Factory Five Racing Forum banner

Another AEM Infinity ECU Install on a Coyote

80K views 443 replies 18 participants last post by  TMScrogins 
#1 · (Edited)
Another AEM Infinity ECU Install on a Coyote and...

Well, I got my roadster back on the road after the dyno disaster late last year. Took the opportunity to replace the destroyed stock coyote motor with an Aluminator XS (this proved to add some unintended consequences, I'll describe in a moment). Here's a link to that saga for those interested... http://www.ffcars.com/forums/45-for...ter-builds/469954-cold-air-intake-coyote.html

After having a new tune built for the Aluminator XS and logging dyno time I began driving the heck out the car over the spring and early summer months. The tune just wasn't quite right down at low RPM, especially at tip-in and tip-out conditions at no load. Clearly more power with this engine but the driveability characteristics just weren't as gentlemanly as my stock coyote motor. So back to the tuner I went for some more tweaks. We were able to improve it but still not quite right. I then pondered going back again to try and get more tuning done but began to think maybe it's time to start tuning this thing myself as the inconvenience to take it to someone continues to mount. I began looking into SCT's Advantage III Pro Racer tuning software since I was setup with the SCT tuner and tune already.

Enter David Borden... David says he's thinking of considering ditching the stock PCM in favor of AEM's Infinity ECU. Not knowing much about their product, I began my research and ultimately decided to ditch the SCT route and Ford PCM and go full hog. Fast forward and we are both sitting in tuning class down at AEM's headquarters (safety in numbers right?). After the whirlwind 2-day tuning training class and product purchases made, on to the install of the new ECU.

What sold me on the Infinity setup was the fact the hardware and software was manufactured and supported by one company. No locked tables, no mystery areas of the ECU, no reverse engineering. Everything is above board. Coupled with a one stop shop to help the likes of a sorry wannabe tuner like me. Couple this with some of the product features that I was really interested in, like Traction Control, and the user-friendly approach of their tuning software interface with the use of wizards, and I was hooked.

The physical install itself, while involved, went without a hitch. I took the opportunity to install tone rings on the rear IRS axles along with wheel speed sensors front and back. I also had to convert the VSS Analog VR signal from the Tremec TKO to a digital signal for the Vehicle Speed input to the ECU (Thank you Dakota Digital), wired in inputs for a 12-position traction control switch as well as some other inputs for engine protection strategies and some other features. Wired in the fuel pump to be controlled by the Infinity ECU. Done with that and it was time to start tuning.

What is really cool about the Infinity ECU is that AEM provides a base tune file for the stock Coyote motor. You basically start with a conservatively safe tune that will run your engine and can then tweak it from there. My problem was that the Aluminator XS uses larger heads and cams along with larger throttle body and fuel injectors. This would need to be compensated for in the tune. The great guys at AEM took my throttle body (Ford Racing Corbrajet Twin 65mm) and one of my fuel injectors (47 lb.) and characterized them for me on their equipment. This gave me the data I needed to build out the modifications to the base tune file.

It was time to start the engine. At this point David Borden was about 1 week ahead of where I was on mine. He had already successfully started, idled and test drove his without much of any issues. The engine started right up with not a lick of problems. However, there were immediate idling issues. Idle would range from over 2,000 RPM down to 400 RPM. The engine would just begin to oscillate as the idle feeback would work overtime to correct the oscillations. Until I could get the idle tuned properly there would be no driving it. With hours and hours of tuning on just the idle with lot's of phone and email support from AEM (those guys have been absolutely fantastic), I still have yet to get the idle working properly. All the issues have been focused on the drive-by-wire tuning. Particularly, we determined that there was some stiction in the throttle body down at just off idle at around 1-2% throttle position. Ultimately, I changed out the throttle body and went to the Super Cobrajet Monoblade TB. Still unsuccessful although through tuning the PIDs and the DBW Bias tables I was able to get the idle within a range of 600 RPM to 1300 RPM using idle feedback to control the idle. Still not acceptable though.

Both of these throttle bodies flow about 1600 CFM (twin) and 1800 CFM (mono). The stock Coyote TB flows about 900 CFM. Since upgrading to the Aluminator XS with the Cobrajet Intake and TB, I've never had nearly as good of idle characteristics as the Boss IM and Accufab 84.5 mm throttle body (1200 CFM). So my current theory is that the Throttle Body is just too large for the N/A application to reasonably expect low-end driveability. The slightest movement of the throttle plate causes a large run up or run down in engine speed as it takes in or restricts such a mass volume of air.

