subaru FB25 engine interchange

Thanks for the pictures they help a lot. They look OK now, so I guess just the preview is distorted.

I have a old style whack it with a hammer impact wrench with screwdriver bit adapters. It digs in and helps with boogered Phillips heads, one last time.

In reply to mildensteve :

I think those pics are with the pickup wheels already changed.  The forester had cutouts and the legacy had the tabs before. 
 

I still don't see anything wrong with the forester engine. But I haven't split it to look at rods or pistons yet. Other than the rocker that fell off, but that wasn't what was causing the knock. I'm hoping the loose clearances aren't what allowed the rocker to fall off. Because the legacy seems to have about the same clearance between the valve caps amd rockers. 

In reply to Skenton :

Oh I have one of those too. But was afraid I might crack some plastic woth it!  
 

oh and these are the new ones in the pics taken apart. Wasn't sure if I made that clear. I'll try to take them further apart and get the old one apart today. 
 

gettinf ready to put the old timing cover on the legacy!  Hoping it goes smoothly. 

jdf60 said:

getting ready to put the old timing cover on the legacy!  Hoping it goes smoothly. 

Don't forget to transfer over the flex plate, crank trigger plate, sensor/bracket, and harmonic balancer to the Forester.  I think they are all different on the Legacy. 

In reply to mildensteve :

Thanks for the reminder. I did transfer all that stuff and wondered if I should leave the legacy balancer just in case some harmonics are slightly different. But I used the forester because it was less rusty. I think they were identical though as far as fit and size. I didn't see a part number to check. I guess the rubber compound could be different. 
 

anyway it's all ready to drop in. I plan to test it out with the tgv placed in open position manually and use the Forster tgv with its riser blocks as dummies as suggested here. That won't be good for cold weather starting though so I'd like to figure out the tgv eventually.  I do have several pics of both disassembled that I'll sort out and post here. The internals definitely look incompatible. Different stops and rotation directions and gear/magnet components. Will be tough to figure out a mechanical solution. 
 

im going to start it initially without a belt or coolant or any accessories. Just oil and wiring connected. In case I have to pull it again for some reason.  Will this cause any ecu or other major problems im not aware of?  This is my first modern engine swap and woth my old stuff I never had to worry about running them this way. Other than just don't run them very long without coolant.  Oh I'll also go ahead and bolt up the flywheel and tranny and trans coolant lines to avoid a mess.

In reply to jdf60 :

I think the different fly wheel and trigger plate means different harmonics for the balancer to cancel out so the matching parts may be important.  I would flush the block with water for a while before installing it.  No telling how long it sat and you don't want to pump the stuff that is in there into the radiator and new water pump.  Personally, since it is so easy to install, I would hook up the radiator and run clean water to start and look for leaks.

Let us know how it all goes.

In reply to mildensteve :

Good idea to flush with water. Since I'm putting the old radiator to seal off trans lines anyway I'll do that before I put in the new rad.
The crankcase was sure clean. I took off the oil pan and cleaned everything and checked rod bearings but It was basically spotless. 

Always nice when the previous owner changed the oil regularly. One thing I have done a couple of times when a project vehicle came with a really dirty oil pan is drain then overfill some with diesel and let it sit for a week to soak off as much grunge as possible. Goes in clean and comes out looking as bad as the original dirty oil. I leave the drain plug out for a day or two to let it drip out as much as possible and then do a quick cheap Dino oil and filter change before putting in the "good" stuff.

In reply to Skenton :

For some reason I can only post 1 photo per post. These are some pics of the magnets removed. 
 

the 2013 is always on the left in these pics. 2015 is on the right with the white plastic magnet cover 

Great! Things I see right off.

The shafts look to be the same size/threads but the flat/index is about a quarter turn different

There are twelve teeth on the '13 vs ten on the '15 gear. I assume the tooth size difference is a photo focus illusion and the actuator motor gears are identical

The cast limit stops are different, taller and asymmetric on the '13 vs shorter and symmetrical on the '15

The limit tabs attach to the gear piece not the magnet piece. The two on the '13 one goes straight out and the other about 90 degrees counter clockwise bends down vs the one visible on the '15 which bends down

The magnets do not engage the cast limit stops and can rotate over and past them

The black '13 magnets are shorter than the white '15 magnets

========

Questions

Did I describe the tabs correctly?

Are the magnet and gear pieces identical left/right or mirror image in design and/or indexing?

