Tuning the AEM Infinity, check out this helpful tool...
 

There seems to be a reasonable interested in the AEM Infinity Stand Alone FI system here on Pelican so figured it was worth posting this.

A great friend of mine Trevor and I cut our teeth on the AEM Infinity system and have found it to be solid platform from both a hardware and software perspective.

One of the things that can a bit time consuming is logging the VE error and applying it to the VE table after a session. This is particularly true when trying to tune in part throttle fueling. The "NewVE" virtual channel works well and goes a long way to help, but you still have to step through the log and make changes one cell at a time.

From using HPTuners, I found the ability to build a table that duplicates the VE load and RPM axis/cells and have it generate a VE error table very useful. This table can then be applied to the existing VE table and correct the current error.. It's somewhat of a manually implemented autotune.

I wanted to figure out how to do the same thing with the AEM Infinity tuning software, but after several hours experimenting, it became apparent the Excel skills to pull it off were not in my toolbox.

When talking with Trevor about the idea, he volunteered to try his hand at it... Little did I know, he is a bit of a Excel guru!

After a few days of work building the formulas and tables etc, he had something for us to try… and it worked! Since then, he’s refined the formulas, added instructions, and some additional features. In addition, and we have tested it fairly well.

How it works...

Drive and log the required channels. Download your log into the AEMData software, export just the channels that are needed to CSV, then copy and paste that data into the spreadsheet.

A copy of the CSV data as its exported from AEMData.
http://lightdance.zenfolio.com/img/s...60242823-5.jpg

The first page of the spreadsheet with instructions.
http://lightdance.zenfolio.com/img/s...59262617-5.jpg

Same data from the export pasted into the spreadsheet.
http://lightdance.zenfolio.com/img/s...59262615-5.jpg

The modified cells populated with the “NewVE” data. Before your data is useable, you need to specify whether your load is throttle or MAP based. You also need to make sure your breakpoints for both load and RPM matching your VE table exactly.
http://lightdance.zenfolio.com/img/s...59263000-5.jpg

Simply copy the data from this table, and paste into your Actual VE table like so.
http://lightdance.zenfolio.com/img/s...59262632-5.jpg

At this point, you want to look at the data and see if it all makes sense. You can even scrub through the log and verify the logs NewVE is matching what you copied in. Wash, rinse, repeat…

Does anyone have an interest in using this? If so, we can throw it up on a google drive for people to check out and use.

Let us know your thoughts.

David

Tuning the AEM Infinity, check out this helpful tool...
 

I am vey interested. I have not gotten to this point yet in my project, but have been researching and reading about the tuning process and this tool fills a gap. I expect to be at a point to start tuning before the end of Jan and would be happy to be a Beta partner.

Also interested!

My ITB project is a bit off into the future but would very much appreciate adding your spreadsheet to my library. Thanks for sharing!

I'd have a real interest in this!

I'm interested. Thanks David

I am hoping to go ITB/EFI and I've been looking at the AEM. This looks like a fantastic idea!

nice David.............

regards,
al

Ok guys, great to hear there is an interest in this! Try this link out to download the file. I will also copy a CSV template that you can use to make sure you export the correct channels to CSV. I'll get those up tonight or tomorrow morning.

https://drive.google.com/open?id=0B0GQlfZHOrn_bUZKNlF6VzVUVVU

Also, you need to make sure you are logging the needed channels. I would recommend that you know it is working from logging to importing into the spread sheet and getting back into the Infinity tuning software. Nothing more frustrating than going out and logging for an hour and find out you missed a channel required.

We may run into some areas that need some tweaking. After you download and take a look, please ask your questions in this thread so everyone will benefit from the dialog.

Looking forward to hearing what everyone thinks of it once its being used.

David

Once you have logged your proper channels, you open your log file in AEM Data.

File > Export CSV. You should see a screen like this

Use these if you are using MAP for load
http://lightdance.zenfolio.com/img/s...62457791-5.jpg

Use these if you use throttle for load
http://lightdance.zenfolio.com/img/s...62457770-5.jpg

You want to uncheck select all/none so nothing is selected. You can verify by check "show exported only"

You can either find and check the channels needed in the image above, or load the CSV template that I will post in the same directory that was posted earlier.

What I need to figure out, is how we deal with the 2nd lambda channel which Im guessing most of you will not be using.

I'll check on this and let you know.

David

Got it, many thanks!!!

Unless I'm missing something here, and I've been pretty heavily involved with reading pro tuner websites, I really appreciate the VEAnalyze Live! and MegaLogViewer from TunerStudio/MegaSquirt.