To test this theory I have ditched the CobraJet intake and throttle body for now in favor of the Boss intake with Accufab throttle body. Just finishing up the install now and should be back to tuning it this weekend. Assuming this fixes the problem I'll hit the road for some data logs and begin working the tuning out for the VE map, ignition, cam timing, and decel. Hopefully be back on the road soon.

Fingers crossed.

Trevor
 

Attachments

See less See more
5
#145 ·
Hi,

Hope u are doing good. I have a question please. Regarding the Dakota digital module, which signal did u use to send to ECU?

I have digital Dakota instruments and it has a speed out that can send 2k or 4K ppm for ECU. So thought maybe I might save some money and not buy the module u have ?

Thanks
Hakeem
 
#146 ·
So you would feed the speed sensor input to the Dakota Digital Speedo gauge. Then use the speed out that you are indicating and feed it into the ECU providing this speed out from the instrument is a digital signal (I would assume that it is). I wouldn't worry too much about the pulse signal rates. The Infinity can calibrate it on the ECU side to match.

I would agree that you do not need a module I have to make this work.
 
#149 ·
How To: Tuning Procedures - Installment #10 (Valet Mode)

The tenth installment - Valet Mode switch

This concludes my "How-To" DIY documents for the Infinity. I hope someone finds this information useful, either now or down the road.

I'm hoping someone forges ahead on some of the other features that I haven't explored and shares with the rest of us. E.g. Boost control (Saul, hint, hint), ITB tuning (Mark G - hint, hint), launch control, shift cut, etc.

Still so much more to learn and explore.

Trevor
 

Attachments

#150 ·
New Tank Installed

Almost back to working on my E-85 tuning. I have pretty much wrapped up the install on my new, custom Boyd fuel tank. It's awesome. Here are a few pics. I'll spend the next couple of weeks finalizing the Gasoline VE tuning with the new injectors (this all assumes the tank fixed my cavitation issues, which I'm sure it has) and then back to a first run with E-85.

Stay tuned...
 

Attachments

#151 ·
Tank works flawlessly. After a 100 mile cruise in the foothills this last weekend with friends tearing up the curvies, zero fuel cavitation issues (pic below of our midway stop at Daffodil Hill). I finally have an upgraded fuel system that is complete and working perfectly. Over the weekend cruise, I pulled 18 different data logs under various driving conditions. This should give me plenty of data to dial in my VE tuning for gasoline while I burn through this tank of gas, and the next tank will be the first run of E-85. I'll post again after my first tank of E-85.

Trevor
 

Attachments

#152 · (Edited)
Great! Looks like a another successful upgrade, good to know.

I had to backtrack a bit and ended up installing mine this last weekend the good news is I can get back to wiring and electrical. My priority now is first start and go cart, my goal is to have her running before the Huntington Beach Cruise In.

So thanks again Trevor for helping me spend my money :wink2:

Saul
 

Attachments

#156 ·
VE Tuning new injectors

Spent a good chunk of time re-working my VE table since installing the new injectors. I really thought the swap wouldn't require much tuning on the VE table. But because I am using Injector Dynamic's characterization tables for offset and flow instead of AEM's provided characterization, the models are different, and since I used AEM's characterization tables on my old injectors, this move from one test/characterization methodology to another probably is the reason.

Anyhow, the VE values really needed to ratchet up on the high load cells and ratchet down on the low load cells (less efficient down low, more efficient up high). So the table extremes are more extreme but the basic graph/curve shape remained pretty close to the same.

Driving around and datalogging does a pretty good job of giving you enough data to tune for volumetric efficiency from about 70 kPa down to 30 kPa. Getting the high load VE values above 70 kPa at higher RPM is extremely difficult and dangerous to do driving around datalogging. Wish I had my own dyno.

