating that work and reselling to others.
With respect to content, I view calibration data the same way I would view jets and emulsion tubes on a correctly calibrated big block Chevy with carbs. There is nothing to stop the buyer from pulling the tops and seeing what his engine builder did. The builder can file off the numbers to slow the process but the owner still has every right to look because he/she owns it. Unless the owner signs something saying he/she will not look, I would think they have a right to look.
I say the very same thing to tuners that ask me if their tunes can be read or copied. Once you ship it, it is out there for all to see. This is one of the reasons I do not tune cars. The other is that it would put me in direct competition with those that distribute my designs.
All the above is just my two cents worth and purely my opinion.
As for what you own and do not own, I tend to believe that a buyer purchases the ability to use a particular calibration data set on a car. I would not feel comfortable duplicating that work and reselling to others.
With respect to content, I view calibration data the same way I would view jets and emulsion tubes on a correctly calibrated big block Chevy with carbs. There is nothing to stop the buyer from pulling the tops and seeing what his engine builder did. The builder can file off the numbers to slow the process but the owner still has every right to look because he/she owns it. Unless the owner signs something saying he/she will not look, I would think they have a right to look.
I say the very same thing to tuners that ask me if their tunes can be read or copied. Once you ship it, it is out there for all to see. This is one of the reasons I do not tune cars. The other is that it would put me in direct competition with those that distribute my designs.
All the above is just my two cents worth and purely my opinion.
