Mustang GT Active exhaust controller standalone

[Filmchaser]

After a million failed google searches, unhelpful tech docs and eventually just trying another million combos. I figured out how to activate the valves for 3 positions. (I did not have active exhaust as an option in my vehicle). I’m pretty excited about the progress and will have installed testing this weekend. I have it setup to be selectable for the modes I wrote comparable to OEM modes. For the moment I’m only varying the position in 1 of the modes until I see what variables I want to plug in in real life. In the variable mode (similar to the mustang sport mode) it connects via Bluetooth to the PCM to monitor RPM. Below a certain rpm it stays in partially closed position. Above a specified RPM it fully opens. I may integrate throttle position later but for now this seems to be a way to get things driveable for testing. If no Bluetooth connection or no RPM signal received it defaults to my version of normal mode. I have 3” pipes and prior to the active exhaust I have full dual valve dumps. So with this addition I now can go from “Ninja” to full out “Karen” and several steps in between. This will do everything I had hoped although I did hope to get full variable positioning.

If anyone has more knowledge/info on these actuators please chime in. I can not get infinite position variability using PWM. I don’t know if they accept CAN or some other protocol signals differently or if they are simply not able to set precise positions other than the 3 I’ve found?

Sponsored

 

[Jaymar]

If I were to guess at how the OEM system operates, and this is wild-ass speculation mind you, I would probably say they only use the 3 positions of the valve. There would be very little to gain in an infinitely variable valve that couldn't be achieved with a much cheaper open, close, partial setup. The position of which would probably map to RPM and engine load (manifold vacuum) to determine when to open. Quiet, normal and sport would just lower those thresholds and go up to the higher degrees of open on the valve until it reached race which is probably open all the way all the time. Congratulations though, it looks like you put some good work into cracking it this far.

 

[boB]

I second what Jaymar wrote you have done a great job on making and programming a standalone controller!
Somewhere, in some thread, someplace in this forum, someone provided a chart showing the opening of the valve according to mode/rpm/throttle. Maybe it was three discreet settings but I thought it was more (thinking is a dangerous pastime). Of course I can't find that thread now.

*Update, found it: https://www.mustang6g.com/forums/threads/2018-active-exhaust-install-guide.101913/page-51
The lookup table implies the valve has many positions, if it really does is another question.

 

[Filmchaser]

Awesome! That’s actually very helpful. Just looking at the maps it does look like 3-4 max positions so I suppose I do have it figured out as much as it’s being used. I may go ahead with my throttle position mapping. I can see how it could help vs or in combination with the rpm. Thanks for adding that.

 

[Mach VII]

Awesome! That’s actually very helpful. Just looking at the maps it does look like 3-4 max positions so I suppose I do have it figured out as much as it’s being used. I may go ahead with my throttle position mapping. I can see how it could help vs or in combination with the rpm. Thanks for adding that.

I'm thinking you didn't interpret the tables correctly as there are way more than 3-4 positions. The first table dictates which position table is used and the rest show how the position is varied based on RPM and throttle position. Ford felt the need to make adjustments between data points as small as 1%, not sure why they would put that effort in if it didn't really do something.

 

[Jaymar]

I'm thinking you didn't interpret the tables correctly as there are way more than 3-4 positions. The first table dictates which position table is used and the rest show how the position is varied based on RPM and throttle position. Ford felt the need to make adjustments between data points as small as 1%, not sure why they would put that effort in if it didn't really do something.

That xx$ may get used in a function against another variable to determine ultimate position. And the position may get rounded from the final calculation in the controller as well.

 

[Mach VII]

That xx$ may get used in a function against another variable to determine ultimate position. And the position may get rounded from the final calculation in the controller as well.

I posted those tables and have HP Tuners device/software to view PCM parameters. There is no other strategy other than the tables I posted though I forgot to mention that when in 'Normal' or 'Wet' mode it selects different tables based on current transmission gear. Then, it is based on throttle position and RPM. There are some settings related to idle but that affects only when idling.

Watch this video, it shows the valves, RPM, and you judge throttle position by listening as you watch RPM's. See if it matches up with the tables I posted...

 

[Jaymar]

I posted those tables and have HP Tuners device/software to view PCM parameters. There is no other strategy other than the tables I posted though I forgot to mention that when in 'Normal' or 'Wet' mode it selects different tables based on current transmission gear. Then, it is based on throttle position and RPM. There are some settings related to idle but that affects only when idling.

Watch this video, it shows the valves, RPM, and you judge throttle position by listening as you watch RPM's. See if it matches up with the tables I posted...

I saw at least 4 distinct positions between open and closed in just a quick view. Tough to judge the transition speeds to see if it is adjusting but having that many positions alone seems to make it more likely to be an encoder system with all positions 0-100% as Mach VII is suggesting rather than a stepper design.

 

[Mach VII]

I read that Mopar actuators count the pulses from PCM to determine correct position, Ford can't be that much different...

 

[Jaymar]

If Filmchaser is getting raw access to the motor at the actuator then there is probably a controller on the CAN network that receives commands from the PCM to put at xx position and that controller translates those to movements. Additionally, if the motor is adjustable across a range of positions there will probably be a motor control and an encoder feedback available at the motor.

 

[boB]

I'm thinking you didn't interpret the tables correctly as there are way more than 3-4 positions. The first table dictates which position table is used and the rest show how the position is varied based on RPM and throttle position. Ford felt the need to make adjustments between data points as small as 1%, not sure why they would put that effort in if it didn't really do something.