I learned a new trick to tune the VE at higher loads at WOT. Pretty much common sense but didn't dawn on me until I read it somewhere over the weekend. One of those duh moments. Anyway here it is. "VE WOT curve should follow the torque curve!" The engine is most efficient when it is building the most torque - duh! Since I had already made a trip to the dyno over a year ago with this engine, I know what the torque curve of the motor looks like. I know where it builds peak torque. At peak torque the Volumetric Efficiency should be at its highest (100%) or thereabouts. My engine builds peak torque at 4500 RPM then dips down slightly and then back up and peaks again at 5500 then trails off. Using the torque data points at 4500 and 5500 from the dyno run as the highest VE values, I then scaled from there the rest of the RPM range for the WOT high load cells.

The attached pics represent the old and new VE tables with only an injector change. Like I said, the curve is more or less the same but you'll notice less green and more yellow and red indicating greater efficiency in those areas. If you look closely at the peak of the new curve, you'll notice that dip in between the two peaks (4500 and 5500 RPM) that match up identically to my torque curve.

One downside to the AEM Infinity is that there is no VE learn tool. Some FI ECU systems on the market today have a VE learn where the computer calculates and populates the table with new VE values as you drive. AEM has hinted that they are working on such a feature but it is not to market yet. They do, however, provide a math calc channel that computes the New VE. So you can manually review data logs and plug in the NEW VE calculated field into the table for each cell where it varies from the existing VE. Very tedious, particularly without a dyno. The tricky part is making sure your Actual Lambda and Target Lambda are in sync and not jumping when taking the NEW VE value. I am going to add another How To - Tuning Procedures document for the VE tuning for those interested.
 

Attachments

#158 ·
Fuel oscillations

Still haven't gotten to my first tank of E-85. I again got side-tracked tuning out another minor nuisance. While tuning the VE table there was one condition that existed where I could feel the engine hunting. Almost a surging on and off. It seemed that fuel was added then cut then added then cut under this condition. So I referred to the datalogs and could see that actual lambda was oscillating while trying to stay on targeted lambda. This would only occur in low-load conditions say around or below 35 kPa (on a slight downhill) with engine speed at a steady state (accelerator being held steady just at tip-in) and the RPM below 3,000. Under higher load conditions I get no oscillations and the datalog looks nice and smooth on lambda to target lambda. See attached pics showing these two opposite scenarios.

At first, I thought maybe my Decel Fuel cut was being activated then deactivated back and forth creating the oscillations. But the data logs showed that was not the issue. The interesting observation here is that with all conditions identical with the exception of load, oscillations would exist or disappear depending solely on load (MAP). This was the hint that my Lambda Feedback PIDs needed to be dialed in. The MAP, RPM, lambda target and VE were all perfect in the data logs. The Lambda FB gains were too high in this RPM range inducing the feedback oscillation, but why only at low load?

Under regular to high load conditions, there is a lot of airflow going through the engine and at a much faster rate. The response time from when a fueling change is made to when you see the fueling change is faster thus resulting in a shorter time requiring a larger feedback gain to try to keep the requested changes in sync. However, under low load conditions, the volume and speed of the airflow is much less and slower thus it adds time. Under these conditions, we need to slow the feedback for the longer time.

Now that I understand the why, I can go about tuning for the solution.

The fix is to slow down the "I" gain sooner (and maybe the "P" gain, but I'll start with "I" gain). The Infinity correlates Lambda FB gain to engine speed creating variable gain parameters, so I can change the gains depending on the RPM. Since the problem exists only below 3,000 RPM, I will slow the "I" gain down below that engine speed, but leave it where it is for higher engine speeds since low load conditions at these speeds is very rare. The tradeoff here is trying to find the right balance. If I slow the gain down too much below 3,000 RPM then I face the opposite problem for high load conditions below 3,000 RPM.

I've made my tuning adjustments. Just need to get the car out for a test drive. Will report back how it went with results.
 

Attachments

#159 · (Edited)
Suggestions to troubleshoot an oil pressure input issue.

Actually prior to the first start I ran the wizard and set the most basic Coyote engine parameters, I also ran the wizard and set the drive by wire but I was having issues with the readings on the Speed Hut oil pressure (dash) gauge and the oil pressure reading through the AEM software on the laptop and add to that the previously stated bonehead mistake, this I believe may have caused problems with the drive by wire. Since I had Nick run an oil pressure sensor "feed" line from the Infinity to the Speed Hut gauge I figured there may be a problem somewhere between the oil pressure sensor and the infinity, so to diagnose I disconnected the Infinity 3 wire at the sensor than ran a temporary 3 wire cable directly from the sensor to the gauge and that took care of the erratic reading on the dash gauge but I still have to figure out were the problem is within the sensor to Infinity to the dash gauge.

