Installed the F/IC today and now I have a miss? (01 Impreza RS)

[brydon]

Brydon,

I haven't been working very closely with the FIC project, but here's a theory or two:


Your issue could be hardware related, having to do with the response time of the circuitry.

It could be software related as well: since the FIC was designed to work with such a broad range of vehicle, it's not unlikely that there are a few sensors we haven't tested yet.

The sensor is exactly the same as the Scion TC.  Even the configuration is similar with a grounded negative side and a positive/negative signal on the positive wire that leads to the ecu.  I say exactly the same because I they look the same, use the same plug, and ohmed out similar.  I can't confirm they are the same but similar enough.  The thing is the TC has alot more teeth then my RS. The RS has 6 total on the crank with larger teeth. My WRX has 26 (I think) and are much smaller but it uses the expact same sensor.

If you don't mind helping us troubleshoot, I'd appreciate more info about the signal you're creating with your function generator. I can test a similar signal on a few of the FICs here at AEM to see how they compare. At what frequency is the FIC having trouble keeping up? What is the signal's voltage and duty cycle?

Well this is my daily driver as my other car is a caged huge cammed track car.  So trouble shooting is not that much of a problem but I have limited time to leave the car down. I also work alot.  i had to borrow the equipment from a friend so I did a little testing but not a lot. Here is what I have. The problem starts about 200hz but only shows up after awhile. It has to be run for awhile and then the problem gets worse and worse. Voltage was changed betwen about 5-12volts. Didn't change anything and the duty cycle was 15% because That is roughly the same as the stock tooth size compaired to crank size.  I however didn't try many duty cylces. I just confirmed what I thought and was done.

If you wouldn't mind putting an oscilloscope to the sensor while it's on the vehicle running the stock ECU (without the FIC), that would be helpful also.

Thanks,
Scott

I have done this but the scope is not mine and I'm not sure if there is a way so save it and send it to you guys. It is very simple.  I would love to steal a scope from my work but they would frown on that.   ...  Anyway it is a sort of saw tooth wave with about 2volts plus and minus at cranking and between ~3 and 7 volts at idle/cruise.

But the very very very odd thing is that it only misses after running it for a while.  This time is alot less when the injectors are hooked up but still there.  When just the crank is hooked up it is longer like 3-4mins of driving but still there. When the injectors are hooked up it is less than a minute.  The even funnier part is when the crank positive is only T'ed on  the problem is still there. I'm not sure why the factory ecu would care if it can still see the signal but the crank positive input is somehow feeding back into the wire.

I have a friend in another state trying my set up with his AEM F/IC on the exact same car as mine.  Hopefully he will let me know in a few days. That would rule out my box and point towards firmware/hardware/software issue.

Oh, and thanks for the reply I appreciate it!

Brydon

[SB]

OK, let me make sure I understand how the signal on the Impreza looks; when the FIC is not installed, is this the factory signal?

Crank Sensor (+)    2-7 volts (positive and negative)
Crank Sensor (-)     0  volts  (grounded to chassis)



If this is the case, here's how I would wire it:

Crank Sensor (+)  >> FIC Crank Mag In (+)     Intercept this signal, do not leave it connected to the ECU
Crank Sensor (-)   >> FIC Crank Mag In (-)      Intercept this signal, do not leave it connected to the ECU 


FIC Crank Mag Out (+)    >>   ECU Crank Input (+)
FIC Crank Mag Out (-)     >>   ECU Crank Input (-)


I'm hoping you've got all this correct already, but I always like to double-check the basics first. If the 3-7V crank signal is coming across the (-) wire from the crank sensor, you should reverse the polarity of the crank signal going into the FIC:  for instance, Crank Sensor (-)  >>  FIC Crank Mag In (+),    and FIC Crank Mag Out (+) >> ECU Crank Input (-).