It does these functions for you.

Tippy, do those tools work with AEM Infinity tuning software ECUs? Had no idea if so...

Phil Tobin, the creator of TunerStudio has mentioned years ago, that if there was enough interest, he'd develop for any ECU platform.

Nicely done. This will certainly help the AEM community.

MegalogViewer should work with any csv file type to import the data. I use a modified CSV file to import GPS and accelerometer data along with the engine output from the Megasquirt ECU. The challenge will be that the imported tune file will have different memory allocation locations for direct VE value and target AFR. Basically, the MS software will be looking in the wrong spot for the tune information from AEM.

The system has a (simplified discussion) function to read the instantaneous AFR, compare this to the target AFR, determine if rich or lean, calculate how much rich or lean (ratio of AFR actual to AFR target) then read the amount of fuel delivered, and scale the existing fuel delivery by the magnitude of offset from target AFR.

In the Megasquirt community there is a real-time live version of this script and also a secondary, post-process version. I like the real time driving version when roughing in a VE table and then the post-process version to fine tune.

Phil Tobin (Author of MLV and TS software) can be found at efianalytics.com. He is easy to communicate with. Since you have the math done for the spreadsheet algorithm you might be closer than you think to having a version that works with AEM. It is worth talking to him about.

BTW, some of the analysis tools are why I prefer the MS systems.

Thanks guys... all credit goes to my friend Trevor for this tool!

Im guessing most people will be running a single wide band channel on a 911. Trevor re-spun the XLS so that it will support a single channel and not require any tweaking of your Exported CSV data to make it work.

The new XLS and AEM Data templates are in the shared folder. I moved the folder on my google drive to better organize the data. If you cant access it, please let me know and I will get a new link posted asap.

Interesting about the TS stuff. Ive always been impressed with MS and its community of developers. It may make sense to ping him about it.

David

Nice work David! Thanks for sharing.

I've been using this for the past week with some basic street tuning, and it's been a huge help for a novice. I'm currently using MAP x RPM, but am thinking about switching to TPS. Or... I've been reading up on a blended TPS/MAP x RPM. Any thoughts?


Sent from my iPad using Tapatalk

Great to hear!! Re your question, what kind of Intake system are you running? If stock Intake, MAP is probably going to work fine for you. If you have ITB or big cams, TPS may be better for you.

One thing to keep in mind, if you choose to use TPS mode, the Infinity uses the TPS position for load, but still uses MAP for fueling calcs and baro/altitude changes. This may be why it works so well on this system.

Even though Im running a stock manifold on my Coyote, we chose to move to TPS based load and the car runs great and Trevor and I have no plans of going back to MAP.

David

I have ITBs and have already switched over to TPS for load on the VE table. With three tuning rounds under my belt today, I hope to arrive a a well tuned base MAP by mid week. I'll post more as that progresses.


Sent from my iPad using Tapatalk

3.0 is done
 

I am using TPS mode on the 3.0 i just tuned....works great. This one is done and runs great.



regards,
ahttp://forums.pelicanparts.com/uploa...1485140479.JPG

Bringing this to the top in case it may help another AEM Infinity user.

David

I just downloaded this and looking forward to finding some time to learn and apply.

Anyone have any new info they would like to share since using this?

Thx

Yes... I suck at excel.

This is an impressive tool, but to work well, your car needs to be in a steady state. In other words, you need to filter out all transients. In my opinion, the best tool to do this is Megalogviewer. You can set filters to remove transient conditions. You can even filter out data from when the engine is cold and not appropriate for developing a VE table. It also produces tables which you can copy/paste into the infinity software plus you can do much much more.

When using a usb drive to datalog, does the laptop still need to be inside the car running infinity tuner at the same time?

No need to have the laptop inside to log via usb, but you have to set what parameters you want to log and what is the log start criteria. After you are done running and want to get the data, then you connect the laptop to download the log from the usb to your computer. In other words, you can’t just plug the usb to the laptop and copy the data.

Also, if you log via the laptop, the parameters logged are those for which you show data on any of the tabs. For example. If you don’t have a table, graph or value displayed, and record data but then want to see that value by modifying the layout to include the value, it will show that the data was not recorded.

OK, thanks, I managed to do this.. and recorded a log on my way home from work. I stalled the car twice, so I ended up with 3 small files.
In order to use the spreadsheet, which fields should I be logging?

Have you figured out a way to use Megalogviewer with the AEM Infinity system?

Re tuning on the street using this tool, its actually quite easy to do. Yes, you do need to be smooth, but if you have it setup with a switch to turn the logging on and off, you simply switch it on, drive smoothly in various load ranges and RPMs and then switch it off for each logging session.

