In reply to Roobasu :
Yeah I didn't have any issues at all with the torque converter. I did swap the flex plate and pickup wheel and reused the forester TC. It seemed to fit up perfectly.
how can you tell the spindle is too long or if the TC is just not slid back in all the way?
Raining again, I spent some more time disassembling and examining TGV valves :-)
If you examine the manifold and draw a line between the TGV actuator mounting bolt hole and then another perpendicular to it across the center of the circular opening you can mark the position of the actuator motor drive gear where the second line crosses the circle toward the center of the manifold and the position of the actuator sensor where the second line crosses the circle toward the outside of the manifold.
With the actuator removed the TGV butterfly can be easily opened and closed by hand which allows marking which portion of the magnet is over the sensor when the TGV butterfly is open/closed.
The short black magnets are positioned symmetrically with one end of the magnet over the sensor when open and the other end over the sensor when closed. For the driver side TGV the clockwise end of the curved black magnet is over the sensor when open and generates a ~5Vdc sensor signal. For the passenger side TGV the counter-clockwise end of the curved black magnet is over the sensor when open and generates a ~5Vdc sensor signal. Although visually identical the two black magnets differ by having the magnetic poles reversed end to end. When closed the sensor signal is ~0Vdc.
The long white magnets are positioned asymmetrically where the ends never pass over the sensor but by different distances on either side of the center line when open/closed and reversed from driver side to passenger side. The two white magnets appear visually identical and do not differ in the orientation of the magnetic poles. This seems to generate different lower voltages for open on driver side and passenger side as well as different nonzero voltages for closed on the two sides. Presumably the newer Legacy computer knows how to handle it. The old Forester computer does not!
I am now experimenting with reusing the black magnets from my old 2011 Forester engine in the new 2016 Legacy manifold as a magic fix to keep the old Forester computer happy. I plan to make a better test setup using a TGV connector cut off the Legacy wiring harness and remeasure the voltages. I will post what I find before and after the magic TGV fix for comparison. Fingers crossed etc.
Edit
Here is a link to the post by CHayNZ that inspired all the TGV magnet discussion
Hi all. Just created a account for this thread. I am currently swapping a 2017 Forester motor into a 2011. I am almost complete. I have swapped the oil pan, the rocker covers, the timing chain cover, the coolant crossover pipe and EGR system, and the camshaft angle tone rings and all sensors and engine wiring harness.
Pretty much the final thing I have to do is figure out the intake swap so this thread has been of great interest to me and love the progress here!
The motor I got was a pzev motor for me 2017 Forester. Of course it has a totally different EGR setup and the coolant lines for the CVT. Swapping over the crossover coolant pipe allowed me to eliminate one of the coolant ports for the CVT and utilize the 2011 EGR system and not have to modify the engine harness. Of course it left the port coming off the right side head needing to be capped. Just a tip if anyone needs it this doorman plug works great to cap off the unused coolant pipe under the head.
Hopefully there is a resolution soon to this tgv issue soon and I appreciate all you guys are doing!
Great tip about the Dorman caps!
Be sure to keep the black magnets out of the old Forester engine TGVs if you have to return it as a core.
I bought some M6 (ID) flat washer that are about 10mm OD at Ace Hardware to use as magnet spacers as recommended by CHayNZ. As it turns out it would probably be better if they are more like 13 1/2mm. Ace had some fender washers too so I will take my caliper and check those. Might need to file the center hole a bit to get 6mm for the shaft mount. No big deal.
Just an FYI, you will need to replace those every 12 months or so. They degrade rapidly with coolant heat, and the proximity to the exhaust will make them time bombs.
I use them to block off extraneous coolant ports on RX-7s running coolant temps around 170-180 most of the time and they fail regularly enough that I keep spares in the car. Subarus run a bit hotter than that.
What works better if there is room is a short length of coolant hose with a pipe plug clamped into the end. That lasts more or less forever but is ugly.