I would not recommend simply tapping onto the crank signal  (whichever side is 3-7V) , because the FIC's crank output signal is usually going to be a different voltage than the sensor. For most factory ECUs this is not an issue: they're searching for an edge only and not the actual voltage output of the crank sensor. However, if the sensor voltage is 0-7V and the FIC outputs 0-5V, your ECU could detect that as an "extra" edge:   0v - 7v - 5v - 0v. That would screw things up quite a bit.

[brydon]

OK, let me make sure I understand how the signal on the Impreza looks; when the FIC is not installed, is this the factory signal?

Crank Sensor (+)    2-7 volts (positive and negative)
Crank Sensor (-)     0  volts  (grounded to chassis)



If this is the case, here's how I would wire it:

Crank Sensor (+)  >> FIC Crank Mag In (+)     Intercept this signal, do not leave it connected to the ECU
Crank Sensor (-)   >> FIC Crank Mag In (-)      Intercept this signal, do not leave it connected to the ECU 


FIC Crank Mag Out (+)    >>   ECU Crank Input (+)
FIC Crank Mag Out (-)     >>   ECU Crank Input (-)

You are correct and that is how it was wired at one time.  Then I just T'ed onto the ECU crank input (-) with the FIC's crank mag in (-) and left the FIC Crank MAG out (-) disconnected. It ran exactly the same.

I'm hoping you've got all this correct already, but I always like to double-check the basics first. If the 3-7V crank signal is coming across the (-) wire from the crank sensor, you should reverse the polarity of the crank signal going into the FIC:  for instance, Crank Sensor (-)  >>  FIC Crank Mag In (+),    and FIC Crank Mag Out (+) >> ECU Crank Input (-).


I would not recommend simply tapping onto the crank signal  (whichever side is 3-7V) , because the FIC's crank output signal is usually going to be a different voltage than the sensor. For most factory ECUs this is not an issue: they're searching for an edge only and not the actual voltage output of the crank sensor. However, if the sensor voltage is 0-7V and the FIC outputs 0-5V, your ECU could detect that as an "extra" edge:   0v - 7v - 5v - 0v. That would screw things up quite a bit.

When I tapped on to the factory ECU's crank (+) input to the ecu I used only the FIC crank input (+) with the FIC crank input (-) disconnected. I just want to get a RPM signal but not actually modify the timing signal.  This was just a test and the FIC caused the factory ecu to miss.

I have also used another engine management that opperates exactly like the FIC for ignition.  It was spliced into the ECU crank sensor (+) wire and output a square wave signal only on the positive side just like the FIC.  So the factory ECU would only see the postive edge of the signal instead of the positve and negativ. It ran great and operated great.  The FIC however does not.  This leads me to beleave that the factory ecu only looks for edges and has nothing to do with voltage.  Infact I'm almost positive of this.

[brydon]

Oh, this is how I had it wired at first but I have tried almost every combo you can imagine.

[brydon]

Bumpin it up for an answer maybe?!?!

[Pirken]

any update ? found out anything or have you returned your Fic ?

[larryinkc]

Did you ever get your FIC to work? I have a similar problem with mine. If this would work it is much better than playing with the MAF signal to control the AF ratio.

I have a 2005 Tiburon, 4cyl with an aftermarker SC kit on it. I got the FIC to be able to use a larger set of primary injectors. When I first installed it I used the MAP load table to control the larger injectors. It had a stumble on acceleration from a dead stop. Logged it and the injector duty cycle was cutting back to 1% for a few pulses, causing the stumble, then running OK.

Next I tried a smaller set of injectors in it and let the ECU long term fuel trim reduce the fuel when running NA. I used the MAP load table to add fuel only when  boosting. Now there is a stumble with mild acceleration from 2700 RPM to a little over 3000 RPM when the boost starts and the FIC first starts adding fuel.  If I use moderate to heavy acceleration there is no stumble. Would really like to figure this out and get it to work. Any ideas?

[larryinkc]

Found this earlier post, wonder if the FIC is causing the stumble/miss when changing from the same cylinder to the most recent cylinder for fuel pulse information.