Since I didn't actually do the wiring into and out from the Infinity I will have to learn how to read the wiring contact points within the Infinity computer and harness so that I can track where the problem is. Also I may have a grounding issue with the barometric pressure sensor on the supercharger intake manifold, so this will be a trial by fire! Perhaps on your Infinity thread you could point me in the right direction?

Thanks!

Saul
Saul,

First I would open up your Infinity Wizard with your loaded calibration file and go to Basic Sensors. There you can check for three things.

First look at the "raw" and "scaled" values on the dialog box under Oil Pressure Sensor setup and see if there are any readings. If there are not, then there must be something wrong with the input. Then double click on the Oil Pressure Sensor setup and it will most likely be assigned to your Analog13 input. This will need to be checked to be sure the signal wire is pinned properly on the Infinity which I'll cover in a moment.

Second, while you have this dialog box open, check to make sure that you have the right sensor assigned. There are about 6 pre-loaded sensors available to choose from. I have an AEM 150 PSI sensor and just simply selected the respective sensor from the drop-down box. If your sensor is not listed then you'll need to contact the sensor manufacturer and get the voltage scaling particulars for your sensor and you can key the scale in here on this dialog.

Third, on the "Basic Sensors" dialog box, you will find a button called Pin Out. This is really cool. By clicking on this it will tell you exactly what wires going in and coming out of the Infinity are pinned to what. This is a live representation of what is going on. So if or when you change wires on the Infinity, it will know which inputs or outputs it's assigned to. Let's assume under the first check I referenced above you find that your oil pressure sensor is assigned to Analog13. When you open up this pinout chart click on the Input_Output bar at the top to resort the list and then quickly find Analog13. Once found you can then identify what pin the oil pressure signal input is assigned to. In my case, it is assigned to C2-18. Once you have this knowledge you can then go right to that wire coming out of the Infinity and trace it down to find the problem if it is a wiring issue.

Lastly, on your barometric pressure sensor, can you describe what it is or is not doing? Perhaps I can help point you in the right direction there if you can give me some additional input. Not knowing much about this issue without a description I would still suggest referencing the same comments above as I described in the first check to see if you are getting raw and scale reads on that sensor.

Trevor
 
#160 ·
Thanks Trevor,

I will follow your advise in a couple of days to see what I can find

Saul
 
#161 ·
Fuel oscillations (continued)

It's been awhile since my last post. Life has been busy and I haven't had a chance to get the car out until this last weekend. Last I posted, I was having some wideband oscillations under low load around 3,000 RPM causing a hunting effect. To solve I throttled back the Lambda Feedback P and I gains. I ramped out the "I" gain quicker at lower engine speeds. Additionally, I had to cut the "P" gain in half at 1500 RPM and below. I've attached the before and after PIDs for reference. Lastly, you can see from the data log at less than 40 kPa MAP (low load) and cruising steady at 3,000 rpm, I have virtually no oscillation. The hunting is now gone when driving.

During this outing, and finally having the VE table all dialed in with the new injectors, I noticed that some of my snap, crackle and pop was back. One more serving of Rice Krispies. However, the problem last time was on a high RPM decel. On the decel the engine would backfire. This has long been resolved (as identified in one of my previous posts). This new behavior is different. Instead of backfiring on the decel, it will backfire just after a high RPM run followed by a quick decel (i.e. clutch in) and then getting back on the accelerator quickly and aggressively. So now it will pop one time on the accel, not the decel.

I am going to focus on working the WallWetting fuel trims (explained in an earlier post) and/or the "Decel Cut Exit Enrich" trim. When the decel cut turns off this trim will add fuel back for a specified amount over a specified period of time and then decay to 0. The idea behind this fuel enrichment is to prevent drivetrain shock. Not sure yet which is the better tuning approach. I will report back once I get this issue figured out.

Stay tuned...