We have tuned several cars on the street this way(with no dyno tuning) and have gotten very solid VE maps using it. It makes tuning the Infinity significantly easier than manually doing it. Clearly if you have a load based dyno things become easier with our without the tool.

Do you mean, you can set it up so that a switch in the car, triggers logging to start and stop? Being a PC noob, I kinda wish there was a video tutorial on how to use AEM data, logging with a switch, and then using the spreadsheet....

Yes. You can use Megalogviewer to analyze data from the infinity and really for general data visualization even unrelated to automotive work. It is surprisingly fast handling massive amounts of data. I have no connection to it but was impressed when I used it.

In summary, you collect the data with the infinity software. Then load into AEMdata which allows you to export whichever parameters you recorded as a csv file. Then Megalogviwer can open it and voila, you can analyze your data and get VE tables you can copy/paste. Whether you use MAP or TPS you can still generate a VE. You can even convert from MAP to TPS if you already have one diales in. The ability to create various filters allows you to drive around and automatically remove transients.

Now, if only you could autotune with it like you can with the megasquirt...that would be fantastic.

I'm having trouble using the spreadsheet. Here is the latest log I took. Maybe someone could have a look for me and run the numbers?

https://drive.google.com/open?id=1fEy0BBFlfvRD_ibsPzM5XQNuWJGWwE-J

Mike, I analyzed your data with Megalogviewer and calculated a NewVE based on your data. You will only see values for which there was data. The color code is weighted to number of samples. The light colors have fewer data samples. I plotted TPS vs RPM because that is the strategy you are using. I also matched your values in the X and Y axis (one extraneous column in the RPM can be ignored). I applied two filters: one removes values where there is wall wetting and the other removes values where the lambda is very high, as I see you used fuel cut. Out of a total of 14,044 data points, 7,895 were filtered out. If I used other filters, you lose more data. Here is where it is important to try to hold steady state values whenever possible.

Here is the screen shot. If you can't read it, I can try emailing an Infinity file which you can import.


http://forums.pelicanparts.com/uploa...1548650513.png

This one shows the difference between NewVE and VE to show you how close or how far you are in the values that were analyzed.

http://forums.pelicanparts.com/uploa...1548650632.png

Wow, thank you for going to that trouble. How long does it take? For some reason, I struggle using the software...

I might alter my VE map now, and do some more logging.

When you say, "try to hold steady state values" do you mean to try to drive logging particular cells at a time to?
The other night, I would drive with my throttle around say 30%, and then let it rev right out through the rest of the rpm range, to try to hit all the cells in that row.

No problem. It took me about an hour because I had to understand your data, your y-axis for VE, see what you logged, etc. Now that it is set, probably 15 minutes.

If you take out fuel cut, you can adjust other areas of your VE table.

By holding steady state, what I mean is you want to have slowly varying changes values. You want many measurements at each value of the table to increase the accuracy of the value. The way I do it is go to an expressway with a high speed limit. To log high rpms I use lower gears and accelerate slowly. There are speeds where you can hold the rpm at a certain value and keep it there. I also go to high gear to log higher Throttle values. WOT can be done in high gear until you are driving too fast.

You can stab the throttle to different values but then hold it steady and let the other parameters creep up. If you want you can email me to iterate on your logs.

The values I use are coolant Temp, engine speed, engine speed rate, map, primaryInjDuty, SpkTiming, throttle, throttle rateVE, wall wetting, wall wetting2, lambda and lambda target. You actually need a subset of these bit those are the ones I export.

For you twin pluggers, how much timing at 80-100% open throttle are you dialing in by 2500-3k rpm ish? And at 30-40% cruise? I'm around 31 degrees with light throttle openings and cruise in that rpm range and 24-25 degrees at 80% and above but some tuners have advised they dial in all WOT timing as soon as 2500rpm on a higher compression twin plug motor, some 26-27 degrees.

Well, thanks for reawakening my tuning addiction. I could, and will, lose days inspecting the scatter graphs. My MAP*RPM vs InjDutyCycle is bizarrely shaped. And I thought I had a great tune on.


Sent from my iPad using Tapatalk

I bought Megalogviewer HD... here is a video showing how I used AEM data, and megalog viewer HD to create a new VE table.... am I doing it right??

https://www.youtube.com/edit?o=U&video_id=-ckhtqY9pV8

Anyone here able/willing to walk me through how to create a filter? If it requires being able to write a formula in a separate language, like writing formulas in excel, then I definately need help!