"The F/IC does a hybrid of the two. In steady state modes (when the fuel pulse in successive cylinder events is within a certin window, meaning no accel/decel pump or other such stuff) the F/IC uses the last fuel pulse for the specific cylinder as the basis for the modified fuel pulse. When the pulse is changing, as it does when the engine load condition is changed suddenly, it uses the most recent cylinders fueling information for the next cylinder.

So in short, it uses the same cylinder fuel info if you are cruising and the most recent when things are changing quickly. You get the best of both worlds and you dont have to program in an accelerator pump or any other band-aid. The factory ECU's accel pump works as intended."

[brydon]

Sorry, I have made no progress. I'm waiting on someone else to install another AEM on his RS. That should prove that it's my box not the car. Although I know it's not the car. 

Sounds like you guys have the same issue as me. Even goes away the same way.  The most interesting part is I tryed running mine with just the power and grounds hooked up. and it is fine of course. Then I hooked up just the crank + input for a rpm and it misses. So the AEM is feeding a signal back into the input for some reason.

[SB]

Brydon,

The FIC's Crank Sensor Mag In (+) input is not meant to be "tapped" onto the same wire that is connected to the ECU. There is some circuitry inside the FIC that alters the voltage on the Cam and Crank sensor lines in order to get a more accurate reading, and this will cause the stock ECU to be very unhappy if it is still connected to those sensors. The Crank MAG-In (+) line is connected to power (not directly, but there's some circuitry that eventually ties it to power) and the Crank MAG-In (-) line is connected to GND (again, not directly, there's some circuitry in between the sensor and GND). To sum it up, you MUST intercept the crank sensor wires, do not tap into them.


According to Alldata, the Crank Sensor (+) for the Impreza is located on ECU pin 1 of Connector B135. Alldata says this line should read +/- 7V when the car is running on the stock ECU. The FIC will only supply +/- 5V, and this is not a problem for most OEM ECUs. Certain Honda ECUs were looking for a much higher voltage, and the voltage was mostly RPM-dependent. It's possible, but not likely, that the Subaru ECU is actually checking the Crank voltage like the older Honda ECUs. The Honda ECU threw a check engine lamp and cut fuel when the RPMs were high but the FIC wasn't sending a sufficiently high voltage.

AEM now produces an FIC that has been modified to supply a higher voltage to the Honda ECU, and it's possible this might work with the Impreza. However, the Hondas problems seemed to be RPM related, as opposed to load related, so it's difficult to say that this will help you or not.


Another thing we're working on is certain years of the Mazda Miata appear to be double-firing the injectors, which is screwing with the FIC's fuel pulse calculations, causing those cars to have issues that seem to be load-related, similar to what you're describing. There is a software update in the works to rectify this, I'll let you know how that goes.



I've been recently assigned to help with FIC testing and troubleshooting, but there are other cars that are a higher priority than the Impreza for now. Hopefully the things we learn from other manufacturers will help solve the Subaru problems as well.

SB

[brydon]

It's weird you same it can't be tapped on to the wire. It acts exactly the same tapped as it does intercepting.  I have also tryed doing pull down resistors on the input just incase the signal was to high. It changes nothing. It just a crossover issue.  An even stranger thing is that if you don't have the injectors hooked up the issue takes longer to show up.  If they are it shows up right away. Capacitance issue anyone?

[SB]

Come to think of it, I still know a few people who own Imprezas. I will try to get one over here for testing, but it will probably be a few weeks until that happens. Please let us know how your friend's car turns out.

SB

[brydon]

Come to think of it, I still know a few people who own Imprezas. I will try to get one over here for testing, but it will probably be a few weeks until that happens. Please let us know how your friend's car turns out.

SB

Will do!

[brydon]

I was sent a beta version of new code for the AEM FIC. It seems to have solved the miss but I have not wired the injectors back in. Although my problem was always with the crank sensor. Thanks to Jeremy for sending it to me. I will let everyone know how it goes but I hope for good success.

[Pirken]

is this a other beta than the B21 ? should be fun to test om that could fix my cut on 3500 rpm