In reply to Pete. (l33t FS) :
I was a little concerned that it wouldn't hold up in the long run so I'll certainly keep an eye out, thanks for the tip! I really wanted to swap over the upper oil pan to fully eliminate it but I just couldn't justify that much work for a single pipe.
In reply to Sodapopdave :
Nice! I replaced a 13 forester engine with a 15 legacy. I didn't need to replace my oil pan. In fact it was a different bolt pattern.
also I didn't see that you replaced your cam tone wheels. I think that is required.
another tip, the vehicle harness may not reach if you had to move the egr plug over like I did you can gain slack by taking the wire hold down from the firewall
I'm still waiting to implement the tgv fix because the CEL stays on and won't allow cruise or traction control to work.
In reply to jdf60 :
Ah yeah, I did replace the cam tone rings/plates/wheels whatever they're call, lol. Luckily my harness fits nicely after swapping to the 2011 coolant crossover pipe. Really I think the only thing left is the TGV's but I can't pass emissions and registration until that's figured out :-/. I may have to just plug in the old TGV bodies temporary.
I'm tempted to try throwing on the entire 2011 intake to see if there is leaks.
Even better if someone could make a adapter plate to properly mate the early FB25's intakes to the late model FB25 heads.
I went back to Ace Hardware and got some 5mm (nominal ID) fender washers. Their OD is very close to the size of the nut and center disc on the magnet piece. In the picture the nut and magnet are on top of washers. Between are a standard and modified washer. I used a chainsaw file to elongate the hole along the line to fit. The threads are a bit less than 6mm and the flat side of the bolt is already narrower than the hole in the washer.
The reason for the washers is that the white magnets are taller so the black magnets need to be raised closer to the Hall effect sensor. CHayNZ estimated an additional 3mm were needed, I have not measured to see exactly how many of these washers will be needed, yet.
Edit - Here is the new test setup using the connector cut off of the new wiring harness now so I can get better voltage measurements. The orange thing between the 5V and signal wires is the 1K Ohm resistor. Get everything all set and then plug the USB cable into a charger block to supply 5Vdc and read the signal voltage then unplug again.
In reply to Skenton :
Awesome! Are you hoping to avoid having to cross the wiring harness from side or side or do you think that will be necessary?
In reply to Sodapopdave :
I plan to swap the 12Vdc pins 4/5 in the actuator connectors to reverse the motor direction the computer "thinks" it is commanding instead of extending and crossing the harness. They are bidirectional DC motors.
Hopefully new voltage measurements tomorrow.
In reply to Sodapopdave
What all did you do to handle temp sensor, cooling and EGR changes. Based on the Dorman cap I assume you ditched the CVT cooler. What did you mean about the crossover pipe?
Pics appreciated.
In reply to Skenton :
The coolant pipe under the intake, I used the 2011 one. That way I didn't have to change the wiring to the EGR or coolant temp sensor.
I put the motor in today, I just plugged in the 2011 TGV assemblies for now, just hanging out in the engine bay. I opened the butter fly valves on the 2017 intake and reinstalled the actuators to keep them open (they are not plugged into anything). This is only until the TGV swap gets figured out but it runs great! Drove for 30 miles and have a pending code for lean bank 1 (P0171) but I did have some hoses loose so hopefully it doesn't pop the CEL.
Here is a pic of the 2011 coolant pipe on the 2017 engine, it was a direct swap.
So you removed the 2017 CVT cooler, EGR cooler and the EGR valve on the manifold, replaced the 2017 crossover pipe with the 2011 crossover pipe which has an internal EGR cooler and EGR valve and temp sensor mounted on the pipe and fewer coolant outlets? The 2011 exhaust recirculation pipes all bolted up correctly and the only coolant port needing to be capped was on the upper oil pan?
In reply to Skenton :
Yes, exactly.