In reply to Skenton :

I think all of your observations are correct. 
 

To answer that last question, I will have to get the other '13 side apart, even my impact screwdriver had trouble with it. I can easily look at the 15 tonight. I'm going to try to start it and observe what the flappers do as the ignition is turned on and while running. Then I'm wondering if I could try to wire up the '15 tgv into the harness reversing the polarity like you suggested. I'm pretty sure  the magnet sensor will not work properly with the different shaped magnets.  Even if the flappers open as they are supposed to.  But maybe it will work enough to still help in cold starting. 

If I understand correctly, the butterfly valves should close at cold start and be open the rest of the time.

Reversing the polarity of TGV motor power ought to let the old Forester computer set the new Legacy TGVs to the correct open/close state with the magnet sensor reading FUBAR as you say. If the computer just throws a code and does not monkey with the motors while trying to "fix" things or go into some sort of "limp mode" you should be good to drive in the mean time.

So far I am having no luck finding a one piece manifold to experiment with for under $100+ I may have to just bite that bullet once I finish getting my motor out. Making progress a bit at a time as energy allows. Then I will be able to examine my '11 TGVs in detail, so if you have to choose, concentrate on the '15 please.

I will study the pictures some more and try to visualize relative magnet sensor motion. Maybe make a paper/cardboard model.

Edit

One difference between swapping the harness left/right and reversing the polarity is which TGV the computer "thinks" it is moving and sensing. I believe the cases are equivalent for our use but I want to think on it some more

In reply to Skenton :

Well I did get everything together and did a test start. It seemed to start up and run pretty good. But the check engine light is on. Which I expected. I haven't read any codes. I still have some work to do replacing the radiator snd flushing the coolant system and buttoning everything up good. 
 

I did notice that the dummy tgv on the passenger side flips open when I turn on the key. It stays open probably because it is warm. The driver side never moved at all. Very strange. Makes me wonder if a plug or wire is damaged on that side. I'll investigate tomorrow.

But at least it started and ran good and has no knocks! 

In reply to mildensteve :

Something I noticed in the harmonic balancer is although they look the same, the 15 legacy version has the stubby oil pump shaft attached and comes that way if you buy a new one. The 13 has the shaft as a separate piece keyed into the balancer. I did use the original al '13 version just because I was using the timing cover it is matched with. 
so far it seems to run smooth but I'm not sure what to do really. I can easily put the '15 balancer/ shaft setup on if needed.  I think both function the same really. 

Congratulations on a successful first start after the swap.

Looking at the TGV pics, the butterflys move about 75 degrees. The bent tab covers an arc of about 30 degrees and the arc between the cast stops is more than 90 degrees, let's call it 105 degrees for now. The bent tab is attached to the gear part with an index slot to "clock" it properly on the shaft. It bounces back and forth between the stops so the thin brass butterflys do not get banged and bent. None of that needs any change since the magnet part mounts on top of it with its own indexing slot. The black magnets may be lower as well as shorter than the white magnets. That's OK because 10mm washers can be inserted on the shaft as needed between the gear part and the magnet part to get the black magnets suitably close to the sensor as part of the swap/modification.

Still pondering exactly how the magnet clocking needs to be changed, but it looks hopeful since the magnets do not need to interact with the stops and rotate over them instead. Once I can compare the left/right magnet part indexing slots I will have a better idea how to proceed.

Edit

Questions:

Looking at actuator pictures, if the line between the mounting holes is vertical on a clock then I think the drive motor gear is very close to 3 o'clock and the sensor 9 o'clock. Does that agree with hands on observation?

How are the magnets attached to their bracket, glued, screwed, riveted etc? Could they be removed/replaced/swapped?

In reply to Skenton :

Yeah that seems acccurate. I can measure the relative positions a little closer if needed.  The actual motors might be interchangeable. They have some differences but look to be basically the same orientation. Although they do rotate opposite. 
Still focusing on getting it back together and running good for the moment (in my little spare time). And then I'll try to read codes and see where I am there.  I still do want to investigate this issue and figure it out once and for all! 

In reply to mildensteve :

If you have a TGV actuator off your old engine's lower manifold could you mark the location of the sensor on the outside of the actuator then place it back on the manifold long enough to transfer the location mark onto the manifold, remove the actuator again and extend the mark onto the face of the gear/magnet cavity and post a straight on picture of it?