Trevor
 

Attachments

#162 ·
Fuel oscillations (continued)

Reporting back on the fuel oscillations with Decel Fuel Cut enabled. I have not been successful in tuning out the quick pop on acceleration just after a high rev decel. The data logs clearly show it is a lean condition when it backfires as you would expect with a fuel cut. I tuned on this every which way conceivable and no luck. Sooooo, I ended up disabling the Decel Fuel Cut feature and no longer have the problem. Not really the solution I wanted but the problem is now gone. The only reason I wanted to use the fuel cut feature was to try to boost my fuel economy.
 
#163 ·
VE Learn Tool for the Infinity (kinda)

So in a previous post, I pointed out that the AEM Infinity does not have a VE learn tool. It does provide for a NewVE calc channel that allows you to easily plug in the new VE value for a given cell in the table rather easily but the process is tedious as you have to cull through the datalog to find those conditions where you would make the VE changes.

David Borden and I have been trying to figure out a way to "automate" or create our own VE Learn tool to help expedite the VE tuning process. This feature is found in some of the other fuel injection platforms. We both pretty much had our VE tables all dialed in. I had mine dialed in for the new injectors and now ready for my first run with E-85. But, we wanted to sweep the cams to try to get more power and responsiveness out of the engines. With one driving and the other "real-time" tuning we began "road dynoing" the cars to dial in the cams. Well, when you make cam timing changes it alters the Volumetric Efficiency of the engine requiring you to go back and re-work (re-tune) the VE table. So began the exercise to come up with a VE calculator that would make quick work of this re-tune requirement every time we made cam timing changes.

I used Excel as the platform to build the tool. What we envisioned was taking a data log from the Infinity and exporting it as a *.CSV file using AEM's AEMData Analysis software. From here we would import into our Excel VE Calculator and filter the data points based on some parameters to exclude data points that don't match our criteria.

Clearly we have to filter out those datapoints that would not be appropriate for calculating NewVE. So I created a lambda variance variable that is user defined that determines the threshold to consider the datapoint valid. We basically only take the datapoints where the average of the two banks of actual lambda (from the widebands) is within X% of the target lambda. We also pull out other datapoints such as any knock events. The spreadsheet replicates where the two actual wideband lambdas and the target lambda all overlap indicating a steady state of the engine long enough to give a valid dataset to calculate a NewVE.

We key in our variables and run the macro and it builds a new VE table from the data. We can then simply copy and paste right into Infinity Tuner VETable. It really works slick. We can take up to 250,000 datapoints and within a couple of minutes have it analyzed and spit out a new VE table.

This has really quickened the VE Tuning process.

If anyone has a need for a tool like this, let me know as I would be glad to share it. I'm sure it can be improved greatly but it has already really made this part of tuning so much easier.

Trevor
 
#164 ·
Time to run E-85

Well, it is finally time. Everything is done and sorted on gasoline regarding my tuning. I spent quite a bit of time finalizing my cam timing then sorted my VE tuning and finally optimized my ignition timing and knock control. On my gasoline ignition map I am running WOT timing of 21.5* at 3000 RPM and it ramps up to a tick over 26* at redline with zero knock events under a number of different driving conditions.

I just picked up 10 gallons of E-85. Tonight, I will empty out all of the gasoline and dump in the E-85. I am hoping to pick up at least 4* more advance timing across the ignition map on E-85.
 

Attachments

#165 ·
E-85 Tuning finished

The ignition timing for E-85 is finished and the tuning is complete. I am now running 30.3* at WOT/7500 RPM and above with about 4.8* additional timing across the entire RPM range from about 25% through 100% throttle.

Data logs reveal zero knock events. The car runs exceptional and very strong on E-85. With both the Gasoline tuning and E-85 tuning complete, I can now mix the fuels for a true FlexFuel setup.

Trevor
 

Attachments

#166 ·
ITBs?

So what's next? Hankering to do something else to the engine and to play further with the ECU (which continues to amaze) and learn more about tuning with it while perhaps improve the look of the engine bay with a little more nostalgic flair (if that's even possible with a Coyote). I am contemplating ITBs for the Coyote. I love the look of ITBs on these cars but I have reservations which have stemmed from the nightmares I have heard about tuning ITBs for driveability. With where my engine and tuning is at right now, I have excellent driveability and if I have to sacrifice that, then I probably wouldn't entertain a stack setup. Afterall, driveability was one of the reasons I went with the Coyote to begin with.