I drove it to work today - no issues besides my upper rad hose blowing off (bad clamp). The pending P0171 code went away but now have a pending P0420 code for low catalytic efficiency - I've seen it before the motor swap so maybe my cat or O2 sensors are going out or slight exhaust leak - no CEL yet and I don't think it's related to the motor swap. Going to keep driving it until all emission systems tests are complete and hopefully the CEL stays off and I can pass smog check here.
Your comment about wanting to swap the upper oil pan to avoid having to cap the coolant line is interesting.
It looks like the big change 2015+ was broaching the casting and adding the curved hose connection at lower right in this picture for the CVT cooler. I believe the Forester and Legacy lower oil pans are different so the 2 pans would need to move as a pair. Other than a bakers dozen bolts and sealant the upper oil pan looks pretty easy.
I may pull mine off and check the compatibility when I get to that point.
In reply to Sodapopdave :
What sealant(s) did you use?
In reply to Skenton :
I used Permatex Ultra Grey for the timing cover and oil pan, I've used it for years without issue.
I used the new test setup to remeasure the stock 2011 Forester TGV signal voltages and confirmed that both sides show about 5Vdc when fully open and about 0V when closed. This is what the forester computer expects to see. Since we are keeping the Forester computer and moving the Forester wiring harness to the Legacy donor engine those are the signal voltages we must convince the Legacy TGVs to provide.
One fly in the ointment is the 2015+ Legacy (2017+ Forester) change from the TGVs closing toward the center of the manifold to closing toward the outside of the manifold. This means the Forester computer commands the newer Legacy TGVs to move the opposite direction than needed. Because of the left/right mirror image design, swapping the Forester harness left/right as documented by CHayNZ corrects the direction at the expense of having to cut/splice/extend the wiring to reach and misleading the computer about which side it is dealing with in diagnostic messages. I believe reversing the polarity to the 12Vdc TGV actuator motor by using a pin extractor to swap pins 4/5 on the TVG connector is a simpler/better solution.
The second issue is that new white Legacy magnets are longer, taller and have the North/South magnetic poles located differently than the old black Forester TGV magnets. CHayNZ said the white magnets are incompatible with the Forester computer and I concur that they cannot be used to provide the necessary signal voltages. Instead the black Forester TGV magnets must be saved and used to replace the white Legacy TGV magnets along with some minor modifications.
Next up, using the new test setup to remeasure the stock Legacy TGV signal voltages and then replace the white Legacy TGV magnets with the black Forester magnets and record the voltages before starting the other modifications.
Edit - Using the new test set up the stock 2016 Legacy TGV signal voltages for the one piece manifold with white magnets on the driver side open/closed are 3.90V/1.25V and on the passenger side 3.86V/1.39V which are, as expected, pretty far off of what the Forester computer wants. I only checked fully open and closed but I think I will go back and look at the middle range a bit too.
Hi I see this thread is still active, I just put in an FB25 from a 2014 Legacy into my 2013 Forester. I switched over pretty much every sensor and did the coil pack boot trick mentioned previously. The engine will start but will not idle, but once I give it gas it runs perfect. I drove it to the gas station assuming it was bad gas but that's not it. Tried different throttle bodies, cam sensors, maf sensors and I'm really stuck. It is throwing a CEL for misfires on all 4 cylinders. I did a leak down test and everything came back good. New grounds and New plugs btw.
In reply to Samseyfert :
Hi Samseyfert, did you also replace the crankshaft position sensor and the plate behind the flywheel and the tone wheels bolted to the end of the intake camshafts behind the timing chain cover? I don't know if that would cause your exact issue but it would certainly affect spark timing.
Edit: My bad, I assumed they were different but apparently they are the same so disregard what I said.
I believe the Legacy 2013/2014 FB25 were pretty much drop in for the 2013 Forester. The "fun" changes started in 2015.
Try swapping the coils and injectors from different cylinders and see if the misfire moves for starters.
Oops - Reread and saw "all 4 cylinders" instead of "cylinder 4" as originally misread.