I cannot tell exactly where the magnets need to be to rotate over the sensor when I am hopping back and forth between pictures at different angles.

It looks like it will just take filing new index slots in the magnet bracket, or drilling the slot out round and holding at the correct angle while tightening the shaft nut.

Edit

A really cheap shot way to reverse the voltage out of some Hall effect sensor is to flip the magnet (or sensor) over end to end swapping the magnet North and South Pole fields on the sensor. If you have a small magnetic pickup tool or some such, see if it is attracted to one end of the black magnet and repelled by the other on the assumption the poles are on the long axis. Bonus, are the magnet poles the same or reversed  right/left. Can you tell how the black magnets attach to the bracket? The white magnets look like the metal magnet slides into the white plastic mount. Hmmm.

Thanks

In reply to Skenton :

I can try when I get home this weekend. It appears you are headed down a similar path as the other successful conversion. 

In reply to Skenton :

If you look back at my pics I did something similar by taking a pic of both the sensor and intake flapper while closed and open. But I didn't mark anything and it wasn't very straight on. in the open orientation both versions are straight up and down, basically parallel with the flow path.  That may or may not help much. 

after I get my codes read with and without the dummy tgv attached, I could send you one or both of my intake risers with tgv attached if that would help  I would just like them back later in case I do need them    Not sure that would help you without the newer version  

question: even if you reversed the magnet pole like you said and moved the magnet position so the sensor would pick it up, the motor still rotates opposite to operate the new style flapper  are you still planning on figuring out how to get the older style ecu to rotate it opposite?  Possibly just reversing the polarity like you've mentioned already  

The motors are bidirectional 12Vdc and the computers drive them both ways by reversing the positive and negative supplied to pins 4/5 on the fly. Kind of like a room light with 2 switchs at different doors that can both turn the light on or off.

In 2015 the Legacy TGV butterfly open/closed rotation was reversed to improve mixing for start up emission reasons at the same time the FB20 was upgraded to the one piece plastic manifold and the new computers were programmed to change the direction to rotate for open/close and what sensor signal meant what. Since the left/right TGV rotation is mirror image clockwise and counter-clockwise we can fool the old Forester computer to rotate in the new direction by either swapping the harness left/right or, I believe,  swapping pins 4/5 in the TGV connector. The problem is generating a sensor open/closed signal on pin 1 that will make the computer happy and prevent codes.

The original poster on this topic swapped harness sides and also moved the Forester actuators and magnets to the Legacy, drilled out the indexing holes and rotated the magnets to "work". I am less than happy with my understanding of the details of that last part. The "incompatible" white magnets comment also bothers me.

Conventionally, a 5Vdc Hall effect throttle sensor that uses both magnet poles will output 2.5Vdc with no net magnetic field applied, which happen when the center of the magnet is over the sensor. Moving toward one pole generates an offset voltage up to -2Vdc for a sum of 0.5Vdc low when fully closed and while moving toward the other pole generates an offset voltage up to +2Vdc for a sum of 4.5Vdc high when fully open. That matches what I recall from the graphs in the factory service manuals.

I believe the same Denso part 012010-6110 was used on the 2013-2014 and 2015-2016 Legacy so the sensor probably did not change. Hence my recent fascination with the magnet details :-)

I am making slow progress getting my engine out. Maybe I should go ahead and pull the manifold so I can at least have the old design to measure and evaluate. Flying blind like this is frustrating and I appreciate the pics and feedback.

Edit

OK, I made a half prices offer for a damaged one piece manifold on Ebay also missing one TGV. If they accept I should be able to do most of my examination and testing.  Knock on wood.

Skenton said:

Maybe I should go ahead and pull the manifold so I can at least have the old design to measure and evaluate.

I pulled my intake manifold system to make access to the flex plate bolts easier. Plus I could get the cherry picker head low enough that I did not have to remove the hood to get the engine out of the bay. Go ahead and do it. You know you want to...

In reply to mildensteve :

Exactly!  That's the fun part. Pulling it out and disassembling everything. It's kind of like opening a Christmas present every time I take a cover or piece off and see what's underneath! Especially when I am trying to figure it why my old forester engine was knocking. (Still no idea)

agreed, hood can stay on. And flex plate bolts were easy to get out with the intake off. The whole process was pretty painless really. Hardest part was stabbing the engine back onto the tranny by myself without a load leveler. Took me an hour to get it lined up just right. Would be much easier with two people or a load leveler of some kind.