The Infinity can easily accommodate ITBs by using Throttle based load (Alpha-N) while still maintaining MAP (Speed Density) for fuel calculations (best of both worlds). So I am optimistic that I can tune it with the Infinity. Just not sure it will actually have as good of driveability characteristics as I am currently enjoying.

I really have no experience with ITBs so I am in the research phase of this potential project. If any of you have any experience to share I would really appreciate it. Just collecting input at this point prior to jumping off. Any advice?
 
#170 ·
Quest for ITBs

In doing research on driveability with ITBs it seems pretty standard to tune ITB setups using Throttle (TPS) instead of MAP for the load. My concern has been abandoning the MAP model of tuning since it has been rock solid and where I live, altitude changes vary widely. So as a test, I retuned my VE table and my Ignition table using Throttle% instead of MAP for my ignition tables and VE table.

I quickly realized that my engine was reaching full airflow saturation at about 50% throttle. I didn't really realize this when I was tuned under MAP load model. But under the throttle model, it became quickly evident. So basically at about 50% throttle and above I reach maximum volumetric efficiency. So I really don't have much control at the top end of the throttle movement. 50% and above is pretty much WOT. This discovery caused me to re-think how I have my DBW throttle curve mapped along with reconsidering my Throttle% breakpoints on my VE table.

Throttle Breakpoints

I had 10 breakpoints on my VE table equally distributed above 50% and 10 breakpoints at or below 50% (pretty much linear from 0% to 100% throttle). I remapped the axis and now have 3 breakpoints above 50% and 17 breakpoints at or below. This has given me much greater granularity and tunability for the area of VE curve below WOT. Now the axis is much more progressive. I Matched the new breakpoints on the ignition tables as well.

DBW Throttle Curve

This led to me rethinking how I have my DBW throttle tuned. I maintained a fairly linear throttle curve. One of the nagging problems that I have had with Coyote with both the Controls Pack PCM and the Infinity ECU has been the "twitchiness" just off idle with the accelerator. Now that I understood that at 50% and above is pretty much where I am reaching maximum airflow saturation, I completely reworked my DBW throttle map for much greater granularity down low off idle. It now emulates a progressive linkage setup and really works better than it ever has it terms of off-idle driveability with all twitchiness eliminated.

And for the drive test regarding tuning using Throttle% on the load axis... It works great. Just as solid as it did when I was tuning it under MAP. This exercise has been a big success and further bolsters my opinion that I should be able to maintain great driveability characteristics with ITBs on this engine. Or at least I hope so.
 
#173 ·
After reading this thread from the very beginning, to its current state, all I can say is wow. I am completely blown away with the knowledge and experience shared here, and had already answered so many questions I have had.

With that said, I am putting the plans together to build a 33', with a Roush supercharged crate motor and now am completely sold on the AEM Infinity set up.

With the versatility on this system, and seeing a supercharged application brought into the mix, I can't wait to start. Thanks to everyone that has contributed, especially Trevor. I have downloaded all your tuning tips and am close to pulling the trigger on my kit. Just need to offload one rental property :).
 
#174 ·
Shoeless,

Let us know if you need any help along the way and I'm glad there are maybe a few golden nuggets in this thread. My primary goal is to hopefully get others hooked on the Infinity ECU. Selfishly, the more of us out there the better forum support we can all provide each other. A big learning curve for all of us... Safety in numbers.
 
#175 ·
You've helped sell me on the idea of an Infinity too. I knew I wanted to be able to tune my future build and after emailing some questions to AEM I think I'm sold on a standalone instead of a tuning unit on the factory ECU.

I've done all my own tuning for intakes, headers, new turbos, etc. on my WRX with RomRaider (RomRaider - Open Source ECU Tools | RomRaider / RomRaider) so between that and your writeups I'm getting excited to start.

Mind if I send you a PM or two just to confirm some things in the software?
 
#176 ·
q4stix,

No problem. PM me with any questions you have. Glad to help if I can.

If you have already had some tuning experience, I think you will do great with this system. If it is any consolation, I had zero ECU tuning experience before travelling down this path a year ago and have found my way through it. That said, anyone with any shred of tuning experience will certainly be a step ahead of me and certainly shorten their learning curve on this system.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top