Edit
FYI, In the past someone had swapped the coil connectors on one side and had misfires on those 2 cylinders.
Lots of folks trade out the rocker covers to avoid shimming the coils. Maybe worth considering.
In reply to Skenton :
The coils are hooked up properly
Thank you all very much, this is an amazing wealth of knowledge. Really saved the day.
So I just changed the engine in my Sons 2011 Subaru Forester. I installed a 2016 engine. The intake ports on the 2011 are smaller then the 2016 block, so I had to use the 2016 intake. I changed the harness, all sensors, valve covers, coils, timing chain cover, flywheel and crank sensor/plate and the water pump. The 2016 engine donor was in a collision, the water pump was pushed in, found that out on start. Really glad I hadn't put coolant in yet, had to hear it start before going farther. I missed the Cam plates, I didn't change them, and the cam solenoids were thrashed. I just finished about 3 hrs ago. 4 codes P0018 - P0016 - P2006 - P2007, Cam sensors and choke sensors. Purrs excellently, even with the codes.
I need to figure out all of the choke variants, I've read so many of them in this thread my brain is over loaded. I have both intakes still, I gotta re read this a few times after some sleep. Sucks that the intake tubes are different.
The bonus coolant line on the passenger side, I took a bolt snug to the size of the hose and put the clamp on it. Make sure the bolt isn't threaded all the way up, you don't want the fluid to follow the threads out of the hose :P. As a kid one of my Dads friends did it for a destruction derby car, if it works on that kinda abuse... It should be good till the hose goes bad, years and years.
Edit/update. So reading through this while a bit more awake. With the polarity differences it says the chokes are closed, but starts a bit slow, and runs great warm I'm guessing its reading closed, but is actually opened. I've decided I am going to use a piece of Aluminum and gasket material to fill the gaps and install the 2011 intake. Like a 1/8in adapter plate if you may. Keep with the 2011 intake to avoid all the conversion craft. < Not an engineer, more Junkyard trial and error :P (I have lexan laying around, I thought of using that at first but it may get soft and leak at operating temps, so Aluminum I will try :P)
We did the install of the engine with about 3in of snow on the ground, outside under an yard pop up pavilion. So all of that will wait till its a bit nicer outside, feeling in the fingers and toes would be nice. The codes can hang around a bit, it runs really well, I put about 150miles on it this week, mostly getting the antifreeze out of the exhaust from the blown up engine. First 70miles or so, it had one hell of a smoke screen, cooking off the antifreeze out of the exhaust o.o .... After life calms a bit I'll rip apart the old engine and see if its salvageable, spare for my Son to learn on.
Been under the weather, went and got some stuff for my lungs and feeling a bit better so I decided to take some more measurements.
To my surprise the 2011 no-name Forester TGV actuator and 2016 Denso Legacy TGV actuators do not seem to be functionally equivalent. When removed from the manifold the Forester actuator signal is 5V with no magnet near it and 0V when either magnetic pole is placed near the sensor. The Legacy actuator signal on the other hand is 1V with no magnet near it and rises to 5V for one pole and drops to less than a volt but not near zero with the other pole. This goes a long ways toward explaining why I could not make accurate predictions about the signal voltages from the Legacy actuators using the Forester magnets.
So, I bet we want to move the forester actuators/sensors to the Legacy manifold along with the black forester magnets if we are going to make the Forester happy with the TGVs on the donor motor.
Getting tired now so maybe more later.
Edit
On that topic, for the 2011 Forester manifold and TGV actuators the distance from the edge of the manifold case to the bottom of the black magnet is about 7 1/4 mm. If I move both the magnets and actuator to the 2016 Legacy manifold that distance needs to be maintained to correctly position the magnet over the sensor. CHayNZ estimated an additional 3mm was needed. I will break out the new manifold and get actual measurements and see how many of the fender washers I need.
You'll need to log in to post.