I suspect there are more than a few valve positions but maybe not as fine as 1%. The mechanism is likely a servo motor, it has four wires: power; ground; and (probably) signal; feedback. We know the PCM can detect if the requested position does not match the reported position, this is my reason for the signal and feedback wire guess.

Most servos use a geartain to drive the output and a potentiometer to measure the rotation. 1% resolution is pretty good (and $$), maybe this one is that good. My guess is the PCM can send 1% resolution but does not check that the feedback is that precise.

This is some good investigation into the workings of the active exhaust! Thanks to all who are doing this work.
If my active exhaust ever acts up I plan to look at the signals on an oscilloscope but for now I am too lazy to do it. ;)

 

[Geosh]

I suspect there are more than a few valve positions but maybe not as fine as 1%. The mechanism is likely a servo motor, it has four wires: power; ground; and (probably) signal; feedback. We know the PCM can detect if the requested position does not match the reported position, this is my reason for the signal and feedback wire guess.

Most servos use a geartain to drive the output and a potentiometer to measure the rotation. 1% resolution is pretty good (and $$), maybe this one is that good. My guess is the PCM can send 1% resolution but does not check that the feedback is that precise.

This is some good investigation into the workings of the active exhaust! Thanks to all who are doing this work.
If my active exhaust ever acts up I plan to look at the signals on an oscilloscope but for now I am too lazy to do it. ;)

Here is the inside. Motor isn’t anything special and doesn’t look like it has a feedback loop.

OP: great work. I went the manual switch route but glad to see this effort isn’t dead.

 

[Filmchaser]

I suspect there are more than a few valve positions but maybe not as fine as 1%. The mechanism is likely a servo motor, it has four wires: power; ground; and (probably) signal; feedback. We know the PCM can detect if the requested position does not match the reported position, this is my reason for the signal and feedback wire guess.

Most servos use a geartain to drive the output and a potentiometer to measure the rotation. 1% resolution is pretty good (and $$), maybe this one is that good. My guess is the PCM can send 1% resolution but does not check that the feedback is that precise.

This is some good investigation into the workings of the active exhaust! Thanks to all who are doing this work.
If my active exhaust ever acts up I plan to look at the signals on an oscilloscope but for now I am too lazy to do it. ;)

Sorry I’m late to reply, life gets hectic. So to kind of recap up to this point. I agree it seems like alot of mapping to only use 3-4 positions but I can also see the possibility of having those conditions be used with (future) actuators. The ones I have I have only been able to get to move to those positions and I’ve tried all types of frequency and duty cycle. So I’m thinking either that’s all they programmed to do using pwm, or there’s another signal that is used. There is no potentiometer that I can see. My best guess at this point is that it uses current sensing to calibrate at first turn on. It moves to the internal stop and the current tells it that is home. Then duty cycle pulse tells it to move. It again reaches its limit at the far end of the cycle. So it could roughly keep consistent positioning based on having those two points. If I’m not wrong, the newer actuators are different and do include an internal feedback for position so that lends itself to the idea that the tables are more detailed for a future purpose. In any case, it is working even if the positioning is somewhat limited. I half thought about making a bracket and just using stepper motors but the challenge of making it work has me plus a screw drive has better holding torque especially in a no power situation.

 

[Filmchaser]

One of the most frustrating parts of this not being able to get any answers on the communication protocol. Like I said I’ve gotten it to work but I’ve had a Ford tech try and a bmw tech try as well as I myself tried to get a direct answer on what signal is used and no one can get a straight answer. I have read in an obscure tech doc that PWM and CAN can be used with some versions of these actuators but that’s all I could find. They’re told “just replace the actuator” or the “PCM handles it” really?? And I couldn’t get anywhere because I’m not an “entity”, “dealership”or “affiliate”.

 

[Filmchaser]

Note-I don’t have the oem actuators to compare, however, from looking up pictures online the below seems to be the case.

A couple of updates:
I was having issues with the valves not fully closing.

1. After much tinkering and having a controller to move positions I realized the BMW actuators do not align with the positioning of the valves using the oem spring. I heard there was guy saying you could use these actuators because they fit but had to “clock it 180”. I opened up and rotated the shaft 180 only to realize after getting it sealed and back together that it’s positioning seems to be based on running to the limit (current spike) and then determine that as 1st position. So, in effect I did nothing but break open the actuator. It reset its positions and I was back to identical directions and positions.
2. I then tried writing different code to invert the position in one side. Still not lining up… in the end was a blessing because that added complexity and additional wiring that I now see I didn’t have to have.
3. Finally out of frustration, I made a jig using an old crescent wrench to bend my own springs.

Success! Both actuators rotate the valves in the sane positions as expected and I have true active exhaust. I’ll continue to tweak the code to adapt for RPM and or throttle position now that I can drive it and see what makes the most sense for my preference. It’s very cool to hear how much difference there is between modes. ( to remind, I have 3” pipes with no resonator).

I’m under the impression that the oem actuators do in fact work differently and it seems that the shaft may have the cut to hold the spring in a 90degree position different than the BMW actuators. This is based on what I found to work and the fact that the springs I made are only a few degrees off to allow correct positions between actuator and valve. Vs the oem spring which essentially creates a 90degree offset.

Anyway, that’s what I’ve found so far. I’ll post some video probably tomorrow of it working on startup
.

Sponsored